conclusion for lung cancer research paper

Lung Cancer Research Results and Study Updates

See Advances in Lung Cancer Research for an overview of recent findings and progress, plus ongoing projects supported by NCI.

Lorlatinib (Lorbrena) is superior to crizotinib (Xalkori) as an initial treatment for people with ALK-positive advanced non-small cell lung cancer, according to new clinical trial results. Treatment with lorlatinib also helped prevent new brain metastases.

The immunotherapy drug durvalumab (Imfinzi) can help people with early-stage small cell lung cancer live longer, results from a large clinical trial show. Three years after starting treatment, nearly 60% of people who received the drug were still alive.

FDA has approved alectinib (Alecensa) as adjuvant therapy for people with lung cancer who have ALK-positive tumors. In a clinical trial, alectinib helped people live longer after surgery without their cancer returning than chemotherapy.

The results of the clinical trial that led to FDA’s 2023 approval of repotrectinib (Augtyro) for lung cancers with ROS1 fusions have been published. The drug shrank tumors in 80% of people receiving the drug as an initial treatment.

A collection of material about the ALCHEMIST lung cancer trials that will examine tumor tissue from patients with certain types of early-stage, completely resected non-small cell lung cancer for gene mutations in the EGFR and ALK genes, and assign patients with these gene mutations to treatment trials testing post-surgical use of drugs targeted against these mutations.

Tarlatamab, a new type of targeted immunotherapy, shrank small cell lung cancer (SCLC) tumors in more than 30% of participants in an early-stage clinical trial. Participants had SCLC that had progressed after previous treatments with other drugs.

For people with lung cancer and medullary thyroid cancer whose tumors have changes in the RET gene, selpercatinib improved progression-free survival compared with other common treatments, according to new clinical trial results.

In the ADAURA clinical trial, people with early-stage lung cancer treated with osimertinib (Tagrisso) after surgery lived longer than people treated with a placebo after surgery. Despite some criticisms about its design, the trial is expected to change patient care.

For certain people with early-stage non-small cell lung cancer, sublobar surgery to remove only a piece of the affected lung lobe is as effective as surgery to remove the whole lobe, new research shows.

Pragmatica-Lung is a clinical trial for people with non-small cell lung cancer that has spread beyond the lungs (stage 4 cancer). The trial will help confirm if the combination of pembrolizumab and ramucirumab helps people with advanced lung cancer live longer.

On August 11, the Food and Drug Administration (FDA) gave accelerated approval to trastuzumab deruxtecan (Enhertu) for adults with non-small cell lung cancer (NSCLC) that has a specific mutation in the HER2 gene. Around 3% of people with NSCLC have this kind of HER2 mutation.

Giving people with early-stage lung cancer the immunotherapy drug nivolumab (Opdivo) and chemotherapy before surgery can substantially delay the progression or return of their cancer, a large clinical trial found.

Atezolizumab (Tecentriq) is now the first immunotherapy approved by FDA for use as an additional, or adjuvant, treatment for some patients with non-small cell lung cancer. The approval was based on results of a clinical trial called IMpower010.

Quitting smoking after a diagnosis of early-stage lung cancer may help people live longer, a new study finds. The study, which included more than 500 patients, also found that quitting smoking delayed the cancer from returning or getting worse.

NCI scientists and their international collaborators have found that the majority of lung cancers in never smokers arise when mutations caused by natural processes in the body accumulate. They also identified three subtypes of lung cancer these individuals.

FDA has approved the first KRAS-blocking drug, sotorasib (Lumakras). The approval, which covers the use of sotorasib to treat some patients with advanced lung cancer, sets the stage for other KRAS inhibitors already in development, researchers said.

Combining the chemotherapy drug topotecan and the investigational drug berzosertib shrank tumors in some patients with small cell lung cancer, results from an NCI-supported phase 1 clinical trial show. Two phase 2 trials of the combination are planned.

Mortality rates from the most common lung cancer, non-small cell lung cancer (NSCLC), have fallen sharply in the United States in recent years, due primarily to recent advances in treatment, an NCI study shows.

In a study of more than 50,000 veterans with lung cancer, those with mental illness who received mental health treatment—including for substance use—lived substantially longer than those who didn’t participate in such programs.

FDA has granted accelerated approval for selpercatinib (Retevmo) to treat certain patients with thyroid cancer or non-small cell lung cancer whose tumors have RET gene alterations. The drug, which works by blocking the activity of RET proteins, was approved based on the results of the LIBRETTO-001 trial.

Osimertinib (Tagrisso) improves survival in people with non-small cell lung cancer with EGFR mutations, updated clinical trial results show. People treated with osimertinib lived longer than those treated with earlier-generation EGFR-targeted drugs.

A large clinical trial showed that adding the immunotherapy drug durvalumab (Imfinzi) to standard chemotherapy can prolong survival in some people with previously untreated advanced small cell lung cancer.

The investigational drug selpercatinib may benefit patients with lung cancer whose tumors have alterations in the RET gene, including fusions with other genes, according to results from a small clinical trial.

FDA has approved entrectinib (Rozlytrek) for the treatment of children and adults with tumors bearing an NTRK gene fusion. The approval also covers adults with non-small cell lung cancer harboring a ROS1 gene fusion.

Clinical recommendations on who should be screened for lung cancer may need to be reviewed when it comes to African Americans who smoke, findings from a new study suggest.

Use of a multipronged approach within hospitals, including community centers, not only eliminated treatment disparities among black and white patients with early-stage lung cancer, it also improved treatment rates for all patients, results from a new study show.

In everyday medical care, there may be more complications from invasive diagnostic procedures performed after lung cancer screening than has been reported in large studies.

The Lung Cancer Master Protocol, or Lung-MAP, is a precision medicine research study for people with advanced non-small cell lung cancer that has continued to grow after treatment. Patients are assigned to different study drug combinations based on the results of genomic profiling of their tumors.

On December 6, 2018, the Food and Drug Administration (FDA) approved atezolizumab (Tecentriq) in combination with a standard three-drug regimen as an initial treatment for advanced lung cancer that does not have EGFR or ALK mutations.

A new study has identified a potential biomarker of early-stage non–small cell lung cancer (NSCLC). The biomarker, the study’s leaders said, could help diagnose precancerous lung growths and early-stage lung cancers noninvasively and distinguish them from noncancerous growths.

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

Lung cancer prediction using machine learning on data from a symptom e-questionnaire for never smokers, formers smokers and current smokers

Roles Conceptualization, Methodology, Writing – original draft

Affiliations Division of Family Medicine and Primary Care, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Solna, Sweden, Academic Primary Health Care Centre, Region Stockholm, Stockholm, Sweden

Roles Data curation, Formal analysis, Methodology, Writing – review & editing

Affiliations Division of Family Medicine and Primary Care, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Solna, Sweden, Regional Cancer Centre Stockholm-Gotland, Region Stockholm, Stockholm, Sweden, Division of Clinical Diabetology and Metabolism, Department of Medical Sciences, Uppsala University, Uppsala, Sweden

Roles Funding acquisition, Methodology, Writing – review & editing

Roles Methodology, Supervision, Writing – review & editing

Affiliation Thoracic Oncology Centre, Karolinska University Hospital, Dept of Oncology-Pathology, Karolinska Institutet, Stockholm, Sweden

Contributed equally to this work with: Lars E. Eriksson, Axel C. Carlsson

Roles Conceptualization, Funding acquisition, Writing – review & editing

Affiliations Division of Nursing, Department of Neurobiology, Care Sciences and Society, Karolinska Institutet, Huddinge, Sweden, School of Health and Psychological Sciences, City, University of London, London, United Kingdom, Medical Unit Infectious Diseases, Karolinska University Hospital, Huddinge, Sweden

Roles Funding acquisition, Methodology, Supervision, Writing – review & editing

* E-mail: [email protected]

• Elinor Nemlander, 
• Andreas Rosenblad, 
• Eliya Abedi, 
• Simon Ekman, 
• Jan Hasselström, 
• Lars E. Eriksson, 
• Axel C. Carlsson

• Published: October 21, 2022
• https://doi.org/10.1371/journal.pone.0276703
• Reader Comments

7 Dec 2023: Nemlander E, Rosenblad A, Abedi E, Ekman S, Hasselström J, et al. (2023) Correction: Lung cancer prediction using machine learning on data from a symptom e-questionnaire for never smokers, formers smokers and current smokers. PLOS ONE 18(12): e0295780. https://doi.org/10.1371/journal.pone.0295780 View correction

The aim of the present study was to investigate the predictive ability for lung cancer of symptoms reported in an adaptive e-questionnaire, separately for never smokers, former smokers, and current smokers.

Patients and methods

Consecutive patients referred for suspected lung cancer were recruited between September 2014 and November 2015 from the lung clinic at the Karolinska University Hospital, Stockholm, Sweden. A total of 504 patients were later diagnosed with lung cancer (n = 310) or no cancer (n = 194). All participants answered an adaptive e-questionnaire with a maximum of 342 items, covering background variables and symptoms/sensations suspected to be associated with lung cancer. Stochastic gradient boosting, stratified on smoking status, was used to train and test a model for predicting the presence of lung cancer.

Among never smokers, 17 predictors contributed to predicting lung cancer with 82% of the patients being correctly classified, compared with 26 predictors with an accuracy of 77% among current smokers and 36 predictors with an accuracy of 63% among former smokers. Age, sex, and education level were the most important predictors in all models.

Methods or tools to assess the likelihood of lung cancer based on smoking status and to prioritize investigative and treatment measures among all patients seeking care with diffuse symptoms are much needed. Our study presents risk assessment models for patients with different smoking status that may be developed into clinical risk assessment tools that can help clinicians in assessing a patient’s risk of having lung cancer.

Citation: Nemlander E, Rosenblad A, Abedi E, Ekman S, Hasselström J, Eriksson LE, et al. (2022) Lung cancer prediction using machine learning on data from a symptom e-questionnaire for never smokers, formers smokers and current smokers. PLoS ONE 17(10): e0276703. https://doi.org/10.1371/journal.pone.0276703

Editor: Ardashir Mohammadzadeh, University of Bonab, ISLAMIC REPUBLIC OF IRAN

Received: May 25, 2022; Accepted: October 11, 2022; Published: October 21, 2022

Copyright: © 2022 Nemlander et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: Data cannot be shared publicly because of limitations in the ethical approval. We are unable to share data sets due to GDPR restrictions in Sweden and the EU. Data is available upon reasonable request from [email protected] .

Funding: The present project was funded by Vetenskapsrådet (ref #2016–01712 and #2019–01222)Vårdalstiftelsen(ref #2014–0044), the Strategic Research Area Health Care Science (SFO-V, ref. #2–2764/2018 and 2020), Cancerföreningen i Stockholm (ref #191092), Sjöbergstiftelsen(ref #2022-01-11:7), Astra Zeneca (unrestricted grant), Zero vision cancer and Stiftelsen Einar Belvén. a) Please clarify the sources of funding (financial or material support) for your study. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. No salaries were paid by any of the funding organizations to any of the authors.

Competing interests: The authors have declared that no competing interests exist.

Abbreviations: CI, confidence interval; GP, general practitioner; NRI, normalized relative influence; PHC, primary health care; RAT, risk assessment tool; SD, standard deviation; SGB, stochastic gradient boosting

1. Introduction

Lung cancer is globally the second most commonly diagnosed cancer with over 2,2 million new cases and the leading cause of cancer death, with an estimated 1.8 million deaths in 2020 [ 1 ]. Based on figures for 2019 from the Global Burden of Disease, the incidence of tracheal, bronchus and lung-cancer was reported to be 29.2, (Globally), 68.9 (Western Europe) and 42.5 (Sweden) cases per 100 000 inhabitants. Corresponding figures of mortality were 26.4, 59.8 and 42.4 respectively [ 2 ]. Fortunately, the most important risk factor for lung cancer, smoking [ 3 ], is declining in most Western countries [ 4 ]. With only 16% smokers, Sweden has the lowest prevalence in Europe, half of the European average of 27% [ 4 ]. Never smokers thus constitutes an increasing part of lung cancer patients in Sweden [ 3 ].

While early detection is crucial for prognosis, early symptoms and signs of lung cancer are often non-specific and common [ 5 – 7 ]. This is challenging for general practitioners (GPs), who must assess the likelihood of lung cancer and prioritize investigations and treatments among large groups of patients with non-specific and common symptoms, as for example fatigue or cough. Methods for assessing patients’ likelihood for cancer prior to the investigation of various symptoms and signs that may raise cancer suspicion are lacking. Risk assessment tools (RATs) for cancer, i.e., tools that translate epidemiological risk factors to applicable individual patient assessments, are lacking in primary health care (PHC).

Despite several studies conducted on various tools that assess patients’ cancer risk based on symptom presentation [ 8 – 10 ], there is insufficient evidence that cancer RATs affects the clinical outcome and more research is recommended in a recent health technology assessment conducted 2020 [ 11 ]. Patients in different health care systems should be studied due to variations in both risk factors, symptom presentation and documentation. The Patient EXperience of Bodily Changes for Lung Cancer Investigation (PEX-LC) study has published a model for predicting lung cancer based on reported symptoms and signs among patients having undergone PHC investigation [ 12 ]. PEX-LC has not yet been stratified on smoking status or applied to an unfiltered PHC population [ 12 ]. The rich material of the PEX-LC study provides a starting point for further studies in a PHC context. The aim of the present study was to investigate the predictive ability for lung cancer of symptoms reported in the PEX-LC study, separately for never smokers, former smokers, and current smokers, to create models for future testing in a PHC population.

2.1 Study design

Participants were recruited among 1200 consecutive patients referred for suspected lung cancer between September 2014 and November 2015 to the lung clinic at the Karolinska University Hospital, Stockholm, Sweden [ 12 ]. Of the 670 patients agreeing to participate, 506 patients were later diagnosed with either lung cancer or no cancer. The remaining 164 patients were excluded due to multiple other diagnoses (primarily previous cancer, or a cancer diagnosis other than lung cancer). Additionally, for the present study, two patients whose smoking status could not be ascertained were excluded, resulting in a study sample of 504 patients, of which 310 (61.5%) were diagnosed with lung cancer and the remaining 194 patients (38.5%)with no cancer, see CONSORT flow diagram in Fig 1 .

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

This figure is based on the CONSORT 2010 flow diagram. As this was not a randomised intervention trial, it has been modified to suit this cohort study accordingly. Primary lung cancer (no other cancer); NSCLC: non-small cell lung cancer (adenocarcinoma, n = 200; squamous cell carcinoma, n = 45; not otherwise specified (NOS), n = 5; other NSCLC (adenosquamous lung carcinoma (n = 4), large cell neuroendocrine carcinoma (n = 3); large cell carcinoma, adenoid cystic carcinoma of the lung, adenoid carcinoma with neuroendocrine differentiation, and mucoepidermoid carcinoma of the lung (n = 1, respectively)); SCLC: Small cell lung cancer (includes one individual with combined SCLC) (n = 24); Other LC: carcinoid, n = 9; no histology, n = 17. * Not meeting inclusion criteria: translator required (n = 50), consent withdrawn/missing (n = 15); missing data (n = 5); other reason such as or pain, illness, or other medical condition (n = 25). 1 Other reasons: Limited time of the visit or lack of resources (staff) at the clinic (n = 47); hospitalisations (n = 34); deaths (n = 20). 2 Other: Medical records non-consent (n = 4); unconfirmed, possible lung cancer (n = 3); undiagnosed cancer (n = 2); death before clinical investigation (n = 1); participant withdrew clinical investigation (n = 2); previous lung cancer (n = 1); incomplete modules (n = 12).

https://doi.org/10.1371/journal.pone.0276703.g001

2.2. Questionnaire

Participants completed the PEX-LC adaptive e-questionnaire on a touch screen before their clinical visit. Research assistants were available for help. The number of questions each patient answered differed depending on their symptoms and sensations, with a maximum of 342 potential items: 285 descriptors indicative of the first symptoms/sensations the patient noticed that had caused a change in their lives, and 57 background variables. Medical records of eventual diagnosis were later retrieved, with a follow-up after questionnaire completion of ≥ 1 year.

The questionnaire has been described in detail elsewhere [ 12 , 13 ]. In short, PEX-LC was tailored to allow participants to complete only those items appropriate for the individual’s onset of symptoms or sensations. Background variables included socio-demographics, comorbidities, and smoking habits. Symptoms and sensations included breathing difficulties, cough, phlegm/expectorates, pain/aches/discomfort, fatigue, voice changes, appetite/eating/taste changes, olfactory changes, and fever/chills/sweating. Finally, other changes were also included, for example general physical condition, malaise, or other emotional changes.

2.3. Smoking status

Smoking status was assessed by asking about current and former smoking habits, as well as recent changes in smoking habits. Based on this, participants were classified as never smokers (smoked < 100 cigarettes in their lifetime), former smokers (daily smokers that quit during the year before commencement in the study), or current smokers. Participants having “other smoking habits” could describe these in free text and based on this were classified into one of the three groups never smokers, former smokers, or current smokers, or denoted as having a missing value for this variable.

2.4. Statistical analyses

All analyses were performed separately for the three groups never smokers, former smokers, and current smokers. Categorical data are presented as frequencies and percentages, n (%), while continuous data are given as means with accompanying standard deviations (SDs). Tests of differences between groups were performed using Pearson’s χ 2 -test for categorical data and one-way ANOVA for continuous data. Stochastic gradient boosting (SGB) [ 14 ], implemented in the R package ‘gbm’ version 2.1.8 [ 15 ], was used to predict if a patient had lung cancer or not. A training-test approach was applied to the data, whereby 70% of the observations were randomly selected for training the SGB model, which was then tested on the remaining 30% of the observations to evaluate its performance. The random selection of patients to include in the training data set was performed using stratification on later diagnosed lung cancer status (Lung cancer/Not lung cancer), to ensure equal proportions of lung cancer cases in the training and test data sets and enough cases in each subgroup. The SGB models used a Bernoulli loss function fitted to 10 000 trees, each having a maximum depth of 5 interactions, with a shrinkage (learning rate) of 0.001, a minimum of 10 observations in the terminal nodes of the trees, and a subsampling rate (bag fraction) of 0.5. The optimal number of trees to use for prediction was estimated using 10-fold cross validation.

Using these trees, the SGB models were applied to the training and test data sets to obtain the individual probabilities of having lung cancer for each patient. Cut-off values for classifying patients in the test data set as having lung cancer or not were then constructed by calculating the value of the percentile of these individual probabilities for the training data set that corresponded to the proportion of patients in the training data set that were known to not have lung cancer. A patient in the test data set was then classified as having lung cancer if the individual probability of having lung cancer obtained from the SGB model was larger than this cut-off value, and otherwise classified as not having lung cancer. The performance of the SGB models were evaluated using area under the receiver operator characteristic (ROC) curve (AUC), confusion matrixes, overall accuracy, sensitivity, specificity, positive predicted value, and negative predicted value [ 14 , 16 , 17 ]. Variable importance was estimated by normalized relative influence (NRI), were the relative influences are normalized to sum to 100 [ 18 ]. All statistical analyses were performed using R version 4.1.0 (R Foundation for Statistical Computing, Vienna, Austria), with two-sided P-values < 0.05 considered statistically significant.

2.5. Ethics

All patients gave their written informed consent to participate before their first scheduled visit. The study was carried out according to the Declaration of Helsinki and data were pseudonymized to protect the privacy of the participants. Approval was obtained from the Stockholm Regional Ethics Review Board (Dnr 2014/1290–32). Data are available upon reasonable request from [email protected] .

3.1. Participant characteristics

Table 1 presents the characteristics of the 504 participants according to smoking status: 87 (17.3%) never smokers, 262 (52.0%) former smokers, and 155 (30.8%) current smokers. The participants were at a mean age of 68.3 years, with 50.6% (n = 255) being males and 83.5% (n = 421) being born in Sweden. About one of three (n = 181; 35.9%) participants had a college/university education, and six out of ten (n = 310; 61.5%) participants had lung cancer. The participants differed significantly regarding age (P < 0.001), education level (P = 0.021), and lung cancer status (P < 0.001), with never smokers being the youngest (mean age 68.3 years), having the highest proportion of participants with a college/university education (n = 42; 48.3%), and having the lowest prevalence of lung cancer (n = 33; 37.9%). Former smokers were the oldest (mean age 70.5 years), while current smokers had the lowest proportion of participants with a college/university education (n = 45; 29.0%) and the highest prevalence of lung cancer (n = 114; 73.5%).

https://doi.org/10.1371/journal.pone.0276703.t001

3.2. Performance of the SGB models

Tables 2 and 3 present confusion matrixes and performance measures, respectively, for predictions of lung cancer status for patients in the test datasets using SGB models from the training datasets, according to smoking status. ROC curves for the three groups are given in Fig 2 . The optimal number of trees to use for the predictions were 976 for never smokers, 1245 for former smokers, and 1472 for current smokers. Overall, the SGB models performed well for never smokers and current smokers, with AUC values of 0.735 and 0.822, respectively, and corresponding overall accuracies of 0.815 and 0.771. The performance was considerable worse for former smokers, with an AUC of 0.604 and an overall accuracy of 0.633. While the sensitivity was high for former and current smokers, with values of 0.816 and 0.829, respectively, the sensitivity of 0.700 for never smokers was low. The specificity of 0.882 for never smokers was, on the other hand, high, while former smokers had a very low specificity of 0.333.

https://doi.org/10.1371/journal.pone.0276703.g002

https://doi.org/10.1371/journal.pone.0276703.t002

https://doi.org/10.1371/journal.pone.0276703.t003

3.3. Variable importance

Of the 73 predictors included in the SGB models, 17 (23.3%) had a non-zero influence for never smokers, 36 (49.3%) for former smokers, and 26 (35.6%) for current smokers. The ten predictors with the highest NRI for the SGB models, according to smoking status, are given in Table 4 . Age was the dominant predictor regardless of smoking status, accounting for 51.0% of the influence on the probability of being diagnosed with lung cancer among never smokers, 35.3% of the influence among current smokers, and 28.6% of the influence among former smokers, while education level and sex came in as the second and third most important predictor, respectively, with NRI values between 6 and 10 percent. Of the other variables, “Breathing worse upon exertion” was among the five most important predictors for all smoking groups, with NRI values > 5.0% for both never smokers and current smokers, while “Antibiotics within the past 2 years”, “Cough varied over the day”, “Voice got hoarser”, and “A cold, flu or pneumonia within the past 2 years” were all among the top ten predictors for all smoking groups, with NRI values between 1.9 and 5.7 percent. Notably, the predictor “Haemoptysis/hematemesis” (blood-mixed/brown sputum) had a non-zero influence for never smokers and smokers. For former smokers, “Haemoptysis/hematemesis”was ranked 26 th among the 36 predictors with non-zero influence, with an NRI of only 0.56%.

https://doi.org/10.1371/journal.pone.0276703.t004

4. Discussion

Among never smokers, 17 predictors contributed to predicting lung cancer with 82% of the patients being correctly classified, compared to 26 predictors with an accuracy of 77% for current smokers and 36 predictors with an accuracy of 63% for former smokers.

4.1. Results in perspective

Several large medical record-based cohort studies carried out in PHC and a prospective cohort study on both medical record data and questionnaires from patients referred to a lung cancer clinic have shown hemoptysis, dyspnea, chest pain, cough, appetite loss and/or weight loss to have predictive capability for lung cancer [ 5 , 6 , 19 – 22 ]. All these variables, except hemoptysis, were important predictors for lung cancer in our models.

Notably, the predictor “Haemoptysis/hematemesis” (blood-mixed/brown sputum), was identified as the most important predictor for lung cancer in a previous study of referred patients [ 21 ], but had in the present study a non-zero influence only for former smokers. However, whereas Walter et al. [ 21 ] recruited patients referred to respiratory clinics with any symptoms suspicious of lung cancer noted in the referral letters, with a response rate of 19.5%, our study recruited participants among 1200 consecutive patients referred to a secondary clinic explicitly for suspected lung cancer, with a response rate of 55.8%, making direct comparisons challenging. Furthermore, the incidence of lung cancer in our study was 62% compared with 19% in Walter et al. [ 21 ]. A study investigating changes in the presenting symptoms of lung cancer from 2000–2017 in the UK PHC found that patients with lung cancer presenting with symptoms of haemoptysis are now rare [ 23 ]. This is consistent with our results even though the patients in our study had already been referred to secondary care with suspected lung cancer and had already passed PHC assessment.

Chest pain had a non-zero influence in all models, with back pain having a non-zero influence also for current and former smokers. However, no model included chest or back pain among the ten variables with the highest importance. Besides age, sex and education level were the most important predictors. The same result was found, when the PEX-LC data were analyzed with smoking as a predictor [ 12 ]. Sex being a variable of such high importance was then conjectured to be due to a higher proportion of smokers among women, but in the present analysis, this result holds also for never smokers. Age and sex are also determining factors for treatment effects of lung cancer [ 24 ].

Previous studies from Korea and Sweden have suggested that the incidence of lung cancer in never smokers has increased [ 25 , 26 ]. A Finnish study pooled seven cohorts and studied five risk factors for lung cancer in over 100 000 never smokers [ 27 ]. They found no general increase of lung cancer in never smokers, although the proportion of adenocarcinoma type of lung cancer among women had increased more sharply during the past 10 years. In contrast to the present study, education level was not predictive for lung cancer in the Finnish study, and height was the only factor associated with lung cancer. Regrettably, we did not have access to data on height in our study and could therefore not evaluate its importance.

4.2. Clinical implications

RATs are much needed in PHC settings. Current national clinical guidelines for lung cancer in Sweden give GPs in PHC little support in finding lung cancer, especially among never smokers. Investigation with chest X-ray or low-dose computerized tomography (CT)-scan is recommended for patients having haemoptysis or chest/shoulder pains without other explanations or if a smoking or former smoking patient coughs or has dyspnoea for > 6 weeks. Our prediction model for never smokers ranked chest pain 15 th among 17 predictors with non-zero influence, with an NRI of only 0.36%. This indicates that chest pain is a less useful predictor for patients who have been referred with suspected lung cancer to secondary care. Whether chest pain as a predictor have a higher influence in PHC should be further investigated.

Lung cancer diagnosis via symptoms and signs are sometimes downgraded in importance compared with screening. However, screening programmes for lung cancer have mostly been targeting high-risk smoking individuals [ 28 ], leaving the increasing group of never smokers without structured guidelines for early detection. Moreover, screening programs have partial uptake and limited sensitivity, and cancers occurring outside the screening age groups must be detected via symptoms and signs. Sweden do not currently have a screening program for lung cancer.

4.3. Limitations and strengths

Limitations of the present study include potential recall bias. Some of the questions in the e-questionnaire could have been more precise. Former smokers are a heterogeneous group including individuals who only smoked sporadically for a short period of time as well as patients who have been heavy smokers but quit > 1 year ago, which should contribute to the obtained model having a lower predictive ability for lung cancer in this group. Adding a question of package years would have facilitated the possibility of subgrouping. The e-questionnaire also lacked systematic questions about non-cigarette forms of tobacco and nicotine use. Due to the stratification according to smoking status and the resulting small sample sizes, the predictive value of rare-occurring descriptors may be underestimated due to chance. Since our study included patients referred from PHC for suspected lung cancer, the usefulness of the prediction models in a general PHC population remains to be investigated.

5. Conclusions

Tools assessing the likelihood of having lung cancer among patients with diffuse symptoms are much needed. This is especially true for never smokers who are often detected in a late stage. Our study presents risk assessment models that may be developed into clinical RATs that can help clinicians in assessing a patient’s risk of lung cancer. We welcome future studies conducted in PHC settings on assessable background variables and symptom combinations and their ability to predict lung cancer in patients with different smoking statuses.

• View Article
• PubMed/NCBI
• Google Scholar
• 3. Cancercentrum i Samverkan. Lungcancer- Nationell kvalitetsrapport för 2019. 2019.
• 14. Boehmke B, Greenwell B. Hands-On Machine Learning with R: CRC Press; 2019.
• 16. Everitt B, Palmer CR, Horton RC. The encyclopaedic companion to medical statistics. 2nd ed. ed. Chichester, West Sussex, U.K: John Wiley & Sons; 2011.

Lung Cancer Detection Using Convolutional Neural Network on Histopathological Images

• October 2020
• International Journal of Computer Trends and Technology 68(10):21-24
• 68(10):21-24

• University of Missouri Columbia
• This person is not on ResearchGate, or hasn't claimed this research yet.

Abstract and Figures

Discover the world's research

• 25+ million members
• 160+ million publication pages
• 2.3+ billion citations
• Kountay Dwivedi
• Anshuman Gupta

• Nitila Gokulkrishnan

• Sonia Kukreja

• Fanny Ramadhani
• Hermawan Syahputra
• Rahel Lina Simanjuntak
• Vina Anggraini

• S. P. Angelin Claret
• Santiago Velastegui
• José De La A
• Brad Timana

• Anubhav Kumar

• Pranay Kumar Mishra
• Mohale Tshwale
• Pius Owolawi
• Moses Olaifa
• Vusi Malele
• Athiramol S
• Sudheep Elavidom M
• Surabhi S Nair
• Asha Susan John

• Harikumar Rajaguru

• Michael Snyder

• Maja Stella

• Worawate Ausawalaithong
• Arjaree Thirach

• Wasudeo Rahane
• Himali Dalvi
• Yamini Magar
• Satyajeet Jondhale
• Adv Neural Inform Process Syst
• Alex Krizhevsky
• Ilya Sutskever
• Geoffrey E. Hinton
• William D. Travis
• Elisabeth Brambilla
• Masayuki Noguchi
• David Yankelewitz
• Gerard A Silvestri
• Michael K Gould
• Mitchell L Margolis
• Frank Detterbeck
• Recruit researchers
• Join for free
• Login Email Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google Welcome back! Please log in. Email · Hint Tip: Most researchers use their institutional email address as their ResearchGate login Password Forgot password? Keep me logged in Log in or Continue with Google No account? Sign up

Advertisement

Lung cancer identification: a review on detection and classification

• Non-Thematic Review
• Published: 09 June 2020
• Volume 39 , pages 989–998, ( 2020 )

Cite this article

• Shailesh Kumar Thakur 1 ,
• Dhirendra Pratap Singh   ORCID: orcid.org/0000-0001-5519-3928 1 &
• Jaytrilok Choudhary 1  

8542 Accesses

114 Citations

1 Altmetric

Explore all metrics

Lung cancer is one of the most common diseases among humans and one of the major causes of growing mortality. Medical experts believe that diagnosing lung cancer in the early phase can reduce death with the illustration of lung nodule through computed tomography (CT) screening. Examining the vast amount of CT images can reduce the risk. However, the CT scan images incorporate a tremendous amount of information about nodules, and with an increasing number of images make their accurate assessment very challenging tasks for radiologists. Recently, various methods are evolved based on handcraft and learned approach to assist radiologists. In this paper, we reviewed different promising approaches developed in the computer-aided diagnosis (CAD) system to detect and classify the nodule through the analysis of CT images to provide radiologists’ assistance and present the comprehensive analysis of different methods.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save.

• Get 10 units per month
• Download Article/Chapter or eBook
• 1 Unit = 1 Article or 1 Chapter
• Cancel anytime

Price includes VAT (Russian Federation)

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

Similar content being viewed by others

An Empirical Analysis of Lung Cancer Detection and Classification Using CT Images

Lung Nodule Detection in X-Ray Images: A New Feature Set

Lung Cancer Detection Using CT Scan Images: A Review on Various Image Processing Techniques

Bray, F., Ferlay, J., Soerjomataram, I., Siegel, R., Torre, L., & Jemal, A. Global Cancer Statistics 2018: GLOBOCAN Estimates of incidence and mortality worldwide for 36 cancers in 185 countries: Global Cancer Statistics 2018. CA: a Cancer Journal for Clinicians, 68 .

(2014). Advances in the early detection of lung cancer using analysis of volatile organic compounds: from imaging to sensors. Asian Pacific Journal of Cancer Prevention, 15 (11), 4377–4384.

Belinsky, S. A., Klinge, D. M., Dekker, J. D., Smith, M. W., Bocklage, T. J., Gilliland, F. D., Crowell, R. E., Karp, D. D., Stidley, C. A., & Picchi, M. A. (2005). Gene promoter methylation in plasma and sputum increases with lung cancer risk. Clinical Cancer Research, 11 (18), 6505–6511.

Article   CAS   Google Scholar  

Doria-Rose, V. P., Marcus, P. M., Szabo, E., Tockman, M. S., Melamed, M. R., & Prorok, P. C. (2009). Randomized controlled trials of the efficacy of lung cancer screening by sputum cytology revisited: a combined mortality analysis from the Johns Hopkins Lung Project and the Memorial Sloan-Kettering Lung Study. Cancer, 115 (21), 5007–5017.

Article   Google Scholar  

Oken, M. M., Hocking, W. G., Kvale, P. A., Andriole, G. L., Buys, S. S., Church, T. R., Crawford, E. D., Fouad, M. N., Isaacs, C., Reding, D. J., Weissfeld, J. L., Yokochi, L. A., O’Brien, B., Ragard, L. R., Rathmell, J. M., Riley, T. L., Wright, P., Caparaso, N., Hu, P., Izmirlian, G., Pinsky, P. F., Prorok, P. C., Kramer, B. S., Miller, A. B., Gohagan, J. K., Berg, C. D., & for the PLCO Project Team. (2011). Screening by chest radiograph and lung cancer mortality: the prostate, lung, colorectal, and ovarian (PLCO) randomized trial. JAMA, 306 (17), 1865–1873.

(2011). Reduced lung-cancer mortality with low-dose computed tomographic screening. The New England Journal of Medicine, 365(5), 395–409.

Shen, W., Zhou, M., Yang, F., Yu, D., Dong, D., Yang, C., Zang, Y., & Tian, J. (2016). Multi-crop convolutional neural networks for lung nodule malignancy suspiciousness classification. Pattern Recognition, 61 .

Dhara, A. K., Mukhopadhyay, S., Dutta, A., Garg, M., & Khandelwal, N. (2016). A combination of shape and texture features for classification of pulmonary nodules in lung CT images. Journal of Digital Imaging, 29 .

Shen, W., Zhou, M., Yang, F., Yang, C., & Tian, J. (2015). Multi-scale convolutional neural networks for lung nodule classification BT - information processing in medical imaging, pp. 588–599

Han, F., Wang, H., Zhang, G., Han, H., Song, B., Li, L., Moore, W., Lu, H., Zhao, H., & Liang, Z. (2014). Texture feature analysis for computer-aided diagnosis on pulmonary nodules. Journal of Digital Imaging, 28 , 99–115.

Wei, G., Cao, H., Ma, H., Qi, S., Qian, W., & Ma, Z. (2017). Content-based image retrieval for lung nodule classification using texture features and learned distance metric. Journal of Medical Systems, 42 , 13.

Wei, G., Ma, H., Qian, W., Han, F., Jiang, H., Qi, S., & Qiu, M. (Feb. 2018). Lung nodule classification using local kernel regression models with out-of-sample extension. Biomedical Signal Processing and Control, 40 , 1–9.

Sergeeva, M., Ryabchikov, I., Glaznev, M., & Gusarova, N. (2016). Classification of pulmonary nodules on computed tomography scans. Evaluation of the effectiveness of application of textural features extracted using wavelet transform of image,” in 2016 18th Conference of Open Innovations Association and Seminar on Information Security and Protection of Information Technology (FRUCT-ISPIT) , pp 291–299.

Mao, K., & Deng, Z. (2016). Lung nodule image classification based on local difference pattern and combined classifier. Computational and Mathematical Methods in Medicine, 2016 , 1091279.

Li, X., Shen, L., & Luo, S. (2018). A solitary feature-based lung nodule detection approach for chest X-ray radiographs. IEEE Journal of Biomedical and Health Informatics, 22 (2), 516–524.

Boroczky, L., Zhao, L., & Lee, K. P. (2006). Feature subset selection for improving the performance of false positive reduction in lung nodule CAD. IEEE Transactions on Information Technology in Biomedicine, 10 (3), 504–511.

Sahu, P., Yu, D., Dasari, M., Hou, F., & Qin, H. (2019). A lightweight multi-section CNN for lung nodule classification and malignancy estimation. IEEE Journal of Biomedical and Health Informatics, 23 (3), 960–968.

Masood, A., Yang, P., Sheng, B., Li, H., Li, P., Qin, J., Lanfranchi, V., Kim, J., & Feng, D. D. (2020). Cloud-based automated clinical decision support system for detection and diagnosis of lung cancer in chest CT. IEEE Journal of Translational Engineering in Health and Medicine, 8 , 1–13.

Google Scholar  

Nasrullah, D., Sang, J., Alam, M., Mateen, M., Cai, B., & Hu, H. (Aug. 2019). Automated lung nodule detection and classification using deep learning combined with multiple strategies. Sensors, 19 , 3722.

Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86 (11), 2278–2324.

Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. In F. Pereira, C. J. C. Burges, L. Bottou, & K. Q. Weinberger (Eds.), Advances in Neural Information Processing Systems 25 (pp. 1097–1105). Curran Associates: Inc.

Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv, 1409.1556 .

Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015). Going deeper with convolutions. In 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 1–9).

He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 770–778).

Chapter   Google Scholar  

Huang, G., Liu, Z., van der Maaten, L., & Weinberger, K. (2007 ). Densely Connected Convolutional Networks .

Zeiler, M., & Fergus, R. (2013). Visualizing and Understanding Convolutional Neural Networks (Vol. 8689).

Xie, S., Girshick, R., Dollár, P., Tu, Z., & He, K. (2017). Aggregated residual transformations for deep neural networks. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 5987–5995).

Su, H., Maji, S., Kalogerakis, E., & Learned-Miller, E. (2015). Multi-view convolutional neural networks for 3D shape recognition.

Book   Google Scholar  

Wu, Z., Song, S., Khosla, A., Yu, F., Zhang, L., Tang, X., & Xiao, J. (2015). 3D ShapeNets: a deep representation for volumetric shapes. In 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 1912–1920).

Ronneberger, O., Fischer, P., & Brox, T. (2015). U-Net: convolutional networks for biomedical image segmentation.

Cao, H., Liu, H., Song, E., Ma, G., Xu, X., Jin, R., Liu, T., & Hung, C. (2019). Multi-branch ensemble learning architecture based on 3D CNN for false positive reduction in lung nodule detection. IEEE Access, 7 , 67380–67391.

Han, C., Kitamura, Y., Kudo, A., Ichinose, A., Rundo, L., Furukawa, Y., Umemoto, K., Li, Y., & Nakayama, H. (2019). Synthesizing diverse lung nodules wherever massively: 3D multi-conditional GAN-based CT image augmentation for object detection. In 2019 International Conference on 3D Vision (3DV) (pp. 729–737).

W. Zhu, C. Liu, W. Fan, and X. Xie, DeepLung: Deep 3D Dual Path Nets for Automated Pulmonary Nodule Detection and Classification . 2018.

Silva, G., Valente, T., Silva, A., Paiva, A., & Gattass, M. (2018). Convolutional neural network-based PSO for lung nodule false positive reduction on CT images. Computer Methods and Programs in Biomedicine, 162 .

Patrice, M., Qi, S., Xu, M., Han, F., Zhao, X., & Qian, W. (2018). CNN models discriminating between pulmonary micro-nodules and non-nodules from CT images. Biomedical Engineering Online, 17 .

Monkam, P., Qi, S., Xu, M., Li, H., Han, F., Teng, Y., & Qian, W. (2019). Ensemble learning of multiple-view 3D-CNNs model for micro-nodules identification in CT images. IEEE Access, 7 , 5564–5576.

Huang, X., Shan, J., & Vaidya, V. (2017). Lung nodule detection in CT using 3D convolutional neural networks,” in 2017 IEEE 14th International Symposium on Biomedical Imaging (ISBI 2017) , pp 379–383.

Jiang, H., Ma, H., Qian, W., Gao, M., & Li, Y. (2018). An automatic detection system of lung nodule based on multigroup patch-based deep learning network. IEEE Journal of Biomedical and Health Informatics, 22 (4), 1227–1237.

Setio, A. A. A., Ciompi, F., Litjens, G., Gerke, P., Jacobs, C., van Riel, S. J., Wille, M. M. W., Naqibullah, M., Sánchez, C. I., & van Ginneken, B. (2016). Pulmonary nodule detection in CT images: false positive reduction using multi-view convolutional networks. IEEE Transactions on Medical Imaging, 35 (5), 1160–1169.

Song, Q., Zhao, L., Luo, X., & Dou, X. (2017). Using deep learning for classification of lung nodules on computed tomography images. Journal of Healthcare Engineering, 2017 , 1–7.

Bhandary, A., Prabhu, A., Rajinikanth, V., Krishnan, P., Satapathy, S., Robbins, D., Shasky, C., Zhang, Y.-D., Tavares, J., & Raja, N. (2020). Deep-learning framework to detect lung abnormality – a study with chest X-ray and lung CT scan images. Pattern Recognition Letters, 129 , 271–278.

Lin, C.-J., Shiou-Yun, J., & Chen, M.-K. (2020). Using 2D CNN with Taguchi parametric optimization for lung cancer recognition from CT images. Applied Sciences, 10 , 2591.

Togacar, M., Ergen, B., & Cömert, Z. (2019). Detection of lung cancer on chest CT images using minimum redundancy maximum relevance feature selection method with convolutional neural networks. Biocybernetics and Biomedical Engineering, 40 .

G. Silva, A. Silva, A. Paiva, and M. Gattass, Classification of malignancy of lung nodules in CT images using Convolutional Neural Network . 2020.

Ren, Y., Tsai, M., Chen, L., Jing, W., Li, S., Liu, Y., Jia, X., & Shen, C. (2019). A manifold learning regularization approach to enhance 3D CT image-based lung nodule classification. International Journal of Computer Assisted Radiology and Surgery, 15 .

Kasinathan, G., Jayakumar, S., Gandomi, A., Manikandan, R., Fong, S., & Patan, R. (2019). Automated 3-D lung tumor detection and classification by an active contour model and CNN classifier. Expert Systems with Applications, 134 .

Zhao, X., Liu, L., Qi, S., Teng, Y., Li, J., & Qian, W. (2018). Agile convolutional neural network for pulmonary nodule classification using CT images. International Journal of Computer Assisted Radiology and Surgery, 13 .

Masood, A., Sheng, B., Li, P., Hou, X., Wei, X., Qin, J., & Feng, D. (2018). Computer-assisted decision support system in pulmonary cancer detection and stage classification on CT images. Journal of Biomedical Informatics, 79 .

Nishio, M., Sugiyama, O., Yakami, M., Ueno, S., Kubo, T., Kuroda, T., & Togashi, K. (2018). Computer-aided diagnosis of lung nodule classification between benign nodule, primary lung cancer, and metastatic lung cancer at different image size using deep convolutional neural network with transfer learning. PLoS One, 13 , e0200721.

Alakwaa, W., Nassef, M., & Badr, A. (2017). Lung cancer detection and classification with 3D convolutional neural network (3D-CNN). International Journal of Advanced Computer Science and Applications, 8 .

Download references

Author information

Authors and affiliations.

Computer Science and Engineering, Maulana Azad National Institute of Technology, Bhopal, India

Shailesh Kumar Thakur, Dhirendra Pratap Singh & Jaytrilok Choudhary

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Shailesh Kumar Thakur .

Ethics declarations

Conflict of interest.

The authors declare that they have no conflict of interest.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Thakur, S.K., Singh, D.P. & Choudhary, J. Lung cancer identification: a review on detection and classification. Cancer Metastasis Rev 39 , 989–998 (2020). https://doi.org/10.1007/s10555-020-09901-x

Download citation

Published : 09 June 2020

Issue Date : September 2020

DOI : https://doi.org/10.1007/s10555-020-09901-x

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Lung cancer
• Classification
• Find a journal
• Publish with us
• Track your research

• Study Protocol
• Open access
• Published: 02 August 2024

Understanding the impact of distance and disadvantage on lung cancer care and outcomes: a study protocol

• Daisy McInnerney 1 ,
• Samantha L. Quaife 1 ,
• Samuel Cooke 2 ,
• Lucy Mitchinson 1 ,
• Zara Pogson 3 ,
• William Ricketts 4 ,
• Adam Januszewski 4 ,
• Anna Lerner 4 ,
• Dawn Skinner 5 ,
• Sarah Civello 3 ,
• Ros Kane 6 ,
• Ava Harding-Bell 7 ,
• Lynn Calman 8 ,
• Peter Selby 9 , 10 ,
• Michael D. Peake 11 , 12 &
• David Nelson 2 , 13  

BMC Cancer volume  24 , Article number:  942 ( 2024 ) Cite this article

33 Accesses

Metrics details

Lung cancer is the third most common cancer in the UK and the leading cause of cancer mortality globally. NHS England guidance for optimum lung cancer care recommends management and treatment by a specialist team, with experts concentrated in one place, providing access to specialised diagnostic and treatment facilities. However, the complex and rapidly evolving diagnostic and treatment pathways for lung cancer, together with workforce limitations, make achieving this challenging. This place-based, behavioural science-informed qualitative study aims to explore how person-related characteristics interact with a person’s location relative to specialist services to impact their engagement with the optimal lung pathway, and to compare and contrast experiences in rural, coastal, and urban communities. This study also aims to generate translatable evidence to inform the evidence-based design of a patient engagement intervention to improve lung cancer patients’ and informal carers’ participation in and experience of the lung cancer care pathway.

A qualitative cross-sectional interview study with people diagnosed with lung cancer < 6 months before recruitment (in receipt of surgery, radical radiotherapy, or living with advanced disease) and their informal carers. Participants will be recruited purposively from Barts Health NHS Trust and United Lincolnshire Hospitals NHS Trusts to ensure a diverse sample across urban and rural settings. Semi-structured interviews will explore factors affecting individuals’ capability, opportunity, and motivation to engage with their recommended diagnostic and treatment pathway. A framework approach, informed by the COM-B model, will be used to thematically analyse facilitators and barriers to patient engagement.

The study aligns with the current policy priority to ensure that people with cancer, no matter where they live, can access the best quality treatments and care. The evidence generated will be used to ensure that lung cancer services are developed to meet the needs of rural, coastal, and urban communities. The findings will inform the development of an intervention to support patient engagement with their recommended lung cancer pathway.

Protocol registration

The study received NHS Research Ethics Committee (Ref: 23/SC/0255) and NHS Health Research Authority (IRAS ID 328531) approval on 04/08/2023. The study was prospectively registered on Open Science Framework (16/10/2023; https://osf.io/njq48 ).

Peer Review reports

Globally, lung cancer is a leading cause of cancer mortality and premature death, particularly within communities experiencing deprivation, and is the third most common cancer in the UK [ 1 , 2 ]. NHS England commissioning guidance for optimum lung cancer care recommends management and treatment by a specialist team, with a concentration of experts in one place, providing access to specialised diagnosis and treatment facilities [ 3 ]. However, the complex and rapidly evolving diagnostic and treatment pathways for lung cancer, along with workforce limitations, make achieving this goal challenging [ 4 ]. There are wide variations and inequalities in lung cancer care and survival outcomes across the UK [ 5 ]. Indeed, the deprivation gap (i.e. the survival difference between individuals from the least deprived compared to the most deprived groups) is highest for smoking related cancers, such as lung cancer, compared to other-cancer types [ 6 ]. Evidence suggests that, in the UK, poorer survival rates for people with lung cancer experiencing deprivation compared to more affluent groups are driven by lower screening and treatment rates [ 7 , 8 ]. Broader, structural inequalities related to tobacco-dependence [ 9 ] also drive higher rates of lung cancer incidence and poorer outcomes for people experiencing socio-economic deprivation [ 10 ]. To close this deprivation gap in lung cancer, it is therefore vital to understand and address the factors underlying these lower treatment rates.

One integral factor to consider is patient engagement. Whilst the term ‘patient engagement’ is widely used and may have different meaning across different contexts [ 11 ], in this study, we define this term as the extent to which an individual patient attends, understands and undergoes each investigation, test and treatment that comprises their personal lung cancer pathway, as was recommended by, and agreed mutually with, their clinical care team. Enhancing and supporting patient engagement can improve patient outcomes and care experiences [ 11 ], and in the context of lung cancer may enable greater adherence to recommended diagnostic and treatment pathways. Most work to date has focused more so on improving lung cancer outcomes [ 12 , 13 ] and the quality of the lung cancer services themselves [ 14 ], whilst neglecting to consider how the individual circumstances of people with lung cancer may impact their engagement with available services and support. For example, factors such as an individual’s location in relation to their lung cancer services; their available resources; their language and culture; their prior experiences of and beliefs about healthcare; and availability of social support, may all influence their engagement with care [ 15 , 16 ].

Such factors can be explored systematically using behaviour change models, like the COM-B framework. This theoretical behaviour system model describes three essential and interacting conditions that determine how likely it is that an individual will perform a behaviour (B): their capability (C), opportunity (O) and motivation (M) [ 17 ]. Understanding these interacting factors in relation to individuals’ engagement with their recommended diagnostic and treatment pathway is crucial to identify how best lung cancer patients can be supported to take part in their recommended pathway. For example, one potentially helpful approach is Pathway Navigation. Cancer Alliances report that appointing Pathway Navigators, who provide tailored, individual support to help patients navigate and thus engage with their complex diagnostic test, appointment, and treatment schedules, can double the number of patients receiving lung cancer treatment within the target of 49 days [ 18 ]. Support with pathway navigation may be particularly crucial for individuals who do not have access to informal support from friend or family carers [ 19 ]. Equally, it is critical to understand the experiences of informal carers who are supporting people with lung cancer, to identify areas where additional support from formal healthcare services may be required.

The location of patients relative to the healthcare services they need to access is a particularly important element to consider in relation to an individual’s engagement in their diagnostic and treatment pathway. For instance, services situated in and serving rural or urban areas, are associated with both distinct and overlapping challenges to engagement. The UK consists of large rural and coastal populations that are often characterised by high levels of economic and social deprivation, limited digital infrastructure, poor mental and physical health, high smoking prevalence, and drug and alcohol misuse [ 20 , 21 ]. The Chief Medical Officer for England has recently recognised the importance of better understanding the impact of place on health as well as the urgent need to address health inequalities in rural and coastal areas [ 22 , 23 ]. The challenges faced by rural and coastal communities are often further exacerbated by poor access to healthcare (i.e. long travel distances, poor transport infrastructure, lack of available services) [ 24 , 25 , 26 , 27 ] and workforce limitations (i.e. poor recruitment and retention of healthcare professionals) [ 28 , 29 ]. Urban areas of the UK also experience high levels of economic and social deprivation but typically in more concentrated areas characterised by diverse ethnic communities [ 30 ]. Urban communities also face significant mental and physical health challenges related to unique health inequalities including high population density and heterogeneity [ 31 ], elevated crimes rates [ 32 ], air pollution [ 33 ], lack of green spaces [ 34 ], and poor and unstable housing [ 35 ]. Whilst healthcare access, infrastructure and workforce are typically more developed in urban areas, highly specialised clinical teams are often situated in different hospital settings, requiring significant patient travel and time commitment. Although distances between centres are often shorter than in more rural settings, urban transport systems can be disparate, expensive, and complex to navigate.

There is a distinct lack of research surrounding lung cancer within UK and European settings compared to other tumour sites [ 36 , 37 ]; especially qualitative inquiries that explore experiences across local settings. Further research is needed to gain an in-depth understanding of the individual-level barriers that urban, rural and coastal people living with and affected by lung cancer in the UK face, and to identify facilitators to support engagement. In this study, we will compare and contrast the challenges faced by people with lung cancer and the friends and family members who support them, in urban North East London, to those of predominantly rural and coastal Lincolnshire. The behavioural-science informed approach, theoretically underpinned by the COM-B model, will enable the identification of modifiable factors amendable to intervention to facilitate equitable engagement with the diagnostic and treatment pathway for lung cancer. The aims of this pragmatic and uniquely translational study are:

To explore how lung cancer patients and their informal carers (close family and friends who support people with lung cancer) characteristics and their location in relation to specialist services impact on their capability, opportunity and motivation to attend and participate in their recommended lung cancer diagnostic and treatment pathway in North East London and Lincolnshire.

To generate translatable evidence from both North East London and Lincolnshire to inform the evidence-based design of a patient engagement intervention to improve lung cancer patients’ and informal carers’ participation and experience of the lung cancer care pathway.

Methods/Design

Study design.

This study will use a cross-sectional qualitative interview study design to explore the experiences of people with lung cancer and their informal carers in urban (North East London) and rural (Lincolnshire) areas of England. Guided by the Medical Research Council’s (MRC) framework for the development and evaluation of complex interventions [ 38 ], this study will be conducted in accordance with the intervention ‘development’ phase of the framework, by generating translatable evidence to inform the evidence-based design of a patient engagement intervention that will aim to better support lung cancer patients in engaging in treatment and care pathways. In this study, we are defining patient engagement as the extent to which an individual patient attends, understands and undergoes each investigation, test and treatment that comprises their personal lung cancer pathway, as was recommended by, and agreed mutually with, their clinical care team. This study will be reported in line with the Consolidated Criteria for Reporting Qualitative Research (COREQ) checklist [ 39 ].

Study setting

The study will be conducted in an urban (North East London) and rural and coastal area (Lincolnshire) of England, United Kingdom. It should be noted that the county of Lincolnshire also has urban areas such as the city of Lincoln, although the county as a whole, is predominantly rural in geography, with a significant coastline to the East. People diagnosed with lung cancer and their informal carer’s will be recruited from two NHS trusts: Barts Health NHS Trust and United Lincolnshire Hospitals NHS Trust. Barts Health NHS Trust consists of five hospitals in the City of London and East London (Mile End Hospital, Newham University Hospital, Royal London Hospital, St Bartholomew’s Hospital and Whipps Cross University Hospital) and serves a population of ~ 2.6 million people within an urban area. United Lincolnshire Hospitals NHS Trust consists of four hospitals that cover the county of Lincolnshire (Lincoln County Hospital, Grantham and District Hospital, Pilgrim Hospital Boston, and County Hospital Louth) and serves a population of ~ 700,000 people across a predominately rural area. In the case of Lincolnshire, some people with lung cancer are referred to Nottingham City Hospital as part of Nottingham University Hospitals NHS Trust (NUH) for treatment. Nottingham City Hospital is located in the city of Nottingham within the East Midlands region of England and is located approximately 43 miles from Lincoln city and 80 miles from the East coast of Lincolnshire. Poor road conditions and a lack of accessible public transport can make traveling from the more rural and coastal parts of Lincolnshire to Nottingham, both costly and time consuming [ 40 ]. NUH staff from Nottingham City Hospital will support the identification and recruitment of people with lung cancer and their informal carers who have been referred from United Lincolnshire Hospitals NHS Trust sites for treatment.

Ethical approval and study registration

The protocol for this study was registered on Open Science Framework on October 16th, 2023 ( https://osf.io/njq48 ). Ethical approval was obtained (REC Ref: 23/SC/0255; IRAS ID:328531) from the NHS Oxford B Research Ethics Committee and the NHS Health Research Authority on August 4th, 2023.

Theoretical approach

An exploratory qualitative approach underpinned by the COM-B Model for Behaviour Change [ 17 ] will be applied. This will enable the identification of factors potentially amendable to intervention to initiate health behaviour change (i.e., to facilitate improved engagement with the recommended lung cancer diagnostic and treatment pathway). A person-centred pragmatic epistemological approach will be taken [ 41 ], unpinned by the view that knowledge is based on experience, whilst recognising the unique knowledge of each individual as created by their unique experiences. The pragmatist epistemology supports combining inductive and deductive approaches, and selection of research methods based on their appropriateness for addressing real-world problems [ 41 , 42 ]. Here, qualitative interviews and combined inductive and deductive framework analysis have been selected. This will enable inductive analysis of individual’s unique experiences, challenges and needs; mapped deductively to domains of behaviour change, to generate in-depth, person-centred, translational insights. These in turn will be applied to inform development of a pragmatic intervention to address existing inequities in engagement with the recommended lung cancer pathway. By prioritising a ‘practical understanding’ of these issues, this approach will allow us to understand and address the unique challenges and practical needs of people with lung cancer and their informal carers in urban, rural, and coastal areas.

Participants

This study will recruit people with a confirmed diagnosis of lung cancer within the last six months from three patient cohorts who are in receipt of (1) surgery (2) radical radiotherapy or (3) currently with advanced disease, including both those having active anticancer treatment and specialist palliative/best supportive care (provided they were eligible for treatment). The criterion of six months was chosen because the insights will inform a patient engagement intervention to be delivered early in the diagnostic and treatment pathway (i.e. during or close to the first lung cancer clinic appointment). It is therefore important that participants can recall their experiences of the earlier phases of the investigation and treatment pathway, whilst balancing this with their treatment burden and ability to participate. This study will also recruit informal carers of people with lung cancer with a confirmed diagnosis within the last six months. People diagnosed outside of this timeframe, who do not have capacity to provide informed consent or who are not able to understand the recruitment materials (i.e., participant information sheet/video and informed consent form) with assistance of an interpreter are not eligible to participate in this study.

This study will aim to recruit up to 60 patients and 30–60 informal carers (at least 15 at each site) across both NHS trusts, split evenly between United Lincolnshire Hospitals NHS Trust and Barts Health NHS Trust. We will use a purposive sampling approach to achieve representation from patients receiving different types of treatment (surgical, radical radiotherapy, or advanced cancer) and informal carers. Once ten people have been recruited, recruitment will be targeted following a maximum variation purposive sampling framework [ 43 ] to ensure diversity within the sample in relation to: gender, ethnicity, age, socioeconomic position, stage of disease at diagnosis and area of North East London or Lincolnshire; and for informal carers, these factors along with the type of caring relationship they have to the patient (e.g., friend, family member). This type of recruitment will allow us to explore the experiences of a diverse set of participants and ensure the findings and recommendations are applicable to the diverse range of individuals who may be referred on a lung cancer pathway. The chosen sample size is in line with norms for qualitative research [ 44 , 45 ]. This sample size is required due to the multi-site nature of the study and the diversity of the population [ 46 , 47 , 48 ]. The sample size is sufficient to achieve appropriate information power for a study which is well-designed, theoretically-grounded, and addressing specific objectives [ 48 ].

Recruitment

Participant recruitment and data collection will run between November 2023 and May 2024. All participants will give their informed consent (i.e. either written or verbal) to take part prior to the start of each interview. At both sites, participants will give consent to a member of the research team who is experienced in qualitative interview methods.

People with lung cancer

Patient lists will be pre-screened for people who meet the eligibility criteria by a member of the direct care team at routine clinic meetings at both NHS sites. For each person who is eligible for the study, a member of their direct care team will give a brief overview of the study during their appointment and ask for their consent for a member of the research team to contact them directly. They will then be given and/or sent an information pack (an invitation letter, information sheet and reply slip) and invited to express an interest in taking part either by post, email, or telephone. The information pack will also contain a link to a video-version of the invitation letter, information sheet and reply slip that can be accessed online. In the case of North East London, the information pack, video-version, invitation letter, information sheet and reply slip will also be made available in Sylheti owing to the large Bangladeshi population living in the geographic area served by Barts Health NHS Trust. If the clinical team do not introduce the study to eligible participants during their appointment, they will receive an invitation and information pack by post. A note will be made on the clinical record once a person has received an invitation to ensure they are not re-invited, and to confirm whether they take part in the study.

Informal carers

The information packs distributed to eligible patients will also include information for informal carers, explaining that they are also invited to take part in a separate interview. The reply slip will include an option for either the patient, informal carer, or both to take part in the interview. The option for informal carers to take part will also be mentioned by health care professionals when they introduce the study during appointments and will be mentioned by a member of the research team on the phone to potential patient participants. Patients whose informal carers’ do not want to or are not able to take part in an interview themselves can still be recruited to the study, as can informal carers whose patients do not want to or are not able to take part in an interview themselves.

Data collection

Interviews will be carried out by researchers (SC and LM) experienced in conducting qualitative research and audio-recorded on an encrypted recorder. Each interview will last approximately 1-hour and will take place face-to-face or via telephone or Microsoft Teams, depending on participant preference. We intend to only interview participants once, however, if they are tired or not feeling well during the interview, a follow up meeting can be arranged to complete the interview. Where possible, interviews with people with lung cancer and informal carers will be conducted separately to minimise social desirability bias [ 49 ]. However, as the interview explores sensitive subjects during an emotionally and physically challenging time of the person with lung cancer and informal carers’ lives (following a recent lung cancer diagnosis), the participant can request their patient/informal carer is present during the interview. In this case, the option to conduct the interview as a dyadic interview will be offered [ 50 ]. For participants who are not able to communicate clearly in English, an interpreter will be arranged to assist with both the phone calls to arrange the interview and the interviews themselves.

Interviews will follow a semi-structured topic guide (Additional file 1 ) developed by the research team, wider steering group and with patient and public involvement. The interviews will explore people with lung cancer and informal carers’ capability (physical and psychological), opportunity (physical and social) and motivation (reflexive and automatic) to participate in the lung cancer pathway, based on their individual characteristics and location in relation to the specialist lung cancer centres. This will include exploring factors associated with navigation of complex travel systems across multiple sites to attend appointments; attending and engaging with key touch-points along the pathway (including diagnostic processes; referral; systemic, radio-therapeutic and surgical treatments; palliative and allied health services); and digital consultations. The questions will be adapted for patient, informal carer, or dyadic interviews.

Demographic information including age, gender, ethnicity, post code (as proxy for region, rural-urban residence and socioeconomic position via Index of Multiple Deprivation score), stage of disease, and performance status will be extracted from the medical records of consenting participants. Additional questions in relation to participant characteristics and health behaviours (e.g. lifestyle, smoking behaviour) will be asked as part of the pre-determined interview schedule. The sample will be described in terms of; age, gender, ethnicity, disease stage, location (e.g. rurality/urban), area-level deprivation (converted from postcode to Index of Multiple Deprivation quintile), and performance status.

Data analysis

Following completion of the interviews, the audio-recordings will be professionally transcribed, and a subset checked for accuracy. Transcripts will be psesudonymised and stored securely on the University of Lincoln’s One Drive and Queen Mary University of London’s Data Safe Haven. The qualitative data analysis software package NVivo will be used to support the analysis. A framework approach to applied thematic analysis, as described by Ritchie and Spencer (1994) [ 51 ], will be used to analyse qualitative data. Framework analysis is well-suited to analytical approaches involving multi-disciplinary team members and will enable comparison and interpretation of patterns of themes both within and between North East London and Lincolnshire. This approach will allow us to identify similarities and differences between the two sites, comparing factors affecting patient engagement in rural and urban settings. It will also enable systematic identification of potentially modifiable factors related to participants’ capability, opportunity and motivation to engage that can be targeted by a patient engagement intervention, as well as interactions between these factors and implementation considerations.

The framework method is a five-stage qualitative analysis process involving; (1) Familiarisation, (2) Identifying a thematic framework, (3) Indexing, (4) Charting, and (5) Mapping and Interpretation [ 51 ]. The coding of data will be guided by an inductive and deductive approach, allowing for a data-driven and theory informed development of an analytical framework. The analytical framework will be developed collaboratively between the North East London and Lincolnshire research teams, in consultation with the broader steering group and Patient and Public Involvement and Engagement (PPIE) representatives. The COM-B model will be used to guide the structure of the analytical framework, enabling the grouping of facilitators and barriers to participants’ engagement (capability, opportunity and motivation to engage) with the lung cancer pathway. The analysis will result in a set of recommendations for the proposed patient engagement tool, drawn from the analysis of both the Lincolnshire and North East London interviews, addressing both (a) common principles; and (b) region-specific recommendations. The quantitative demographic data will be summarised and presented as ranges and percentages to describe the overall sample.

Study management and oversight

The study conduct is overseen by a national steering committee, who meet bimonthly to monitor progress, ensure alignment between research sites and with overall project aims, and contribute to results interpretation; application to intervention development; and dissemination. The steering committee is made up of funding body representatives from Cancer Research UK; researchers with expertise in qualitative methods, behavioural science and health inequalities from both Lincolnshire and London; clinicians (oncologists, respiratory physicians, and nurses); NHS cancer pathway managers and administrators; and PPIE representatives. The steering committee are also responsible for delivering a parallel quantitative service evaluation project, that the results of this qualitative study will inform. Alongside the national steering committee, two regional study management groups have been established to manage operational processes at both sites and inform data interpretation and PPIE collaboration.

Reflexivity

Qualitative research is contextual and we as a diverse team of clinical and non-clinical researchers, healthcare professionals and people with lived experience, recognise the importance of reflexivity as a crucial strategy in the process of generating knowledge via qualitative research [ 52 , 53 ]. Reflexivity is considered a major foci for quality control and understanding how it may influence a study should be carefully considered [ 52 ]. Where researchers clearly describe the contextual intersectional relationships between the participants and themselves, this can improve the robustness of the study and generate a deeper understanding of the findings [ 53 ]. This study takes place within two distinct geographic settings, the predominantly rural and coastal county of Lincolnshire and urban North East London. The context is the delivery of lung cancer care in both these settings and the experiences of people diagnosed with lung cancer care and their informal carers who reside in both Lincolnshire and North East London. Both areas have unique social and environmental contexts but are linked by inequalities in lung cancer care. North East London is the London region with the highest level of deprivation and an ethnically diverse community, with over two-thirds of the community from a minority ethnic group. These factors are associated with higher lung cancer mortality and challenges navigating complex healthcare pathways [ 5 , 6 ]. Lincolnshire is not as ethnically diverse with the majority of the population being White British although there is a sizeable Central and Eastern European community. Access to lung cancer care or oncology care for people living in rural and coastal areas is hindered by the uneven geographic distribution of workforce and services [ 37 , 54 ]. We therefore have site-specific research teams that possess a wealth of subject-specific and methodological expertise as well as individual and collective experiences of residing and/or working in these two sites. Reflexivity will also be carefully considered throughout the study by maintaining reflexive logs to document evolving thoughts, biases, and personal reflections during data collection. These will be shared with the wider team at regular team meetings to promote wider reflexivity insights and will be used to help frame and contextualise the interpretation of and meaning of data. The teams will meet regularly throughout data analysis and interpretation, with meetings minuted and reflected upon, to inform the iterative analysis approach and provide core contextual reflections that will be reported with study findings.

Patient and public involvement and Engagement (PPIE)

This study was developed in response to a need identified by Cancer Research UK based on foundational PPIE focus groups, and PPIE is embedded throughout the study lifecycle. The protocol was developed and co-authored by a public contributor with lived experience as a lung cancer carer (AH-B). The study documents (invite, information sheet, consent form and draft interview questions) have also been reviewed by two people living with lung cancer and one carer using a PPIE consultation sheet (Additional file 2 ) and facilitated by our NHS colleagues. Two region-specific PPIE groups have been established; members have lived experience of living with lung cancer, or as informal carers providing support to individuals living with lung cancer. These groups will be consulted at key points throughout the study lifecycle and will play a crucial role during the conduct of this research through providing unique perspectives, support, and guidance. We will also work with these groups to report our findings in accessible formats informed by patients and informal carers’ needs. There will also be opportunities for patients and informal carers to support the dissemination of our findings to clinical and non-clinical audiences. Where appropriate, there will be opportunities for interested members to collaborate and co-author both academic and non-academic outputs. The involvement of PPIE members will extend beyond the conclusion of this study and will play an integral role is shaping and refining the patient engagement tool throughout its subsequent development phases.

Dissemination

We will publish the study findings in peer-reviewed scientific journals and present them at appropriate national and international conferences. Accessible summaries will also be produced and disseminated to people with lung cancer, informal carers, and healthcare professionals. A detailed dissemination plan for this study has been created and agreed upon by the study steering committee. PPIE representatives contributed to the dissemination plan development to ensure our findings will be shared in an inclusive and community-focused way.

This qualitative cross-sectional study will address an urgent need to better understand the experiences and difficulties of lung cancer patients and their informal carers who reside in urban, rural and coastal areas of the UK. More specifically, this study will gather important insight into the capability, opportunity, and motivational factors that may influence lung cancer patients’ engagement in optimal care pathways in these settings. Recent systematic mapping of global cancer screening, prevention, and diagnosis research between 2007 and 2020 points towards a clear disparity in the volume of cancer research across tumour site, with 61% of included studies ( n  = 1762) conducted in colorectal, breast and cervical cancer [ 36 ]. Despite being the leading cause of cancer related deaths globally [ 2 ], only 6.4% percent of studies were in relation to lung cancer [ 36 ]. Furthermore, evidence suggests that our understanding surrounding the development of and engagement in optimal care pathways for people with lung cancer remains in its infancy across the broader health systems [ 4 ], highlighting the need to better understand how individuals engage in these pathways across local settings [ 4 ]. Whilst the world’s largest independent cancer research organisation, Cancer Research UK have prioritised lung cancer research over the last decade [ 55 ], our understanding of lung cancer within a UK context predominantly stems from epidemiological and quantitative inquiries. There remains a dearth of evidence that explores the qualitative experiences of people living with lung cancer who reside in both rural and urban areas in the UK. This is particularly evident in rural settings with recent review evidence identifying only a limited number of qualitive studies ( n  = 9 studies) undertaken in rural areas none of which were from a UK or European setting [ 56 ].

Cancer care pathways are becoming increasingly challenging to deliver and engage with due to their rapidly evolving and complex nature [ 57 ], as well as the multi-faceted individual-level barriers faced by patients unique to urban and rural settings. By identifying and understanding these factors, the study findings can inform the development of tailored services to enable more personalised and patient-centred lung cancer care. Indeed, evidence generated by this study will directly inform the development of a patient engagement intervention that will aim to support lung cancer patients to optimally engage with their recommended care pathway. The MRC has published, and recently updated, their guidance surrounding the development and evaluation of complex interventions, presenting a framework of four phases: (1) development, (2) feasibility/piloting, (3) evaluation, and (4) implementation [ 38 , 58 ]. The current study forms an integral element as part of the ‘development’ phase of the MRC framework, with the qualitative interview findings iteratively integrated with insights from a series of region-specific key stakeholder workshops with a range of healthcare professionals; service managers and co-ordinators; and PPIE representatives. This approach will ensure that experiences and perceptions are gathered from stakeholders across the care continuum to inform robust, patient-centred and theory-and-evidence-based intervention development, with core implementation factors considered throughout. Once we have developed the key components of the patient engagement intervention, we plan to then undertake iterative feasibility and acceptability testing in late 2024 / early 2025. This will be followed by intervention evaluation where we will ascertain the impact of the intervention on key quantitative indicators of pathway engagement, as well as qualitative exploration of patient and carer experience.

Data availability

The dataset(s) that will support the conclusions of this article will be included within the article and its additional file(s).

Abbreviations

Capability Opportunity Motivation Behaviour Model

General Data Protection Regulation

National Health Service

Patient and Public Involvement and Engagement

United Kingdom

United Lincolnshire Hospitals NHS Trust

Oliver AL. Lung Cancer: Epidemiology and Screening. Surg Clin North Am. 2022;102(3):335–44.

Article   PubMed   Google Scholar  

Cancer Research UK. Lung Cancer Statistics. https://www.cancerresearchuk.org/health-professional/cancer-statistics/statistics-by-cancer-type/lung-cancer . Accessed 19 Nov 2023.

NHS England. National Optimal Lung Cancer Pathway: For suspected and confirmed lung cancer: Referral to treatment. https://www.cancerresearchuk.org/sites/default/files/national_optimal_lung_pathway_aug_2017.pdf . Accessed 20 Nov 2023.

Otty Z, Brown A, Sabesan S, Evans R, Larkins S. Optimal care pathways for people with Lung Cancer- a scoping review of the literature. Int J Integr Care. 2020;20(3):14.

Article   PubMed   PubMed Central   Google Scholar  

Conibear J, Nossiter J, Foster C, West D, Cromwell D, Navani N. The National Lung Cancer audit: the impact of COVID-19. Clin Oncol. 2022;34(11):701–7.

Article   CAS   Google Scholar  

Cancer Research Uk. Deprivation gradient for cancer mortality https://www.cancerresearchuk.org/health-professional/cancer-statistics/mortality/deprivation-gradient#heading-One . Accessed 19 Nov 2023.

Forrest LF, Adams J, Rubin G, White M. The role of receipt and timeliness of treatment in socioeconomic inequalities in lung cancer survival: population-based, data-linkage study. Thorax. 2015;70(2):138–45.

Crosbie PAJ, Gabe R, Simmonds I, Hancock N, Alexandris P, Kennedy M et al. Participation in community-based lung cancer screening: the Yorkshire Lung Screening Trial. Eur Respir J 2022;60(5).

Payne NWS, Brown KF, Delon C, Kotrotsios Y, Soerjomataram I, Shelton J. Socio-economic deprivation and cancer incidence in England: quantifying the role of smoking. PLoS ONE. 2022;17(9):e0272202.

Article   CAS   PubMed   PubMed Central   Google Scholar  

Office for National Statitics. Likelihood of smoking four times higher in England’s most deprived areas then least deprived. https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/drugusealcoholandsmoking/articles/likelihoodofsmokingfourtimeshigherinenglandsmostdeprivedareasthanleastdeprived/2018-03-14 . Accessed 25 Nov 2023.

Higgins T, Larson E, Schnall R. Unraveling the meaning of patient engagement: a concept analysis. Patient Educ Couns. 2017;100(1):30–6.

Petty WJ, Paz-Ares L. Emerging strategies for the treatment of small cell lung Cancer: a review. JAMA Oncol. 2023;9(3):419–29.

Li Y, Yan B, He S. Advances and challenges in the treatment of lung cancer. Biomed Pharmacother. 2023;169:115891.

Article   CAS   PubMed   Google Scholar  

Lynch C, Harrison S, Butler J, Baldwin DR, Dawkins P, van der Horst J, et al. An International Consensus on actions to improve Lung Cancer Survival: a modified Delphi Method among Clinical experts in the International Cancer Benchmarking Partnership. Cancer Control. 2022;29:10732748221119354.

Davis RE, Jacklin R, Sevdalis N, Vincent CA. Patient involvement in patient safety: what factors influence patient participation and engagement? Health Expect. 2007;10(3):259–67.

World Health Organisation. Patient Engagement: Technical Series on Safer Primary Care. https://iris.who.int/bitstream/handle/10665/252269/9789241511629-eng.pdf?sequence=1 Accessed 20 Nov 2023.

Michie S, van Stralen MM, West R. The behaviour change wheel: a new method for characterising and designing behaviour change interventions. Implement Sci. 2011;6:42.

United Kingdom Lung Cancer Coalition. Pathways Matter: A review of the implementation of the national optimal lung cancer pathway. https://www.uklcc.org.uk/sites/default/files/2021-06/UKLCC-Pathways-Matter-Report.pdf . Accessed 20 Nov 2023.

Health Education England. Care Navigation: A Competency Framework. 2016. https://www.hee.nhs.uk/sites/default/files/documents/Care%20Navigation%20Competency%20Framework_Final.pdf . Accessed 27 Dec 2023.

Bird W. Improving health in coastal communities. BMJ. 2021;374:n2214.

Philip L, Williams F. Remote rural home based businesses and digital inequalities: understanding needs and expectations in a digitally underserved community. J Rural Stud. 2019;68:306–18.

Article   Google Scholar  

Witty C. Chief Medical Officer’s annual report 2021: Health in coastal communities. https://assets.publishing.service.gov.uk/media/60f98750e90e0703bbd94a41/cmo-annual_report-2021-health-in-coastal-communities-accessible.pdf . 2021. Accessed 10 Nov 2023.

Witty C. Chief Medical Officer’s annual report 2023: Health in an ageing society. https://assets.publishing.service.gov.uk/media/65562ff2d03a8d000d07faa6/chief-medical-officers-annual-report-2023-web-accessible.pdf . 2023. Accessed 10 Nov 2023.

Department for Environment Food & Rural Affairs: Delivering for rural England - the second report on rural proofing. 2022. https://assets.publishing.service.gov.uk/media/6312001de90e077b73f776d7/Delivering_for_rural_England_-_the_second_rural_proofing_report.pdf . Accessed 27 December 2023.

Todd A, Copeland A, Husband A, Kasim A, Bambra C. Access all areas? An area-level analysis of accessibility to general practice and community pharmacy services in England by urbanity and social deprivation. BMJ Open. 2015;5(5):e007328.

Feng X, Feng Z, Astell-Burt T. Perceived public transport infrastructure modifies the association between public transport use and mental health: multilevel analyses from the United Kingdom. PLoS ONE. 2017;12(8):e0180081.

McPeake K, Jeanes L, Nelson D, Selby P, Cooke S, Gussy M, et al. Developing a ‘Living with Cancer’ programme in a rural and coastal setting: experiences of collaborative and innovative co-production across an Integrated Health System. J Cancer Policy. 2023;38:100452.

Charter L, Scully R, Laidlaw A. The word that comes to mind is polymath’: medical students’ perceptions of rural postgraduate careers. Rural Remote Health. 2021;21(3):6750.

PubMed   Google Scholar  

Maclaren AS, Locock L, Skea Z. Valuing place in doctors’ decisions to work in remote and rural locations. Future Healthc J. 2022;9(3):248–51.

PubMed   PubMed Central   Google Scholar  

Ministry of Housing Communities & Local Government. The English Indices of Deprivation 2019 (IoD2019). 2019. https://assets.publishing.service.gov.uk/media/5d8e26f6ed915d5570c6cc55/IoD2019_Statistical_Release.pdf . Accessed 11 Nov 2023.

Fecht D, Fortunato L, Morley D, Hansell AL, Gulliver J. Associations between urban metrics and mortality rates in England. Environ Health. 2016;15(Suppl 1Suppl 1):34.

Department for Environment Food & Rural Affairs. Key Findings, Statistical digest of rural England. 2024. https://www.gov.uk/government/statistics/key-findings-statistical-digest-of-rural-england/key-findings-statistical-digest-of-rural-england . Accessed 12 Nov 2023.

Kelly FJ, Fussell JC. Air pollution and public health: emerging hazards and improved understanding of risk. Environ Geochem Health. 2015;37(4):631–49.

Public Health England. Improving access to greenspace: A new review for 2020. 2020. https://assets.publishing.service.gov.uk/media/5f202e0de90e071a5a924316/Improving_access_to_greenspace_2020_review.pdf . Accessed 13 Nov 2023.

Nijman J, Wei YD. Urban inequalities in the 21st century economy. Appl Geogr. 2020;117:102188.

O’Dowd EL, Merriel SWD, Cheng VWT, Khan S, Howells LM, Gopal DP, et al. Clinical trials in cancer screening, prevention and early diagnosis (SPED): a systematic mapping review. BMC Cancer. 2023;23(1):820.

Nelson D, Selby P, Kane R, Harding-Bell A, Kenny A, McPeake K, et al. Implementing the European code of cancer practice in rural settings. J Cancer Policy. 2024;39:100465.

Skivington K, Matthews L, Simpson SA, Craig P, Baird J, Blazeby JM, et al. A new framework for developing and evaluating complex interventions: update of Medical Research Council guidance. BMJ. 2021;374:n2061.

Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. Int J Qual Health Care. 2007;19(6):349–57.

Rocque GB, Williams CP, Miller HD, Azuero A, Wheeler SB, Pisu M, et al. Impact of Travel Time on Health Care costs and Resource Use by Phase of Care for older patients with Cancer. J Clin Oncol. 2019;37(22):1935–45.

Kaushik V, Walsh CA. Pragmatism as a Research Paradigm and its implications for Social Work Research. Social Sci. 2019;8(9):255.

Patton M. Qualitative Research and Evaluation Methods, 3rd Edition edn: Thousand Oaks: Sage Publications; 2002.

Benoot C, Hannes K, Bilsen J. The use of purposeful sampling in a qualitative evidence synthesis: a worked example on sexual adjustment to a cancer trajectory. BMC Med Res Methodol. 2016;16:21.

Dworkin SL. Sample size policy for qualitative studies using in-depth interviews. Arch Sex Behav. 2012;41(6):1319–20.

Hennink M, Kaiser BN. Sample sizes for saturation in qualitative research: a systematic review of empirical tests. Soc Sci Med. 2022;292:114523.

Jenkins EK, Slemon A, Haines-Saah RJ, Oliffe J. A guide to multisite qualitative analysis. Qual Health Res. 2018;28(12):1969–77.

Philbin MM, Parish CL, Kinnard EN, Reed SE, Kerrigan D, Alcaide ML, Cohen MH, Sosanya O, Sheth AN, Adimora AA, Cocohoba J. Multisite study of women living with HIV’s perceived barriers to, and interest in, long-acting injectable antiretroviral therapy. JAIDS J Acquir Immune Defic Syndr. 2020;84(3):263–70.

Malterud K, Siersma VD, Guassora AD. Sample size in qualitative interview studies: guided by Information Power. Qual Health Res. 2016;26(13):1753–60.

Bergen N, Labonté R. Everything is perfect, and we have no problems: detecting and limiting Social Desirability Bias in qualitative research. Qual Health Res. 2020;30(5):783–92.

Morgan DL, Ataie J, Carder P, Hoffman K. Introducing dyadic interviews as a method for collecting qualitative data. Qual Health Res. 2013;23(9):1276–84.

Ritchie J, Spencer L, Bryman A, Burgess RG. Analysing qualitative data. In.: Routledge, London; 1994.

Berger R. Now I see it, now I don’t: researcher’s position and reflexivity in qualitative research. Qualitative Res. 2015;15(2):219–34.

Dodgson JE. Reflexivity in qualitative research. J Hum Lactation. 2019;35(2):220–2.

Levit LA, Byatt L, Lyss AP, Paskett ED, Levit K, Kirkwood K, et al. Closing the Rural Cancer Care Gap: three institutional approaches. JCO Oncol Pract. 2020;16(7):422–30.

Cancer Research UK. Making Discoveries. Driving Progress. Bringing Hope. Our Research Strategy. 2022. https://www.cancerresearchuk.org/sites/default/files/cancer_research_uk_-_our_research_strategy.pdf . Accessed 22 Nov 2023.

Ali N, Nelson D, McInnerney D, Quaife SL, Laparidou D, Selby P, et al. A systematic review on the qualitative experiences of people living with lung cancer in rural areas. Support Care Cancer. 2024;32(3):144.

Brice SN, Harper P, Crosby T, Gartner D, Arruda E, England T, et al. Factors influencing the delivery of cancer pathways: a summary of the literature. J Health Organ Manag. 2021;35(9):121–39.

O’Cathain A, Croot L, Duncan E, Rousseau N, Sworn K, Turner KM, et al. Guidance on how to develop complex interventions to improve health and healthcare. BMJ Open. 2019;9(8):e029954.

Download references

Acknowledgements

We would like to acknowledge the people living with lung cancer who supported us with the review of the study documents. We would also like to thank Brian Knowles (Senior Research and Evaluation Manager, Cancer Research UK) for commissioning the research and Cancer Research UK for funding it.

Cancer Research UK is a registered charity in England and Wales (1089464), Scotland (SC041666) and the Isle of Man (1103). This research was funded by the Social and Behavioural Research Team, Cancer Research UK (PICATR-2022/100019; PICTAR-2022/100017). http://www.cancerresearchuk.org/ . The protocol manuscript underwent peer review by the funding body as part of the grant application.

Author information

Authors and affiliations.

Centre for Cancer Screening, Prevention, and Early Diagnosis, Wolfson Institute of Population Health, Queen Mary University of London, Charterhouse Square, London, EC1M 6BQ, UK

Daisy McInnerney, Samantha L. Quaife & Lucy Mitchinson

Lincoln Institute for Rural and Coastal Health, College of Health and Science, University of Lincoln, Lincoln, UK

Samuel Cooke & David Nelson

Lincoln County Hospital, United Lincolnshire Hospitals NHS Trust, Lincoln, UK

Zara Pogson & Sarah Civello

Barts Health NHS Trust, London, UK

William Ricketts, Adam Januszewski & Anna Lerner

Pilgrim Hospital, United Lincolnshire Hospitals NHS Trust, Boston, UK

Dawn Skinner

School of Health and Social Care, College of Health and Science, University of Lincoln, Lincoln, UK

Swineshead Patient Participation Group, Swineshead Medical Group, Boston, UK

Ava Harding-Bell

Centre for Psychosocial Research in Cancer, School of Health Sciences, University of Southampton, Southampton, UK

Lynn Calman

School of Medicine, University of Leeds, Leeds, UK

Peter Selby

Lincoln Medical School, College of Health and Science, University of Lincoln, Lincoln, UK

Glenfield Hospital, University of Leicester, Leicester, UK

Michael D. Peake

Cancer Research UK, London, UK

Macmillan Cancer Support, London, UK

David Nelson

You can also search for this author in PubMed   Google Scholar

Contributions

DN and DMcI are Co-Principal Investigators of the study. DN, DMcI and SLQ led on the initial funding applications with support from SC, ZP, WR, AJ, DS, SC, RK, AH-B and PS. DN and DMcI led on the ethics application to the NHS Research Ethics Committee and Health Research Authority with support from SLQ. The study design and methods were developed and modified by all co-authors (DN, DMcI, SC, LM, ZP, WR, AJ, DS, SC, RK, AH-B, LC, AL, PS, SLQ, MP). DN, DMcI, SLQ, LM and SC drafted the first version of the manuscript. All authors critically reviewed the protocol and approved the final manuscript.

Corresponding author

Correspondence to Lucy Mitchinson .

Ethics declarations

Ethics approval and consent to participate.

The study protocol and supporting documents were approved by the NHS Oxford B Research Ethics Committee and the NHS Health Research Authority (REC Ref: 23/SC/0255; IRAS ID: 328531, August 4th, 2023). All participants are required to give their informed consent before they are recruited to the study.

Consent for publication

Not applicable.

Competing interests

MP is a Specialist Clinical Advisor to Cancer Research UK who have funded the study.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1

Supplementary material 2, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

McInnerney, D., Quaife, S.L., Cooke, S. et al. Understanding the impact of distance and disadvantage on lung cancer care and outcomes: a study protocol. BMC Cancer 24 , 942 (2024). https://doi.org/10.1186/s12885-024-12705-9

Download citation

Received : 27 February 2024

Accepted : 25 July 2024

Published : 02 August 2024

DOI : https://doi.org/10.1186/s12885-024-12705-9

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Lung cancer
• Patient experience
• Informal carer
• Rural health
• Urban health
• Qualitative research
• COM-B model

ISSN: 1471-2407

• Submission enquiries: [email protected]
• General enquiries: [email protected]

• DOI: 10.1007/s12672-024-01171-1
• Corpus ID: 271643533

Antibody–drug conjugates treatment of small cell lung cancer: advances in clinical research

• Yuan Meng , Xuerui Wang , +5 authors Fanming Kong
• Published in Discover Oncology 1 August 2024

51 References

Antibody–drug conjugates (adcs) for cancer therapy: strategies, challenges, and successes.

• Highly Influential

Targeting the Notch signaling pathway and the Notch ligand, DLL3, in small cell lung cancer.

Brief report: a phase 1 study evaluating rovalpituzumab tesirine (rova-t) in frontline treatment of patients with extensive-stage small cell lung cancer., the expression of three negative co-stimulatory b7 family molecules in small cell lung cancer and their effect on prognosis, rovalpituzumab tesirine as a maintenance therapy following first-line platinum-based chemotherapy in patients with extensive-stage small cell lung cancer: results from the phase 3 meru study., sacituzumab govitecan, a trop-2-directed antibody-drug conjugate, for patients with epithelial cancer: final safety and efficacy results from the phase 1/2 immu-132-01 basket trial., immunotherapy in small-cell lung cancer: from molecular promises to clinical challenges, antibody drug conjugates for cancer therapy, upregulation of trop-2 quantitatively stimulates human cancer growth, a non-mutated trop2 fingerprint in cancer genetics, related papers.

Showing 1 through 3 of 0 Related Papers

• Open access
• Published: 29 July 2024

Surgical intervention for lung cancer in patients aged 75 and above: potential associations with increased mortality rates—a single-center observational study

• Andrey Kaprin 1 , 2 ,
• Oleg Pikin 3 ,
• Andrey Ryabov 3 ,
• Oleg Aleksandrov 3 , 4 ,
• Denis Larionov 3 &
• Airat Garifullin 3  

Journal of Cardiothoracic Surgery volume  19 , Article number:  471 ( 2024 ) Cite this article

173 Accesses

1 Altmetric

Metrics details

Lung cancer, which is diagnosed two to three times more frequently in patients over the age of 70, is a leading cause of cancer-specific mortality. Given the elevated risk of morbidity and mortality, surgical intervention may not always be the most appropriate primary treatment option. This study aims to evaluate specific risk factors associated with postoperative morbidity and mortality in elderly patients and to optimize patient selection therefore improving surgical outcomes.

Patients and methods

The study encompassed a cohort of 73 patients aged 75 and above who underwent surgical treatment for non-small cell lung cancer (NSCLC) at the Department of Thoracic Surgery of the P. Hertsen Moscow Oncological Research Institute between 2015 and 2021. All patients underwent preoperative evaluation, including PET/CT staging and functional assessment, carried out by a multidisciplinary team comprising thoracic surgeons, anesthesiologists, and other medical specialists.

The investigation revealed a postoperative mortality rate of 5.5% and a postoperative morbidity incidence of 16.4%, with occurrences of atrial fibrillation in 41.6%, persistent air leak in 33.3%, and pneumonia in 25% of complicated cases. At the one-year follow-up, 88% of patients remained free from relapse, whereas at three years, this rate stood at 66%. During the follow-up period, 16 patients (22%) passed away, with a median survival duration of 44 months. Survival rates at one year, three years, and five years were 71%, 66%, and 35%, respectively. Multivariate analysis disclosed several significant factors predicting a complex postoperative period, including stage IIIb ( p  = 0.023), pN1 ( p  = 0.049), pN2 ( p  = 0.030), and central location ( p  = 0.007). Additionally, overall survival was primarily influenced by a Charlson comorbidity index of 6 ( p  = 0.044), stage Ia2 ( p  = 0.033), and the necessity for thoracotomy ( p  = 0.045).

Each case of lung cancer in patients aged 75 and older necessitates an individualized approach. Given the higher mortality rate relative to younger patients, comprehensive risk assessment and preoperative management of underlying comorbidities are imperative, with the involvement of anesthesiologists, intensive care physicians, cardiologists, and other relevant specialists as needed.

Peer Review reports

Introduction

In contemporary society, the steady increase in life expectancy is a prevailing trend, and it is projected that ongoing medical advancements will further contribute to the ageing of the population. By 2050, approximately 17% of the population is anticipated to be over 65 years old [ 1 ]. It is noteworthy that lung cancer, the most prevalent and deadliest of all cancers, predominantly affects individuals aged 70 and older [ 2 ]. For instance, while an 80-year-old US citizen has an expected life span of nine years, the median survival time for lung cancer patients in this age group is merely 14 months. A series of questions related to age-related variations in tumour biology, the influence of concomitant medical conditions on the tolerability of cancer treatment or surgery, and the potential benefits of stereotactic body radiation therapy (SBRT) in enhancing long-term survival in localised, resectable lung cancer necessitate thorough investigation. Nonetheless, a common tendency persists to exclude elderly patients from research studies or set restrictive age criteria at 65 years [ 3 ]. Elderly patients often face exclusion from studies or are denied surgical interventions not only due to diminished functional capabilities but also on account of their age and the perceived risk of complications [ 4 ]. Surgical procedures in elderly patients are characterised by an array of risk factors, encompassing age-related physiological changes resulting in declining organ function and reduced reserve capacity, the presence of comorbid medical conditions, polypharmacy, cognitive and functional impairments, nutritional status concerns, insufficient social support, and issues related to caregiver availability [ 5 ]. In numerous instances, major lung resections become impractical due to reduced lung and cardiovascular capacity [ 6 ]. The conventional methods of preoperative assessment may prove inadequate, necessitating a more comprehensive evaluation that includes exercise tolerance testing, which is sometimes inaccessible. It is well documented that the incidence of complications increases proportionally with age [ 7 ]. To optimise patient selection and, consequently, enhance surgical outcomes, it becomes of paramount importance to comprehend the factors that can impact outcomes in this patient demographic.The main objective of this study was to examine the outcomes of surgical procedures in patients over 75 years of age with non-small cell lung cancer (NSCLC) and to identify the main risk factors that contribute to a complicated postoperative course and reduced long-term survival rate. This article adheres to the STROBE reporting checklist.

We gathered data on patients aged 75 and above who underwent lung cancer surgery at the Department of Thoracic Surgery of the P. Hertsen Moscow Oncology Research Institute between 2015 and 2021. Inclusion criteria encompassed individuals aged 75 years or older, a confirmed diagnosis of non-small cell lung cancer (NSCLC), and surgeries performed by our clinical team. The choice of setting the age threshold at 75 for elderly lung cancer patients served several important purposes. First, it adheres to the widely accepted medical and demographic definition of the elderly population, as individuals aged 75 and above often experience heightened age-related physiological changes and comorbid conditions, substantially influencing their health and surgical outcomes. Second, employing a specific age threshold, such as 75, ensured a more uniform study group, enhancing the similarity among patients in terms of age-related factors and health profiles. Consequently, this approach bolstered the internal validity of the study by minimising potential confounding variables. The decision to pursue surgery was made collaboratively by a multidisciplinary tumour board for all cases. Patients with small cell lung cancer and those who had undergone exploratory surgery were excluded from the study. Ethical considerations were paramount; thus, the study was conducted in strict adherence to the Declaration of Helsinki, gaining approval from the P. Hertsen Moscow Oncology Research Institute Ethics Committee. Additionally, participating patients provided written consent, and all clinical data were anonymized. The study received approval from the Institutional Review Board (#034-B, 10.02.2022). Exclusion criteria encompassed patients with incomplete data, those who underwent diagnostic surgical procedures such as mediastinoscopy, and individuals who declined participation in the study.

Preoperative assessment

Perioperative data encompassed a range of factors, including age, clinical and pathological stage according to the 8th TNM classification, preoperative forced expiratory volume in the first second (FEV1, %), Charlson's Comorbidity Index (CCI) score, date of operation, histology details, radicality of resection, resection margins, growth pattern, extent of surgery, duration, blood loss, type of incision, and bronchial resection. Additionally, postoperative complications were meticulously documented. Patients were categorised into two groups based on the presence or absence of complications. An initial univariate analysis was followed by a multivariate model to elucidate the risk factors associated with adverse outcomes and long-term survival.

Preoperative assessments consisted of an electrocardiogram (ECG), echocardiogram, spirometry test, and venous Doppler ultrasound. Patients with emphysema and respiratory volume depletion underwent assessments of diffusing capacity of the lung for carbon monoxide and VO2max. Individuals with a reduced ejection fraction received further evaluation, involving a Holter ECG monitor and an exercise stress test.

For staging purposes, 18-FDG PET/CT scans were conducted in conjunction with brain MRI. All patients were staged in accordance with the TNM 8th edition. Central tumours with an endobronchial component necessitated endoscopic biopsy, while peripherally located lesions were suspected of lung cancer based on dynamic tumour enlargement and increased PET/CT uptake. In such cases, wedge resection or segmentectomy with a frozen section was performed to rule out benign lesions or tuberculosis. Decisions regarding the extension of parenchymal resection and the performance of mediastinal lymphadenectomy were contingent upon pathological findings. Indications for thoracotomy included central tumours involving the main or lobar bronchi, mediastinal lymphadenopathy, or an inability to tolerate single-lung ventilation. The selection of candidates for surgery followed international guidelines, with preference for surgery in case of atelectasis and tumour disintegration.

Following anatomical pulmonary resection, patients were admitted to the intensive care unit (ICU) for 24 h. This observation period was reduced to 2 h after wedge resection or segmentectomy. Removal of the pleural drain occurred once the outflow volume diminished to 300 ml within 24 h, provided that the fluid exhibited no chylous or blood characteristics, and air leakage had ceased. All patients underwent rehabilitation following the fast-track protocol employed in our department. This protocol encompassed preoperative patient education, thoracic epidural anaesthesia, early extubation, and the encouragement of early mobilisation. In cases of persistent fluid accumulation, a series of repeat thoracenteses were performed. The decision for discharge was made when the patient was fully rehabilitated and no longer required analgesics or antimicrobial treatment.

Postoperative complications were assessed based on the Clavien‒Dindo classification [ 8 ], and the preoperative Charlson Comorbidity Index (CCI) score was documented for all patients [ 9 ]. To examine any potential bias in the advantages of minimally invasive surgery, a comparison was made between the preoperative and histological characteristics of patients who underwent thoracotomy and those who underwent video-assisted thoracoscopic surgery (VATS).

Power analysis

Before initiating the study, a power analysis was conducted to determine the required sample size for the chi-square test. The primary objective was to assess the association between two categorical variables, namely the presence of a postoperative complication and a 30-day lethal outcome. The following parameters were considered in the power analysis. The effect size was estimated based on pilot data. A moderate effect size was anticipated. A significance level (α) of 0.05 was chosen, indicating a 5% probability of Type I error. A power of 0.80 was selected, representing an 80% probability of detecting an association if it exists. The power analysis indicated that a minimum sample size of 32 participants was required to achieve the desired level of statistical power. This sample size was used as the basis for participant recruitment and data collection.

Statistical analysis

The primary data acquisition, calculation of assessed parameters, systematisation, and data aggregation were executed utilising our proprietary open-source medical information system developed in C#, employing ASP. Net Core, and backed by PostgreSQL.

For quantitative variables adhering to a normal distribution, the mean and standard deviation were reported alongside a 95% confidence interval. In instances where variables deviated from a normal distribution, descriptive statistics included the median with interquartile range. Categorical data are presented as absolute and relative frequencies. Comparisons of frequencies in multifield contingency tables were conducted via Pearson's chi-square test. When assessing three or more groups concerning a quantitative variable without a normal distribution, the Kruskal‒Wallis test was employed, followed by Dunn’s criterion with Holm correction as a post hoc method. In cases involving three or more groups with a quantitative variable exhibiting a normal distribution, a one-way analysis of variance was conducted, accompanied by the Tukey test as a post hoc method, assuming equal variances. Survival time was calculated from the date of surgery to the last follow-up date using the Kaplan‒Meier method. Variations in survival were analysed through log-rank analysis. Missing data were handled based on the missing at random assumptions. A Cox proportional hazard regression model was applied to perform a multivariate analysis of prognostic factors, while a logistic regression model was employed to identify risk factors for postoperative complications. After multivariate regression, we constructed the Receiver Operating Characteristic (ROC) model to analyze the performance of our predictive model. This involved plotting the true positive rate against the false positive rate and calculating the area under the curve (AUC) to assess the model's discriminative ability. In addition to plotting the ROC curve, we identified the optimal threshold that maximizes both sensitivity and specificity. In addition to plotting the ROC curve, we identified the sensitivity and specificity for predicting the lethal outcome in NSCLC patients aged 75 and above. Surgical mortality encompassed all patients who passed away within the initial 30 days post surgery or during the same hospital stay. Statistical analyses were executed using IBM SPSS Statistics v.26 for Windows (IBM Corporation), StatTech v. 2.8.3 (Developer—StatTech LLC, Russia), and GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA).

The sample size for this study was established after the exclusion of 13 elderly patients who had undergone invasive staging procedures such as mediastinoscopy and parasternal mediastinotomy to assess lymph node involvement, as well as 9 patients diagnosed with small cell lung cancer. These exclusion criteria were implemented to enhance the homogeneity of the patient group under analysis by removing individuals with a history of prior surgical procedures or neoadjuvant treatment. The primary reason for nonparticipation was associated with incomplete data (31.6%). Among the 139 elderly patients who underwent surgery in our department for suspected non-small cell lung cancer (NSCLC), a total of 73 patients were included in the analysis for this study (Fig.  1 ).

STROBE flow chart. n, number of patients; NSCLC, non-small cell lung cancer; SCLC, small cell lung cancer

Following the exclusion of patients meeting the criteria, the age of the study participants ranged from 75 to 84 years (median = 78, interquartile range: 77–79), with 45.2% being female and 54.8% male. The decision to proceed with surgery was made collectively by an extended tumor board, including an oncologist, a thoracic surgeon, a radiotherapist, an anesthesiologist, an intensive care physician, and a cardiologist. Out of the participants, 10 (13%) reported complications related to vascular disease, comprising 6 (8%) who had experienced myocardial infarctions and 4 (5%) with a history of cerebrovascular accidents. The Charlson Comorbidity Index (CCI) scores ranged from 5 to 9. Further details regarding baseline patient characteristics are presented in Table  1 .

In our patient cohort, 40 (54%) of the patients exhibited a decline in FEV1 below 90%, a typical age-related occurrence [ 10 ]. The mean FEV1 was 88 ± 16%, with a 95% CI of [84, 92]. Notably, the lowest rate (65%) was observed in a long-term smoker with central adenosquamous carcinoma associated with emphysema, while the highest rate (152%) was observed in a patient with peripheral adenocarcinoma devoid of pulmonary comorbidities.

Stage I lung adenocarcinoma was diagnosed in 37 (50.6%) of the patients, with lymph node metastases identified in 25 (32.9%). Among the 9 (12.9%) patients with pN2, preoperative detection of enlarged or hypermetabolic lymph nodes was observed in only 4 (44%) patients, indicating a 7% rate of occult N2 disease. Tumour size ranged from 1 to 14 cm (median = 3.6 cm, interquartile range: 2.1–4.5). In three patients, mediastinal lymphadenectomy was not conducted due to significant comorbidities.

The mean duration of the surgical procedures was 184 ± 58 min, with a 95% CI of [170, 198]. The longest operation, lasting 330 min, was noted during VATS S1 + 2 segmentectomy for adenocarcinoma of the left upper lobe. Blood loss varied from 50 to 750 ml (median = 150 ml, interquartile range: 2.1–4.5). The median duration of chest drainage was 3 days (interquartile range: 2–4). The most prolonged air leakage was observed after a right upper lobectomy with bronchial sleeve resection, followed by a diagnosis of pneumonia in the operated lung that responded to antibiotic treatment. The mean length of hospital stay was 12 ± 2 days, with a 95% CI of [9, 14], and was significantly longer in patients with complications (median = 14.5 days; U = 170, p  = 0.003) but shorter after VATS (median = 9.5 days; U = 228, p  < 0.001).

Adjacent organ and structure resection was performed in 15 (20.5%) patients, including sleeve lobectomy in 6 (8.3%), vagus nerve resection in 2 (3%), pericardial resection in 2 (3%), wedge lung resection in 2 (3%), chest wall resection in 2 (3%), and pulmonary artery resection in 1 (1.5%) patient.

Out of the 73 patients, a total of 12 (16.4%) experienced complications, with 4 of them succumbing during their hospital stay, resulting in a 30- and 90-day mortality rate of 5.5%. Notably, the majority of these complications occurred in patients who had undergone thoracotomy, comprising 9 of the 12 cases (75%). The distribution of complications was as follows: 3 (25%) in 2015, 1 (8.3%) in 2016, 1 (8.3%) in 2017, 4 (33.3%) in 2018, 1 (8.3%) in 2019, and 2 (16.6%) in 2021. According to the Clavien‒Dindo classification, grade II complications were diagnosed in 5 (6.8%) patients, and grade IIIa complications were diagnosed in 2 (2.7%) patients. The most prevalent complication was atrial fibrillation, observed in 5 (7%) patients, one of whom succumbed to a massive pulmonary embolism. Persistent air leakage, defined as lasting for more than 5 days, was detected in 4 (5.4%) patients, and in one case, it led to severe pneumonia and death on postoperative day 32. Pneumonia was diagnosed in 3 (4.1%) patients, accounting for 25% of all complications. One patient with hypocoagulation underwent emergency thoracotomy due to hemothorax several hours after the initial operation but eventually succumbed to erosive pulmonary haemorrhage on postoperative day 21. Table 2 summarises the univariate analysis of the risk factors for complications.

Patients subjected to thoracotomy exhibited a significantly higher incidence of complications than their counterparts who underwent VATS, as demonstrated by a chi-square test (χ2(1) = 7.21, p  = 0.007) (Fig.  2 ).

Surgical access, lymph node involvement and complication rate. df, degree of freedom; p, p value; pN, pathologic N; VATS, video-assisted thoracic surgery; X 2 , chi-square

However, it is worth noting that the VATS group exhibited a lower stage ( p  = 0.024), less mediastinal lymph node involvement ( p  = 0.048), fewer combined resections ( p  = 0.001), a higher frequency of lobectomy ( p  = 0.042), and a greater prevalence of adenocarcinoma ( p  < 0.001). Furthermore, the median tumour volume in the thoracotomy group was significantly higher at 3.6 cm than in the VATS group at 2.3 cm (U = 301, p  < 0.001).

Patients with stage IIIb had a significantly higher complication rate than those with localised stages (χ2(1) = 7.42, p  = 0.006). Patients with central tumour location were more prone to complications than those with peripheral growth (OR 8.0, 95% CI [1.844, 34.712], p  = 0.001). Figure  2 also illustrates the increase in complications related to mediastinal lymph node involvement and the type of surgical access.

Quantitative characteristics such as age, duration of surgery, blood loss, preoperative FEV1, tumour size, and CCI score did not show a statistically significant difference between complicated and uncomplicated patients ( p  > 0.005).

The multivariate analysis revealed the most significant predictors of a complicated postoperative period (Fig.  3 ).

Multivariate analysis of postoperative complications. CI, confidence interval; pN, pathologic N

The area under the receiver operating characteristic (ROC) curve was 0.930 ± 0.052, 95% CI [0.827, 1], indicating a high capacity of the model to correctly classify positive and negative cases. The model's statistical significance ( p  < 0.001) underscores its strong performance. With a sensitivity of 83.3% and specificity of 82.8%, the model demonstrates a balanced ability to accurately identify true positive and true negative cases, highlighting its clinical utility.

The mean follow-up duration was 25 ± 28 months, 95% CI [20.02, 76.08], ranging from 1 to 82 months. Local relapse occurred in 8 (11%) of the followed patients. At 1 year of follow-up, 88% of patients remained relapse-free (95% CI [0.75, 1]), and at 3 years, this rate was 66% (95% CI [0.42, 0.9]). During the follow-up period, 16 (22%) patients passed away, with a median survival of 44 months (95% CI [25, 69]). The 1-year survival rate was 71% (95% CI [0.54, 0.87]), the 3-year survival rate was 66% (95% CI [0.49, 0.84]), and the 5-year survival rate was 35% (95% CI [0.13, 0.57]). Notably, overall survival after VATS was significantly higher than that after thoracotomy (54 vs. 25 months, p  = 0.021; Fig.  4 ).

Overall survival stratified by the surgical access method used. Me, overall survival; VATS, video-assisted thoracic surgery

Multivariable proportional hazards regression analysis revealed that patients with stage I disease and a lower Charlson Comorbidity Index (CCI) experienced better outcomes, while undergoing thoracotomy was the sole significant predictor of poorer outcomes (Table  3 ).

The area under the receiver operating characteristic (ROC) curve was 0.861 ± 0.068, 95% CI [0.729, 0.994], signifying a reasonably good ability of the model to classify positive and negative cases. The model exhibited statistical significance ( p  < 0.001). With a sensitivity of 75% and specificity of 81.2%, the method displayed a balanced capacity to identify true positive and true negative cases, indicating its clinical utility.

Surgery for lung cancer in elderly patients can be safely conducted at a high-volume cancer centre with substantial experience in both minimally invasive and open surgical techniques. However, it is important to acknowledge that the mortality rate remains notably high in this patient population. A crucial aspect of this process is the meticulous preoperative assessment conducted by a multidisciplinary tumour board, which includes the involvement of anesthesiologists and critical care physicians. The choice between video-assisted thoracoscopic surgery (VATS) and open surgery (thoracotomy) often hinges on patient selection. It is essential to comprehensively evaluate the extent of comorbidities since the long-term outcome for elderly patients with localised lung cancer is strongly influenced by the initial performance level, which can be affected by concurrent medical conditions.

The World Health Organization defines individuals aged 75 or older as elderly [ 11 ]. Despite the increasing popularity of stereotactic body radiation therapy (SBRT), surgical intervention continues to be the primary approach for the treatment of stage I-II lung cancer, primarily due to its favourable outcomes concerning local disease control and overall survival [ 12 ]. Nonetheless, it is crucial to recognize that surgery in this age group is associated with elevated morbidity and mortality rates [ 13 ]. Several studies have reported varying morbidity rates, ranging from 20 to 63%, and mortality rates spanning from 0 to 16% (Table  4 ).

In our study, the postoperative morbidity rate was observed to be 16.4%, a figure that appears slightly lower than the rates reported in previously published series. Several factors may contribute to this disparity. A pivotal element of the fast-track recovery protocol involves the early encouragement of physical activity, which is facilitated by the absence of a chest tube. The prompt removal of the chest tube has been associated with decreased pain levels, a reduced risk of pneumonia, and a lower likelihood of pleural cavity contamination [ 19 ]. However, in certain cases where refractory postoperative hydrothorax persists, necessitating repeat thoracentesis, this intervention is typically well tolerated and curative. In our study, we did not categorise these cases as complications, which might have led to a somewhat lower overall morbidity rate compared to other studies. Another contributing factor is the reduced pain levels and infrequent use of opioid analgesics, which likely play a role in reducing the incidence of pneumonia, responsible for only 4% of complications in our study. The investigation also underscored an increased morbidity associated with thoracotomy compared to VATS, emphasising the substantial impact of the surgical approach on patient outcomes. These findings highlight the importance of further exploration and potential refinement of surgical methods, particularly when dealing with elderly patients, to minimise morbidity and optimise patient safety.

It is well known that the overall mortality rate after lung cancer surgery varies between 1.6% and 3.7% [ 20 , 21 , 22 , 23 ]. However, this rate significantly escalates when operating on older patients, as evident in our study, where it reached 5.5%, and in a study by Naunheim et al., reporting a rate of 16%. The notable increase in mortality among older patients remains a key limiting factor in the decision-making process for surgery, particularly when considering extended resections and dealing with patients with poor functional performance. In our study, four out of twelve patients who experienced complications did not survive, resulting in a mortality rate of 33% among those who encountered postoperative complications.

Atrial fibrillation emerges as one of the most prevalent adverse events during the postoperative period. The incidence of atrial fibrillation is typically low at 0.37–1% after general surgery but can surge to 10–20% after lung parenchyma resection [ 24 , 25 ]. The rates increase even further following anatomical lung resections, reaching 34.2% after pneumonectomy and 4.5% after lobectomy [ 26 ]. Typically, atrial fibrillation is diagnosed on postoperative days 2–3 and has the potential to progress to systemic hypotension, heart failure, and even pulmonary embolism due to the formation of thrombi in atriae. In our study, atrial fibrillation was detected in 42% of complicated cases and in 7% of all enrolled patients.

Regrettably, one patient in our study died of a subsequent massive pulmonary embolism. When dealing with elderly patients, it becomes imperative to conduct a thorough preoperative assessment to rule out any possible cardiac issues and initiate early anticoagulant prophylaxis in high-risk individuals. In cases where significant neurological impairment or cognitive problems arise due to a prior cerebrovascular accident, a comprehensive discussion with the patient's family becomes a necessary step. Additionally, alternative treatment options, such as stereotactic body radiation therapy (SBRT), should be considered, and a judicious decision should be made in such circumstances.

Various tools are available to assess surgical risk, including the NSQIP, SURPASS, Lee, and ASA scoring systems. One of the commonly used tools for evaluating the cumulative severity of comorbidities is the Charlson Comorbidity Index [ 27 ]. This index calculates adjusted risk scores based on age and 19 of the most common and significant comorbidities in patients undergoing treatment. It is frequently employed in surgical and chemotherapy planning for a broader oncology population [ 28 ]. In our study, the CCI score ranged from a minimum of 5 points, attributed to patients over 70 years old with only the presence of oncology, to a maximum of 10 points, observed in an 83-year-old patient with a history of penile and prostate cancer, emphysema, postinfarction cardiosclerosis, type 2 diabetes, asthma, and an abdominal aortic aneurysm. Given the high surgical risk and the subpleural location in this particular case, we opted to maximise lung parenchyma preservation and expedite the surgery by performing a VATS wedge resection, despite the presence of lung adenocarcinoma. Remarkably, the patient survived for four years without recurrence but eventually succumbed to COVID-19. While univariate analysis revealed that CCI did not predict the complication rate, it was effective in estimating the overall survival of patients in the study. However, patients with a higher number of preoperative complications may undergo a reduced scope of surgical resection, ultimately reducing surgical trauma and the incidence of postoperative complications.

One of the most accurate methods for assessing the functional capacity of a patient scheduled for pulmonary resection is the measurement of VO2max. A higher VO2max generally correlates with better cardiopulmonary fitness and can serve as a valuable prognostic indicator for surgical outcomes [ 29 ]. It aids in patient selection, risk stratification, and the development of personalised preoperative optimization strategies, ensuring that individuals with lower VO2max receive appropriate interventions to enhance their functional capacity and mitigate potential postoperative complications. However, VO2max was not incorporated into the current study due to the absence of this facility in the clinic.

The milestone work of Ginsberg and Rubinstein [ 30 ] laid the foundation for our understanding of lung cancer resections. Their seminal study demonstrated that lobectomy represents the least radical type of lung resection for lung cancer, and patients who underwent sublobar resection exhibited significantly lower overall and recurrence-free survival rates. While this classic paper, published in 1995, has been widely cited over 640 times, significant advancements in medical sciences, including the advent of PET/CT, new TNM classifications, and VATS, have ushered in substantial changes in the diagnosis, treatment modalities, and comprehension of lung cancer.

Recent studies have challenged the notion that segmentectomy represents a compromise operation, suggesting that it may even be beneficial for lesions less than 2 cm, both in terms of complication rates and long-term survival [ 31 , 32 ]. Notably, the results of the CALGB 140503 study, which randomised 701 patients with stage IA lung cancer to lobar and sublobar resections, demonstrated that after a 7-year follow-up, sublobar and lobar resections yielded similar disease-free and overall survival rates. The overall survival rates after 5 years were 80.3% for sublobar resection and 78.9% for lobar resection [ 33 ].

In our study, segmentectomy was performed in 19.2% of cases, emerging as the second most common procedure following lobectomy (68.5%). Remarkably, the complication rates did not exhibit significant differences between these two groups. Wedge resection has largely faded from the spectrum of treatment options, as it is no longer deemed radical enough and is reserved for severely ill patients, as previously described. Pneumonectomy is also witnessing a decline in popularity due to the poor tolerance of subsequent hemodynamic and ventilatory changes. In cases where pneumonectomy is being considered, it is essential to thoroughly evaluate the feasibility of sleeve resection, a procedure that necessitates a two-stage approach: first during preoperative planning and subsequently during intraoperative assessment. The open sleeve lobectomy technique should be integrated into the standard practice of every thoracic surgery department that manages lung cancer patients.

The introduction of VATS for anatomic lung resection in the early 1990s significantly transformed the landscape of lung cancer surgery. VATS has demonstrated safety, reduced morbidity, and equivalent efficacy when compared to traditional open surgery. This minimally invasive technique has been embraced as the standard surgical approach for early-stage lung cancer and is increasingly employed in more advanced cases, including in elderly patients [ 34 , 35 ].

In our study, the VATS group exhibited fewer complications than the open approach group (5.3% vs. 28.6%; p  = 0.007). However, it is important to consider that this discrepancy may partly result from selection bias. VATS procedures were predominantly performed on peripheral small adenocarcinomas without mediastinal lymph node involvement. In contrast, meticulous lymphadenectomy during thoracotomy in the presence of N1 and N2 disease might account for potential damage to vagal parasympathetic nerve branches, leading to a higher incidence of atrial fibrillation in this group [ 36 ]. Nonetheless, VATS demonstrated superior results to the open approach, regardless of the stage, and was more favourable for elderly patients due to enhanced postoperative recovery and reduced pain levels. The multivariate analysis underscored that thoracotomy was an independent negative predictor of overall survival.

The 8th edition of the TNM classification, introduced by the Union for International Cancer Control in 2017, brought several significant updates [ 37 ]. One of the notable changes was the reclassification of T3N2 patients as stage IIIb, rendering them nonsurgical candidates. Our study included three stage IIIb patients who underwent surgery before the implementation of the new TNM classification. In some cases, surgery may still be considered for advanced-stage patients with the intention of performing salvage procedures for conditions such as abscess, empyema, or massive pulmonary haemorrhage. However, such cases are associated with a heightened complication rate (60% in our study), larger tumour sizes (average of 8 cm), and more extensive lung resections (pneumonectomy or bilobectomy performed in 100% of these cases). Furthermore, complications in these situations carried a 100% fatality rate. Therefore, decisions to operate under such circumstances should be made with the primary goal of saving the patient's life, and the associated risks must be carefully weighed against potential benefits. Nevertheless, in our series, the number of deaths in patients with IIIb and N2 was very low and insufficient for an accurate analysis.

This study exhibits certain potential limitations that warrant consideration when interpreting the findings. First, all patients underwent surgery at a single, high-volume cancer centre renowned for its adeptness in managing patients afflicted by severe comorbidities and offering a spectrum of oncological services. Each surgical procedure was conducted by seasoned thoracic surgeons who had transcended the learning curve and had completed 40–50 VATS lobectomies annually.

Second, it is worth noting that the number of cases included in the analysis might not be adequate for a comprehensive evaluation of survival rates. This insufficiency could introduce selection bias, which may undermine the generalizability of the study's conclusions. The potential bias introduced by collinearity among key variables in the logistic regression models. Notably, variables such as tumour stage, N status, and distance to hilum exhibited a degree of interdependence, raising concerns about the robustness of our models. While we implemented rigorous model selection techniques, including cross-validation, to mitigate the risk of overfitting, the inherent collinearity remains a challenge. Estimation of overall survival in the context of elderly cancer patients is complex and is potentially influenced by comorbidities. While our primary focus was on cancer-related endpoints, the study did not delve into a detailed examination of specific comorbidities and their impact on overall survival. But the prognostic power of CCI was validated and confirmed. Further propensity score matching or a prospective randomised trial is needed.

In each case of lung cancer in patients above 75 years of age, an individualised approach is essential. The elevated mortality rate underscores the necessity for a comprehensive and meticulous assessment of risk factors. Preoperative intervention to ameliorate existing comorbidities should involve the active participation of anesthesiologists, intensive care physicians, cardiologists, and other relevant specialists, as dictated by the specific case requirements.

The convenience of a thorough discussion within the extended tumour board is of paramount importance in ensuring optimal patient care. Especially in the presence of negative prognostic factors such as stage IIIb, lymph node metastases, central tumour location, squamous cell histology, and the requirement for thoracotomy, the absolute risk of postoperative complications experiences a substantial increase. This underscores the criticality of precision and thorough planning in the management of elderly lung cancer patients.

Clinical practice points

Increased mortality and morbidity rates in elderly patients with lung cancer require a careful and detailed assessment of risk factors and preoperative compensation for any existing comorbidity with the participation of an anesthesiologist, intensive care physician and cardiologist and other related specialists if necessary. In the presence of the identified risk factors, such as stage IIIb, lymph node metastases, central tumour location, squamous cell histology, and the necessity for thoracotomy, shared decision-making should be undertaken. This involves providing a thorough explanation of the associated risks, benefits, and alternative treatment modalities, including options such as SBRT or chemoimmunotherapy.

Availability of data and materials

The data underlying this article are available in the Dryad Digital Repository at https://doi.org/10.5061/dryad.kh1893289.

He W, Goodkind D, Kowal PR. An aging world: 2015. Washington, DC: U.S. Government Publishing Office; 2016.

Google Scholar  

Siegel RL, Miller KD, Jemal A. Cancer Statistics, 2017. CA Cancer J Clin. 2017;67(1):7–30. https://doi.org/10.3322/caac.21387 .

Article   PubMed   Google Scholar  

Hutchins LF, Unger JM, Crowley JJ, Coltman CA Jr, Albain KS. Underrepresentation of patients 65 years of age or older in cancer-treatment trials. N Engl J Med. 1999;341(27):2061–7. https://doi.org/10.1056/NEJM199912303412706 .

Article   CAS   PubMed   Google Scholar  

Okami J, Higashiyama M, Asamura H, et al. Pulmonary resection in patients aged 80 years or over with clinical stage I non-small cell lung cancer: prognostic factors for overall survival and risk factors for postoperative complications. J Thorac Oncol. 2009;4(10):1247–53. https://doi.org/10.1097/JTO.0b013e3181ae285d .

Berry MF, Hanna J, Tong BC, et al. Risk factors for morbidity after lobectomy for lung cancer in elderly patients. Ann Thorac Surg. 2009;88(4):1093–9. https://doi.org/10.1016/j.athoracsur.2009.06.012 .

Husain ZA, Kim AW, Yu JB, Decker RH, Corso CD. Defining the High-Risk Population for Mortality After Resection of Early Stage NSCLC. Clin Lung Cancer. 2015;16(6):e183–7. https://doi.org/10.1016/j.cllc.2015.04.007 .

Balduyck B, Hendriks J, Lauwers P, Sardari Nia P, Van Schil P. Quality of life evolution after lung cancer surgery in septuagenarians: a prospective study. Eur J Cardiothorac Surg. 2009;35(6):1070–5. https://doi.org/10.1016/j.ejcts.2009.01.050 .

Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg. 2009;250(2):187–96. https://doi.org/10.1097/SLA.0b013e3181b13ca2 .

Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373–83. https://doi.org/10.1016/0021-9681(87)90171-8 .

Thomas ET, Guppy M, Straus SE, Bell KJL, Glasziou P. Rate of normal lung function decline in ageing adults: a systematic review of prospective cohort studies. BMJ Open. 2019;9(6):e028150. Published 2019 Jun 27. https://doi.org/10.1136/bmjopen-2018-028150 .

Dyussenbayev A. Age Periods Of Human Life. Advances in Social Sciences Research Journal. 2017;4(6): 258–63. https://doi.org/10.14738/assrj.46.2924 .

Ettinger DS, Wood DE, Aisner DL, et al. NCCN Guidelines® Insights: Non-Small Cell Lung Cancer, Version 2.2023. J Natl Compr Canc Netw. 2023;21(4):340–50. https://doi.org/10.6004/jnccn.2023.0020 .

Takayuki N, Keiko T, Junji U, et al. Advanced Non-Small Cell Lung Cancer in Elderly Patients: Patient Features and Therapeutic Management. Biomed Res Int. 2018;2018:8202971. Published 2018 Apr 30. https://doi.org/10.1155/2018/8202971 .

Naunheim KS, Kesler KA, D’Orazio SA, Fiore AC, Judd DR. Lung cancer surgery in the octogenarian. Eur J Cardiothorac Surg. 1994;8(9):453–6. https://doi.org/10.1016/1010-7940(94)90013-2 .

Port JL, Mirza FM, Lee PC, Paul S, Stiles BM, Altorki NK. Lobectomy in octogenarians with non-small cell lung cancer: ramifications of increasing life expectancy and the benefits of minimally invasive surgery. Ann Thorac Surg. 2011;92(6):1951–7. https://doi.org/10.1016/j.athoracsur.2011.06.082 .

Mun M, Kohno T. Video-assisted thoracic surgery for clinical stage I lung cancer in octogenarians. Ann Thorac Surg. 2008;85(2):406–11. https://doi.org/10.1016/j.athoracsur.2007.10.057 .

Hope WW, Bolton WD, Kalbaugh CA, Blackhurst DW, Stephenson JE, Taylor SM. Lung cancer resection in octogenarians: a reasonable approach for our aging population. Am Surg. 2007;73(1):22–4. https://doi.org/10.1177/000313480707300106 .

Okada A, Hirono T, Watanabe T. Safety and prognosis of limited surgery for octogenarians with non-small cell lung cancer. Gen Thorac Cardiovasc Surg. 2012;60(2):97–103. https://doi.org/10.1007/s11748-011-0880-3 .

Nomori H, Horio H, Suemasu K. Early removal of chest drainage tubes and oxygen support after a lobectomy for lung cancer facilitates earlier recovery of the 6-minute walking distance. Surg Today. 2001;31(5):395–9. https://doi.org/10.1007/s005950170128 .

Knott-Craig CJ, Howell CE, Parsons BD, Paulsen SM, Brown BR, Elkins RC. Improved results in the management of surgical candidates with lung cancer. Ann Thorac Surg. 1997;63(5):1405–10. https://doi.org/10.1016/s0003-4975(97)00252-x .

Ginsberg RJ, Hill LD, Eagan RT, et al. Modern thirty-day operative mortality for surgical resections in lung cancer. J Thorac Cardiovasc Surg. 1983;86(5):654–8.

Damhuis RA, Schütte PR. Resection rates and postoperative mortality in 7,899 patients with lung cancer. Eur Respir J. 1996;9(1):7–10. https://doi.org/10.1183/09031936.96.09010007 .

Myrdal G, Gustafsson G, Lambe M, Hörte LG, Ståhle E. Outcome after lung cancer surgery. Factors predicting early mortality and major morbidity. Eur J Cardiothorac Surg. 2001;20(4):694–699. https://doi.org/10.1016/s1010-7940(01)00875-2 .

Hassanzadeh DM, Kamandi M, Feiz DH, Dezyani J, Keyvani V. Investigating the effect of inflammation on atrial fibrillation occurrence by measuring highly sensitive C-reactive protein (hs-CRP). J Cardiothorac Med. 2014;2:167–71. https://doi.org/10.22038/JCTM.2014.2613 .

Passman RS, Gingold DS, Amar D, et al. Prediction rule for atrial fibrillation after major noncardiac thoracic surgery. Ann Thorac Surg. 2005;79(5):1698–703. https://doi.org/10.1016/j.athoracsur.2004.10.058 .

Bagheri R, Yousefi Y, Rezai R, Azemonfar V, Keshtan FG. Atrial fibrillation after lung surgery: incidence, underlying factors, and predictors. Kardiochir Torakochirurgia Pol. 2019;16(2):53–6. https://doi.org/10.5114/kitp.2019.86355 .

Article   PubMed   PubMed Central   Google Scholar  

Zhao L, Leung LH, Wang J, et al. Association between Charlson comorbidity index score and outcome in patients with stage IIIB-IV non-small cell lung cancer. BMC Pulm Med. 2017;17(1):112. Published 2017 Aug 15. https://doi.org/10.1186/s12890-017-0452-0 .

Singh N, Singh PS, Aggarwal AN, Behera D. Comorbidity Assessment Using Charlson Comorbidity Index and Simplified Comorbidity Score and Its Association With Clinical Outcomes During First-Line Chemotherapy for Lung Cancer. Clin Lung Cancer. 2016;17(3):205–13.e1. https://doi.org/10.1016/j.cllc.2015.10.002 .

Salati M, Brunelli A. Risk Stratification in Lung Resection. Curr Surg Rep. 2016;4(11):37. https://doi.org/10.1007/s40137-016-0158-x .

Ginsberg RJ, Rubinstein LV. Randomized trial of lobectomy versus limited resection for T1 N0 non-small cell lung cancer. Lung Cancer Study Group. Ann Thorac Surg. 1995;60(3):615–23. https://doi.org/10.1016/0003-4975(95)00537-u .

Cao C, Gupta S, Chandrakumar D, Tian DH, Black D, Yan TD. Meta-analysis of intentional sublobar resections versus lobectomy for early stage non-small cell lung cancer. Ann Cardiothorac Surg. 2014;3(2):134–41. https://doi.org/10.3978/j.issn.2225-319X.2014.03.08 .

Lim TY, Park S, Kang CH. A Meta-Analysis Comparing Lobectomy versus Segmentectomy in Stage I Non-Small Cell Lung Cancer. Korean J Thorac Cardiovasc Surg. 2019;52(4):195–204. https://doi.org/10.5090/kjtcs.2019.52.4.195 .

Altorki N, Wang X, Kozono D, et al. Lobar or Sublobar Resection for Peripheral Stage IA Non-Small Cell Lung Cancer. N Engl J Med. 2023;388(6):489–98. https://doi.org/10.1056/NEJMoa2212083 .

Roviaro G, Varoli F, Vergani C, Nucca O, Maciocco M, Grignani F. Long-term survival after videothoracoscopic lobectomy for stage I lung cancer. Chest. 2004;126(3):725–32. https://doi.org/10.1378/chest.126.3.725 .

Dziedzic R, Marjanski T, Binczyk F, Polanska J, Sawicka W, Rzyman W. Favourable outcomes in patients with early-stage non-small cell lung cancer operated on by video-assisted thoracoscopic surgery: a propensity score-matched analysis. Eur J Cardiothorac Surg. 2018;54(3):547–53. https://doi.org/10.1093/ejcts/ezy101 .

Park HW, Shen MJ, Lin SF, Fishbein MC, Chen LS, Chen PS. Neural mechanisms of atrial fibrillation. Curr Opin Cardiol. 2012;27(1):24–8. https://doi.org/10.1097/HCO.0b013e32834dc4e8 . PMID:22139702;PMCID:PMC3279730.

Lim W, Ridge CA, Nicholson AG, Mirsadraee S. The 8th lung cancer TNM classification and clinical staging system: review of the changes and clinical implications. Quant Imaging Med Surg. 2018;8(7):709–18. https://doi.org/10.21037/qims.2018.08.02 .

Download references

Acknowledgements

Not applicable.

No funding.

Author information

Authors and affiliations.

Federal State Budget Institution National Medical Research Radiology Centre of the Ministry of Healthcare of the Russian Federation (FSBI NMRRC), Moscow, Russia

Andrey Kaprin

Peoples’ Friendship University of Russia, Moscow, Russia

P. Hertsen Moscow Oncology Research Institute, 3, 2 Botkinskiy Proezd, Moscow, 125284, Russia

Oleg Pikin, Andrey Ryabov, Oleg Aleksandrov, Denis Larionov & Airat Garifullin

National Medical Research Center of Phthisiopulmonology, Moscow, Russia

Oleg Aleksandrov

You can also search for this author in PubMed   Google Scholar

Contributions

Andrey Kaprin—Conceptualization, Project administration.  Andrey Ryabov—Conceptualization, Project administration, Writing – review & editing.  Oleg Pikin—Conceptualization, Project administration, Writing – review & editing.  Oleg Aleksandrov—Investigation, Formal Analysis, Writing – original draft.  Airat Garifullin—Writing – review & editing.  Denis Larionov—Data curation, Writing – review & editing.  All authors have read and agreed to the published version of the manuscript.

Corresponding author

Correspondence to Oleg Aleksandrov .

Ethics declarations

Ethics approval and consent to participate.

The study was conducted in strict adherence to the Declaration of Helsinki, gaining approval from the P. Hertsen Moscow Oncology Research Institute Ethics Committee. The study received approval from the Institutional Review Board (#034-B, 10.02.2022).

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Kaprin, A., Pikin, O., Ryabov, A. et al. Surgical intervention for lung cancer in patients aged 75 and above: potential associations with increased mortality rates—a single-center observational study. J Cardiothorac Surg 19 , 471 (2024). https://doi.org/10.1186/s13019-024-02922-5

Download citation

Received : 02 November 2023

Accepted : 15 June 2024

Published : 29 July 2024

DOI : https://doi.org/10.1186/s13019-024-02922-5

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Pulmonary surgical procedure
• Complications
• Perioperative period

Journal of Cardiothoracic Surgery

ISSN: 1749-8090

• General enquiries: [email protected]

• Search Menu
• Sign in through your institution
• Supplements
• Cohort Profiles
• Education Corner
• Author Guidelines
• Submission Site
• Open Access
• About the International Journal of Epidemiology
• About the International Epidemiological Association
• Editorial Team
• Editorial Board
• Advertising and Corporate Services
• Journals Career Network
• Self-Archiving Policy
• Dispatch Dates
• Contact the IEA
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

I. mortality and population data, ii. retrospective and prospective studies, iii. studies on pathogenesis, iv. other laboratory investigations, v. interpretation.

• < Previous

Smoking and lung cancer: recent evidence and a discussion of some questions *

• Article contents
• Figures & tables
• Supplementary Data

Jerome Cornfield, William Haenszel, E. Cuyler Hammond, Abraham M. Lilienfeld, Michael B. Shimkin, Ernst L. Wynder, Smoking and lung cancer: recent evidence and a discussion of some questions, International Journal of Epidemiology , Volume 38, Issue 5, October 2009, Pages 1175–1191, https://doi.org/10.1093/ije/dyp289

• Permissions Icon Permissions

Summary This report reviews some of the more recent epidemiologic and experimental findings on the relationship of tobacco smoking to lung cancer, and discusses some criticisms directed against the conclusion that tobacco smoking, especially cigarettes, has a causal role in the increase in broncho-genic carcinoma. The magnitude of the excess lung-cancer risk among cigarette smokers is so great that the results can not be interpreted as arising from an indirect association of cigarette smoking with some other agent or characteristic, since this hypothetical agent would have to be at least as strongly associated with lung cancer as cigarette use; no such agent has been found or suggested. The consistency of all the epidemiologic and experimental evidence also supports the conclusion of a causal relationship with cigarette smoking, while there are serious inconsistencies in reconciling the evidence with other hypotheses which have been advanced. Unquestionably there are areas where more research is necessary, and, of course, no single cause accounts for all lung cancer. The information already available, however, is sufficient for planning and activating public health measures. – J. Nat. Cancer Inst . 22: 173–203, 1959.

“The sum total of scientific evidence establishes beyond reasonable doubt that cigarette smoking is a causative factor in the rapidly increasing incidence of human epidermoid carcinoma of the lung.”

The consideration of the accumulated scientific evidence has led to the acceptance of a similar viewpoint by responsible public health officials in Great Britain, the Netherlands, Norway, and the United States. This consensus of scientific and public health opinion does not mean that all problems, regarding smoking and lung cancer have now been solved or that valid questions and reservations about some aspects of the subject do not remain. An excellent collection of primary references and opinions expressing both “sides” of the question was issued by a committee of the House of Representatives 3 which sought to examine the claims of filter-tip cigarette advertisements.

The general acceptance of the cigarette-lung-cancer relationship has not decreased research interest but has accelerated research in this and in such related fields as respiratory physiology and environmental carcinogens, and on the effect of tobacco smoke in a wide range of physiological and pathological reactions.

The result is that considerably more information has been published or has become available through other media. Included in the recent scientific evidence are the following:

Additional retrospective studies 4 , 5 , 6 on men with lung cancer and on matched controls have appeared. All show an association between cigarette smoking and epidermoid-undifferentiated lung cancer.

Additional retrospective studies on women 7 , 6 also show the association.

The first results of a third large prospective study 8 , which included 200,000 United States veterans who were observed for 30 months, duplicate closely the reported findings of the Hammond-Horn 9 and the Doll-Hill 10 studies.

Analyses by Kreyberg and others 11 , 12 substantiate that, epidemiologically, primary lung cancer must be divided into epidermoid-undifferentiated and adenocarcinoma. The latter is much less related to smoking and, so far as is know at present, to other carcinogenic inhalants.

Additional findings have become available on the impingement of tobacco-smoke particles in the bronchi of animals, ciliary paralysis, and penetration of unidentified fluorescent materials into the bronchial cells. 13 , 14 , 15

Additional data have been published 16 , 17 on the more frequent occurrence of hyperplastic and metaplastic changes in the lungs of smokers as compared with the lungs of nonsmokers. Hyperplastic and metaplastic changes have been produced in bronchi of dogs exposed to direct contact with tobacco “tars” 18 and in bronchi of mice exposed to tobacco smoke. 19

Additional confirmations have been obtained on the induction of cancer of the skin in mice painted with tobacco-smoke condensates. 20 , 21 , 22 , 23 , 24

Progress continues on the isolation and identification of chemical constituents in tobacco smoke, including compounds of the carcinogenic polycyclic type. 23 , 25 , 26 , 27 , 28

The growing and consistent body of evidence has had no noticeable effect upon the viewpoint of a small but important group of individuals who would deny a causal role of cigarette smoking in cancer of the lung. Among these critics are Little 29 and Hartnett 30 , spokesmen for the American tobacco industry. Berkson 31 , 32 has been critical of many aspects of the statistical studies, and his reservations are, in part, also evident in papers by Neyman 33 and Arkin 34 . More general objections by Fisher 35 , 36 , Greene 37 , Hueper 38 , Macdonald 39 . Rigdon 40 , and Rosenblatt 41 have been published.

We have reviewed the criticism that have been made regarding the cigarette-lung cancer relationship in the light of new evidence. In this review we have several objectives: a) to point out recorded facts that directly answer some of the criticisms; b) to define more precisely some inadequacies of information, with the hope that this will lead to further research. The particular references we have used were selected because in our opinion the criticism was well stated; it is not our intention to reply to any specific publication or to any specific critic. Our view is that all valid questions should be answered. However, some questions may not be relevant, or there may be no information presently available for an answer. In the latter case, we believe that a distinction should be made between data that are unavailable and data that have been found to be contradictory.

For convenience, we have divided the criticisms and answers into five major topics, as follows: (I) Mortality and population data; (II) Retrospective and prospective studies; (III) Studies on pathogenesis; (IV) Other laboratory investigations; and (V) Interpretation.

The rising death rate from lung cancer in all countries that have sufficiently detailed mortality statistics is the most striking neoplastic phenomenon of this century. That this increase is a fact and not a spurious result of statistical classification is now commonly accepted An entirely contrary view is held by only a few persons 40 , though there are dissenting opinions 42 , 38 regarding the extent and time relationship of this recorded increase.

Obviously, the case for the etiologic role of cigarette smoking would be seriously compromised if it could be demonstrated that the lung-cancer rate over the past half century had been stationary, particularly after 1920 when much of the rise in cigarette consumption, instead of other forms of tobacco, occurred 43 .

In a recent review, Rigdon and Kirchoff 44 document that primary lung cancer was first recognized as an entity during the early part of the 19 th century, and that its occurrence has increased steadily since then, as manifested by the recorded relative frequency with which it was recognized in the clinic and at necropsy. This is undoubtedly correct but does not constitute evidence against a true increase in the incidence of the disease during the whole, or a more recent part, of the last 100 years.

Hueper 38 , accepting a true increase in the incidence of lung cancer, regards an increase dating back to 1900, or before the widespread use of cigarettes, as evidence against the cigarette-lung-cancer relationship. His contention would have crucial import only if it were maintained that cigarette smoking is the sole cause of lung cancer.

The vital statistics and the necropsy data that support the presumption of a real increase in lung-cancer risk certainly apply to the years after 1920. Because of the uncertainties associated with changes in diagnostic accuracy, no firm conclusions can be reached on whether the rate of increase in lung-cancer mortality has, in truth, accelerated since 1920.

Effect of Aging

Rosenblatt 41 has raised the question about the effect of the aging population on the lung-cancer rate. This particular point has been investigated by the use of age-adjusted rates. Dunn 45 has noted that only one sixth of the over-all increase in lung-cancer mortality among males in the United States (from 4 to 24 deaths per 100,000 males between 1930 and 1951) could be attributed to an aging population. Similar findings 46 have been presented for England and Wales where observations on lung-cancer mortality date back to 1900; the 1953 mortality rate for both sexes, 34 per 100,000 population, was 43 times the corresponding 1900 rate, 0.8 per 100,000 population. Allowance for increased average age of the population could account for only half this rise in lung-cancer mortality, with a 24-fold difference between 1900 and 1953.

Also, an aging population, does not affect the age-specific death rates and cannot account for the phenomenon of increasingly higher lung-cancer mortality at all ages throughout the lifespan, which has occurred among successively younger groups of males born in the United States and England and Wales since 1850. A similar but less pronounced “cohort displacement” has been shown for females.

Diagnostic Factors

Little 29 and others 40 have raised the important question on whether better diagnostic measures and more complete reporting have resulted in a spurious increase in the recorded attack rate. Several special features of the increase in lung-cancer mortality would be difficult to account for on diagnostic grounds. These include the continuous rising ratio of male to female deaths, the increasing lung-cancer mortality rate among successively younger cohorts, and the magnitude of the current, continuing, increase in lung-cancer mortality 46 . By 1955, among white males, 50 to 64 years of age, in the United States, more deaths were attributed to lung cancer than to all other respiratory diseases combined.

Gilliam 42 has made a careful study of the potential effect of improved diagnosis on the course of the lung-cancer death rate. Even assuming that 2 percent of the deaths certified in past years as tuberculosis or other respiratory disease were really due to lung cancer, he concluded that “… all of the increase in mortality attributed to cancer of the lung since 1914 in United States white males and females cannot be accounted for by erroneous death certification to other respiratory diseases without unreasonable assumptions of age and sex differences in diagnostic error.” His computations reduced the respective 26-fold and sevenfold increase in lung-cancer mortality among males and females, between 1914 and 1950, to the more modestly estimated dimensions of fourfold and 30 percent, respectively. These estimates are certainly the lower bound on the magnitude of the true rate of increase during this period.

The Copenhagen Tuberculosis Station data, examined by Clemmesen et al . 47 , provide the greatest measure of control on the diagnostic improvement factor. In a tuberculosis referral service, used extensively by local physicians, where diagnostic standards and procedures including systematic bronchoscopy remained virtually unchanged between 1941 and 1950, the lung-cancer prevalence rate among male examinees increased at a rate comparable to that recorded by the Danish cancer registry for the total male population. This can be regarded as evidence that the reported increase in Danish incidence is not due to diagnostic changes.

Necropsy Data

Most necropsy data agree with mortality data on the increase in lung-cancer risk. To establish this point we referred to a necropsy series summarized by Steiner 48 , and returned to the original sources for evidence on the nature of changes over time. Since an existing compilation was chosen, the results do not represent a culling of autopsy series for data favorable to this thesis. The findings from 13 series are summarized in text- figure 1 as the proportion of lung cancers in relation to all necropsies. The relative frequency in terms of total tumors or total carcinomas yielded results which would lead to substantially the same inferences.

Mortality and necropsy data have their own virtues and weaknesses. Death certificates provide a complete report of deaths, but do not emphasize a high quality of diagnostic evidence, while the reverse holds true for necropsies. However, since both approaches lead to the same inferences, neither great variation in the quality of diagnostic evidence nor the unrepresentative nature of some of the necropsy observations can be viewed as plausible interpretations of the results. The alternative conclusion of a real increase in lung-cancer risk remains.

Urban-Rural Differences

Emphasis has been placed on the alleged incompatibility of the excess lung-cancer mortality, among urban residents, with the cigarette-smoking hypothesis 38 , 49 . Mortality data from several countries indicate strongly that lung-cancer rates are much higher in cities than in rural areas, and the observation that urban males in general have higher lung-cancer mortality than rural males is undoubtedly correct.

The assertion of Macdonald 39 that “ … country people smoke as much, if not more, than do city people …” is not borne out by the facts 50 . Nevertheless, the evidence indicates that adjustment for smoking history could account for only a fraction of this urban-rural difference 51 .

However, this does not establish the converse proposition that control of residence history in the analysis of collected data would account for the excess lung-cancer risk among cigarette smokers. Evidence now in hand weighs strongly against this last assertion. Stocks and Campbell 67 , in their report on lung-cancer mortality among persons in Liverpool, the suburban environs, and rural towns in North Wales, showed that heavy smokers have higher lung-cancer rates when urban and rural males were studied separately. Mills and Porter 52 reported similar findings in Ohio. These results agree with the experience of the Hammond-Horn 9 study, which revealed markedly higher death rates for bronchogenic carcinoma among smokers regardless of whether they lived in cities or in rural areas. No contradictory observations are known to us.

Sex Differences

The sex disparity in lung-cancer mortality has also been cited 35 , 7 as grounds for discarding the cigarette-smoking hypothesis. In this connection it should be noted that persons advocating this line of argument have minimized sex differences in smoking habits to a degree not supported by available facts. A survey of smoking habits in a cross section of the United States population 50 demonstrated that men, on the average, have been smoking for longer periods than women. The sex differences in tobacco use were especially pronounced at ages over 55, when most lung-cancer deaths occur; 0.6 percent of United States females in this age group have been reported as current users of more than 1 pack of cigarettes daily compared to 6.9 percent of United States males. British data 53 also revealed much lower tobacco consumption among females, particularly for the years before World War II.

The present data contrasting the experience by sex would appear to support the cigarette-smoking hypothesis rather than discredit it. When differences in smoking habits are considered, it is possible to reduce the observed fivefold excess lung-cancer mortality among males to the 40 percent excess mortality which prevails for many other causes of death 51 . One intriguing finding from these studies is that the estimated death rates for female nonsmokers agree closely with the death rates derived from retrospective studies on male nonsmokers 7 .

Evidence for Other Etiological Factors

Etiologic factors of industrial origin, such as exposure to chromates and coal gas, are well established 46 . Excess lung-cancer risks among such groups as asbestos workers who develop asbestosis, appear likely 46 . One epidemiologic study 54 of British, World War I, veterans exposed to mustard gas and/or with a wartime history of influenza revealed virtually no excess lung-cancer risk among these groups.

The existence of other important lung-cancer effects associated with such characteristics as socioeconomic class cannot be questioned. Cohart 55 found that the poorest economic class had a 40 percent higher lung-cancer incidence than the remaining population of New Haven, Connecticut. Results from the 10-city morbidity survey 56 have revealed a sharp gradient in lung-cancer incidence, by income class, for white males, which is consistent with Cohart's findings. Since cigarette smoking is not inversely related to socioeconomic status, we can agree with Cohart “… that important environmental factors other than cigarette smoking exist that contribute to the causation of lung cancer.” These and other findings are convincing evidence for multiple causes of lung cancer. It is obviously untenable to regard smoking of tobacco as the sole cause of lung cancer.

Two points should be made: The population exposed to established industrial carcinogens is small, and these agents cannot account for the increasing lung-cancer risk in the remainder of the population. Also, the effects associated with socioeconomic class and related characteristics are smaller than those noted for smoking history, and the smoking-class differences cannot be accounted for in terms of these other effects.

Special population Groups

Haag and Hanmer 57 reported that employees in 9 processing plants of the American Tobacco Company, with an above-average proportion of smokers, had a lower mortality than the general population of Virginia and North Carolina for all causes and for cancer and cardiovascular diseases, but no higher mortality for respiratory cancer and coronary disease. They concluded: “The existence of such a population makes it evident that cigarette smoking per se is not necessarily or invariably associated with a higher risk of lung cancer or cardiovascular diseases or with diminished longevity.”

The group studied by Haag and Hanmer was too small to yield significant results on respiratory cancer. Moreover, a major flaw in the conclusion has been pointed out by Case 58 . It is well known that mortality comparisons cannot be drawn directly between employee groups and the general population, since the death rates for many groups of employed persons are lower than death rates for the general population with age, sex, and race taken into consideration. This is true because there is a strong tendency to exclude from employment those persons who have acute or chronic diseases or who are seriously disabled from any cause and those employees who develop permanent disabilities from disease or other causes are usually discharged, retired, or dropped from the list of regular employees. Reasons of this nature undoubtedly account for the deficit in deaths from all causes noted in the group of employees under consideration.

A different picture is provided by the Society of Actuaries 59 who made a study for 1946 through 1954. The death claims for employees of the tobacco industry were reported to be slightly higher than, and the permanent disability claims were reported to be over three times as high as, those for employees in nonrated industries as a whole. This latter comparison indicates that the basic assumption of the Haag and Hanmer study is incorrect. Also, interpretation of group comparisons in this field should account separately for the experience of smokers and nonsmokers. We hope that Haag and Hanmer will supplement the report to provide data for smokers and nonsmokers in the study population.

The association between smoking and lung cancer has now been investigated and reported by at least 21 independent groups of investigators in 8 different countries, who employed what is known as the retrospective method 1 , 4 , 5 , 6 , 7 , 46 . In these studies, patients with lung cancer, or their relatives, were questioned about their smoking history and other past events, and the answers compared with those of individuals without lung cancer who were selected as controls. Although these 21 studies have certain features in common, they varied greatly in the methods of selecting the groups, the methods of interview, and other important aspects.

The association between smoking and lung cancer was further investigated in two countries by three independent groups 8 , 9 , 10 , using the prospective method. In these studies, large groups were questioned on smoking habits and other characteristics, and the groups were observed for several years for data on mortality and causes of death. The three prospective studies also varied in several important details including the type of subjects, the selection of subjects, and the method of obtaining information on smoking habits.

In each of these studies, an association was found between smoking and lung cancer. In every investigation where the type of smoking and lung cancer. In every investigation where the type of smoking was considered, a higher degree of association was found between lung cancer and cigarette smoking than between lung cancer and pipe or cigar smoking. In every instance where amount of smoking was considered, it was found that the degree of association with lung cancer increased as the amount of smoking increased. When ex-cigarette smokers were compared with current cigarette smokers, it was found that lung-cancer death rates were higher among current cigarette smokers than among ex-cigarette smokers.

A number of investigators 60 have criticized the retrospective method but, for the most part, the specific points of criticism apply only to some of the studies and not to others. Some features of the three prospective studies on smoking also have been criticized. Again, certain of the points of criticism apply to one or another of the three prospective studies but not to all three. Specifically, doubts raised as to the validity of the early findings of the prospective studies have been eliminated by the persistence of the findings in the later phases of the same studies.

The validity of the findings on these extensive investigations has been questioned in regard to two major aspects: 1) the methods of selection of the study groups, and 2) the accuracy of information regarding smoking habits and the diagnosis of lung cancer.

Selection of Study Groups

Neyman 33 pointed out that a study based on a survey of a population at some given instant of time may yield misleading results. Suppose that a study is made on a day when all patients with lung cancer and a group of people without lung cancer are questioned about their smoking habits. If smokers with lung cancer live longer than nonsmokers with lung cancer, there would be a higher proportion of smokers in the lung-cancer group than in the control group – this would follow without questioning the proposition on which the model is based. However, only two of the retrospective studies were conducted in a way approximating an “instantaneous survey” procedure, so that this criticism does not apply to most of the studies. Furthermore, this difficulty is completely avoided in prospective studies.

Berkson 31 indicated that people with two specific complaints are more likely to be hospitalized than people with only one of these complaints. If a retrospective study were conducted exclusively on hospital patients an association would be found between these two specific complaints, even if there were no association between the same two complaints in the general population. This would influence the results if smokers with lung cancer are more likely to be hospitalized than nonsmokers with lung cancer. However, Berkson showed that this difficulty is trivial if a high percentage of people with either one of these two conditions is hospitalized, which is the situation with lung-cancer patients. Furthermore, one retrospective study 67 included all lung cancer patients who were in the study area, including those not hospitalized; another retrospective study 61 was based on individuals who died of lung cancer and other diseases regardless of whether they had been hospitalized or not. This difficulty does not arise in prospective studies.

In all but one of the 21 retrospective studies, the procedure was to compare the smoking habits of lung-cancer patients with the smoking habits of a control group who did not have lung cancer. Hammond 60 , Berkson 31 , and others have pointed out the grave danger of bias if the control group is not selected in such a way as to represent (in respect to smoking habits) the general population which includes the lung-cancer patients. Subsequent events have proved that this criticism is well founded, though the direction of the bias in most studies turned out to yield an underestimate of the degree of association between cigarette smoking and lung cancer. The reason was that in most of the retrospective studies the control group consisted of patients with diseases other than lung cancer. The choice of such a control group is tantamount to assuming that there is no association between smoking and diseases which resulted in hospitalization of the control subjects. This was an incorrect assumption since other studies have indicated an association between smoking and a number of diseases, such as coronary artery disease, thromboangiitis obliterans, and cancer of the buccal cavity.

Doll and Hill 62 , recognizing the possibility of bias in a control group selected from hospital patients, obtained an additional control group by ascertaining the smoking habits of the general population in a random sample of the area in which their hospital was located. The largest percentage of smokers (particularly heavy smokers) was found in the lung-cancer group, the smallest percentage of smokers was found in the general population sample, and an intermediate percentage of smokers was found in the hospital-control group. Similar results have been reported in a recent study of women 7 .

Berkson 31 pointed out that the criticisms in regard to selection bias in the retrospective studies are also applicable to the earlier findings in a prospective study. Suppose that, in selecting subjects for a prospective study, sick smokers are overrepresented in relation to well smokers and/or well non-smokers are overrepresented in relation to sick nonsmokers. In this event, during the earlier period after selection, the death rate of the smokers in the study would be higher than the death rate of the nonsmokers in the study, even if death rates were unrelated to smoking habits of the general population. If smoking is unrelated to death from lung cancer (or other causes), the death rate of the smokers would tend to equalize with that of the nonsmokers as the study progressed. Thus, the bias would diminish with time, and a relationship due to such bias would disappear. This general principle is well known to actuaries and is one of the cornerstones of the life insurance business.

Hammond and Horn 9 , recognizing this possible difficulty, excluded from the study all persons who were obviously ill at the time of selection. As expected, the total death rate of the study population was low and very few deaths from lung cancer occurred during the first 8 months after selection. The total death rate, and particularly the death rate from lung cancer, rose considerably in the subsequent 3 years. What is more important, the observed association between cigarette smoking and lung cancer was considerably higher in the latter part than in the early part of the study, and the association between cigarette smoking and total death rates was also somewhat greater in the latter part of the study. This showed that the original bias in the selection of the subjects was slight and that it yielded an underestimate of the degree of association between smoking and death rates.

This particular problem was not encountered in the prospective studies of Doll and Hill 10 who could observe the death rates of all physicians in Great Britain (nonresponders as well as responders to the smoking questionnaire). The prospective study of Dorn 8 also had a defined population of veterans holding insurance policies, and nonresponders were observed as well as responders. Moreover, these two studies also showed that higher mortality from lung cancer among smokers was more evident during the later period than in the earlier period of observation. Thus, in the course of time, there was no disappearance of any selection bias factors that may have been introduced into the original study groups.

The subjects for the Hammond and Horn prospective study 9 were selected by volunteer workers with specific instructions on how it should be done. Mainland and Herrera 63 have suggested that the volunteer workers may have introduced a bias in the way they selected the subjects. The foregoing evidence of persistence and accentuation of the differences between smokers and nonsmokers, in time, effectively counters purposeful, as well as unknown, sources of such selection.

Accuracy of Information

Berkson 31 , 32 has remarked that the two major variables considered in all these studies – the ascertainment of smoking habits and the diagnosis of disease – are both subject to considerable error. The accuracy of diagnosis is not a major problem in retrospective studies because the investigator can restrict his study to those patients whose diagnosis of lung cancer has been thoroughly confirmed. This feature has been taken into consideration in several retrospective studies. It is more of a problem in prospective studies since all deaths that occur must be included, and certainly some of the diagnoses will be uncertain. However, in all three prospective studies, the total death rate was found to be higher in cigarette smokers than in nonsmokers and found to increase with the amount of cigarette smoking. If some of the excess deaths associated with cigarette smoking and ascribed to lung cancer were actually due to some other disease, then it means that: a) the association between cigarette smoking and lung cancer was somewhat overestimated, but b) the association between smoking and some other disease was somewhat underestimated. The reverse would be true if some of the excess deaths associated with cigarette smoking and ascribed to diseases other than lung cancer were actually due to lung cancer. Hammond and Horn 9 found that the association with cigarette smoking was greater for patients with a well-established diagnosis of lung cancer than for patients with less convincing evidence for a diagnosis of lung cancer. This suggests that inaccuracies in diagnosis resulted somewhat in an underestimate of the degree of association between smoking and lung cancer.

The study on physicians, by Doll and Hill 10 , in which presumably the clinical and pathologic evidence of the cause of death would be somewhat more than in the general population considered by Hammond and Horn and by Dorn, yields almost identical risks to lung cancer by smoking class.

In regard to information about smoking, Finkner et al . 64 have made a thorough study of the accuracy of replies to questionnaires on smoking habits. Their results indicate that replies are not completely accurate but that most of the errors are relatively minor – very few heavy smokers are classified as light smokers. Random and independent errors simply tend to diminish the apparent degree of association between two variables. A national survey of smoking habits in the United States 50 yielded results on tobacco consumption that were consistent with figures on tobacco production and taxation.

On two occasions several years apart, Hammond and Horn 9 and Dorn 8 questioned a proportion of their subjects. The results indicated close reproducibility in the answers.

Hammond 60 and others 39 have questioned the reliability of the retrospective method on the grounds that the illness may bias the responses given by the patient or his family when they are questioned about smoking habits, and that knowledge of the diagnosis may bias the interviewer. This possible difficulty was minimized in several of the 21 retrospective studies on smoking in relation to lung cancer. For example, in the study conducted by Levin 65 , all patients admitted to a hospital during the course of several years were questioned about their smoking habits before a diagnosis was made. Only a small proportion later turned out to have lung cancer, though many had lung disease symptoms or lung diseases other than lung cancer. Doll and Hill 10 also showed that patients whose diagnosis of lung cancer was subsequently established to be erroneous had smoking histories characteristic of the control rather than of the lung-cancer group. Furthermore, a larger percentage of cigarette smokers have been found among patients with epidermoid carcinoma of the lungs than among patients with adenocarcinoma of the lungs 66 . This could hardly have resulted from bias either on the part of the patient or on the part of the interviewer.

Multiple Variables

Arkin 34 , Little 29 , Macdonald 39 , and others have criticized the studies of cigarette-lung cancer relationship on the grounds that only smoking habits were really investigated, and that numerous other possible variables were not considered.

This criticism may seem especially appropriate in view of the accepted fact that no single etiologic factor has been proposed for any neoplastic disease. The criticism may also be valid in relation to any one of the retrospective and prospective studies. However, in the aggregate, quite a number of other variables have been specifically investigated or can be inferentially derived. Of course, all studies considered the basic factors of age and sex; some dealt with geographic distribution 67 , occupation 68 , urban or rural residence 67 , marital and parous status 7 , and some other habits such as coffee consumption 7 .

The Doll and Hill 10 prospective study was confined to a single professional group, physicians. Thus there could be no great variation attributable to occupation or socioeconomic status. Stocks and Campbell 67 put particular emphasis on the study of air pollution and occupational exposure and included a number of other factors in addition to smoking. It is evident, in the Hammond-Horn 9 study and other investigations, that there is a consistent relationship between urban residence and a higher mortality due to lung cancer. The important fact is that in all studies, when other variables are held constant, cigarette smoking retains its high association with lung cancer.

The only factors that may show a higher correlation with lung cancer than heavy cigarette smoking are such occupations as those of the Schneeberg miners and manufacturers of chromate 46 . We are not acquainted with actual studies of these and related occupation groups in which cigarette and other tobacco consumption is also considered. Such studies, we suggest, would be useful additions to our knowledge of other etiologic agents of the interplay between multiple causes in human pulmonary cancer.

Inhalation of Smoke

If cigarette smoking produces cancer of the lungs as a result of direct contact between tobacco smoke and the bronchial mucosa, smokers who inhale cigarette smoke should be exposed to higher concentrations of the carcinogens than noninhalers and therefore have a higher risk to the development of lung cancer. The retrospective study of Doll and Hill 62 , however, elicited no difference between patients with lung cancer and the controls in the proportion of smokers who stated that they inhaled. Fisher 35 , Hueper 38 , and Macdonald 39 have emphasized this point as contradictory to the smoking-lung-cancer relationship, and, of course, it is. Unfortunately, this particular finding was not reinvestigated in the prospective study of Doll and Hill 10 .

Three authors, Lickint 69 , Breslow et al . 68 , and Schwartz and Denoix 4 , however, did find the relative risk of lung cancer to be greater among inhalers than among noninhalers when age, type, and amount of smoking were held constant. It must be admitted that there is no clear explanation of the contradiction posed by the Doll-Hill 62 findings, though a number of plausible hypotheses could be advanced. More experimental work is required, including some objective definition and measurement of the depth and length of inhalation.

Hammond 70 has recently queried male smokers about their inhalation practices. He found that very few pipe and cigar smokers inhale; that most men inhale who smoke only cigarettes; and that there are proportionally fewer inhalers among men who smoke both cigars and cigarettes than among men who smoke only cigarettes. These findings are compatible with the view that differences in inhaling account for the fact that the lung-cancer death rate of cigar and pipe smokers is less than the lung-cancer death rate of cigarette smokers; and that the lung-cancer death rate of men who smoke both cigars and cigarettes is somewhat lower than the lung-cancer death rate of men who smoke only cigarettes.

Upper-Respiratory Cancer

Rosenblatt 41 has drawn attention to the fact that increased consumption of cigarettes has not been accompanied by an increase in upper-respiratory cancer similar to that noted in cancer of the lung and bronchus. Hueper 38 also has expressed doubts about the causative role of cigarette smoking on the basis that cigarette smoking is not associated with cancer of the oral cavity or of the fingers, which are often stained with tobacco tar.

The premise that a carcinogen should act equally on different tissues is not supported by experimental or clinical evidence 71 . Carcinogens, which produce liver tumors in animals, may be noncarcinogenic when applied to the skin. Coal soot, accepted as etiologically related to carcinoma of the scrotum in chimney sweeps, does not increase the risk to cancer of the penis. There is no a priori reason why a carcinogen that produces bronchogenic cancer in man should also produce neoplastic changes in the nasopharynx or in other sites. It is an intriguing fact, deserving further research, that carcinoma of the trachea is a rarity, whereas carcinoma of the bronchus is common among individuals exposed to chromates, as well as among chronic cigarette smokers.

Several studies have established the association of all types of tobacco smoking, including cigarettes, with cancer of the oral cavity 72 . However, the relative risk of developing cancer of the mouth is greater for cigar and pipe smokers than for cigarette smokers. The risk of laryngeal cancer is increased by smoking and an equal risk exists among cigarette, cigar, and pipe smokers 73 . The per capita consumption of cigars and pipe tobacco has decreased since 1920, while cigarette smoking has increased 43 .

These associations contrast sharply with the findings on lung cancer, which have consistently shown that cigarette smokers have much higher risks than either cigar or pipe smokers. Since 1920 the increase in tobacco consumption has been primarily due to the rise in cigarette consumption 43 , and the stabler rates for intra-oral and laryngeal cancer, while the lung-cancer rates have increased steeply, can be considered compatible with the causal role of cigarette smoking in lung cancer.

Effect of Tobacco Smoke on Bronchial Mucosa

Statements by Hartnett 30 , Macdonald 39 , and others 31 , 29 imply that the relationship of cigarette smoking and lung cancer is based exclusively on “statistics” and lacks “experimental” evidence. The differentiation between various methods of scientific inquiry escapes us as being a valid basis for the acceptance or the rejection of facts. Nevertheless it is true that historically the retrospective studies on lung cancer preceded the intensive interest in laboratory investigations stimulated by the statistical findings.

Hilding 13 has shown experimentally that exposure to cigarette smoke inhibited ciliary action in the isolated bronchial epithelium of cows. Kotin and Falk 15 obtained essentially the same results in experiments on rats and rabbits. Hilding 14 further showed that inhibition of ciliary action interfered with the mechanism whereby foreign material is ordinarily removed from the surface of bronchial epithelium. In addition, he found that foreign material deposited on the surface tended to accumulate in any area where the cilia have been destroyed. Auerbach et al . 16 found that the small areas of the bronchial epithelium where ciliated columnar cells were absent appeared more frequently in smokers than in nonsmokers. Chang 17 found that cilia were shorter, on an average, in the bronchial epithelium of smokers than in that of nonsmokers.

These studies have demonstrated the existence of a mechanism whereby foreign material from any source (e.g. tobacco smoke, industrial dusts, fumes from automobile exhausts, general air pollutants, and, perhaps, pathogenic organisms) is likely to remain in contact with the bronchial epithelium for a longer period in smokers than in nonsmokers.

Auerbach and his associates 16 studied the microscopic appearance of the bronchial epithelium of patients who died of lung cancer and patients who died of other diseases. Each of these two groups of patients was classified according to whether they were nonsmokers, light smokers, or heavy cigarette smokers. Among the cancer patients there were no nonsmokers. Approximately 208 sections from all parts of the tracheobronchial tree from each patient were examined. Many areas of basal cell hyperplasia, squamous metaplasia, and marked atypism, with loss of columnar epithelium were found in the tracheo-bronchial tree of men who had died of lung cancer. Almost as many such lesions were found in heavy cigarette smokers who had died of other diseases; somewhat less were found in light cigarette smokers; and much less in nonsmokers. Chang 17 has reported similar findings in the bronchial epithelium of smokers compared with nonsmokers.

The chief criticism of Auerbach's study has concerned terminology. Following the definition previously set forth by Black and Ackerman 74 , Auerbach et al . used the term “carcinoma- in-situ ” to describe certain lesions with marked atypical changes and loss of columnar epithelium. Whether this is an appropriate term may be questioned, but it is not relevant to the validity of the findings. Certainly there are no data to indicate what proportion of these morphologically abnormal areas would progress to invasive carcinoma.

The recent findings of Auerbach et al . and Chang have been reproduced experimentally in animals. Rockey and his associates 75 applied tobacco ”tar” directly to the bronchial mucosa of dogs. Within 3 to 6 weeks, the tar-treated surface became granular and later developed wart like elevations. Upon microscopic examination, hyperplasia, transitional metaplasia, and squamous metaplasia were found in these areas. Leuchtenberger et al . 19 exposed mice to cigarette smoke for periods up to 200 days. The bronchial epithelium was then examined microscopically. Bronchitis, basal-cell hyperplasia, and atypical basal-cell hyperplasia were found in the majority of the animals and squamous metaplasia in a few. Further work and longer periods of observation are necessary to establish whether some of these lesions would progress to frank neoplasia.

Skin Cancer in Rodents

One of the links in the total evidence for the causal relationship of cigarette smoking and lung cancer is the demonstration that tobacco smoke condensates (usually referred to as “tars”) have the biologic property of evoking carcinoma in certain laboratory animals, particularly mice. The production of skin cancer in mice, following repeated, long-term applications of tobacco tar, has now been reported from at least six different laboratories 20 , 21 , 22 , 23 , 24 , 76 . It is undeniable that some investigators did not obtain positive results, perhaps because the dose and other experimental conditions were different, or because the complex tobacco tars probably varied widely in their composition. The negative results of Passey et al . 18 have been quoted by Hueper 38 and others, but a more recent experiment by Passey 24 with Swiss strain mice did lead to the appearance of at least two carcinomas after repeated applications of tobacco-smoke condensate.

Little 29 indicated that “… the extrapolation to the human lung of results obtained by painting of or injection into the skin of mice is decidedly questionable”. Direct extrapolation from one species to another is, of course, not justified. Nevertheless, results in animals are fully consistent with the epidemiologic findings in man. A quotation from Kotin 49 is appropriate: “The chemical demonstration of carcinogenic agents in the environment and their successful use for the production of tumours in experimental animals do not prove or even especially strongly suggest a like relationship in the instance of man. When, however, a demonstrable parallelism exists between epidemiologic data and laboratory findings, greater significance accrues to both. Medical history is replete with examples in which laboratory findings have been proved ultimately to have their counterpart in the human experience. Exceptions have been very few.”

Greene 37 , while discounting the significance of the induction of skin carcinoma in Swiss mice because of the constitutionally “high differential susceptibility” of the strain, believes that the failure to induce neoplasms in embryonic transplants exposed to tobacco tar is more important evidence. Greene's interesting technique does produce positive results when pure chemicals such as benzo[α]pyrene are used, and this chemical has been recovered from some samples of tobacco-smoke condensate. We are not acquainted with reports of neoplasms arising in embryonic tissue that has been exposed in vitro to coal tar, another crude mixture that contains carcinogens.

The high frequency of carcinoma induction reported by Wynder et al . 76 has not been achieved by other investigators, who reported that no more than 20 percent of animals, and usually considerably less, developed carcinoma of the skin. The presence of cocarcinogenic materials in tobacco-smoke condensates has been demonstrated by Gellhorn 22 and by Bock and Moore 20 . To the mouse data are now added the data on the induction of skin cancer in some rabbits painted with tobacco-smoke condensate 77 ; this condensate, when combined with a killed suspension of tubercle bacilli, and introduced into a bronchus, produced a carcinoma of the bronchus in one rat 78 .

Since malignant neoplasms have been obtained in several strains of mice, and a few neoplasms have been produced in rabbits and rats, the issue of strain or species limitation to the reaction is more difficult to maintain. It is, of course, a fact that many agents shown to be carcinogenic to the skin of mice have not been proved carcinogenic to man. In most instances there is simply no experience with such agents in man, so that lack of proof really represents lack of data, pro and con.

The Problem of Dosage

Little 29 has further questioned the applicability of animal data to man, as follows: “Tobacco smoke or smoke condensate has failed to produce cancer even on the skin of susceptible strains of mice when applied in the quantity and at an exposure rate that would simulate conditions of human smoking.”

The differences in species, tissues, and conditions between the induction of neoplasms on the skin of mice and in the bronchi of man, preclude fine comparisons of dose and time relationships.

Bronchogenic Cancer in Animals

The pulmonary adenomatous tumor in mice, rats, and guinea pigs cannot be compared with the bronchogenic carcinoma in man 71 . Until a few years ago, the experimental induction of epidermoid carcinoma had been achieved only in a few mice by passing strings impregnated with carcinogenic hydrocarbons through the lung. Epidermoid carcinoma of the lung was consistently produced in rats by beryllium 79 , by carcinogenic hydrocarbons introduced as fixed pellets into bronchi of rats 80 , and by inhalation of radioactive particles 81 .

Little 29 has noted that “… prolonged exposure of the lungs of rodents to massive doses of cigarette smoke has failed to produce bronchogenic cancer.” This remains true at the time of this report, although it can be questioned whether any animal receives as large a dose of cigarette smoke through indirect exposure as a human being does by voluntary deep inhalation. Therefore the failure may be a technical one, which may be solved by further experimentation. The early results of Leuchtenberger et al . 19 suggested that this may be achieved.

Carcinogens in Tobacco Smoke

The isolation and identification of specific chemical constituents in tobacco smoke, which are carcinogenic for the pulmonary tissue of man, is an important area for research.

It has been clear for some time that combustion or pyrolysis of most organic material, including tobacco, will form higher aromatic polycyclics of established carcinogenic activity 28 . A number of higher aromatic polycyclics have been identified and isolated ( 23 , 25 , 26 , 27 ). These materials include benzo[ e ]pyrene, benzo[ a ]pyrene, dibenz[ a,h ]anthracene, chrysene, and, most recently, a newly established carcinogen, 3,4-benz-fluoranthene. Whether these compounds are equally involved in human pulmonary carcinogenesis is, of course, conjectural.

Little 29 has implied that a specific constituent must be found to account for the biologic activity of tobacco smoke. This is not necessary. The situation is similar to the establishment of the carcinogenic activity of tar, which was accepted before the isolation of benzo[ a ]pyrene by Kennaway and his coworkers. In this instance, also, benzo[ a ]pyrene is most probably not the only carcinogen in the complex mixture called tar, and there are strong indications that some noncarcinogenic components in tar may have cocarcinogenic effects.

Three interpretations of the observed association of lung cancer and cigarette smoking are possible: 1) that cigarette smoking “causes” lung cancer, either (a) through the direct carcinogenic action of smoke on human bronchial epithelium or (b) by a more indirect mode of action such as making the individual susceptible to some other specific carcinogenic agent in the environment; 2) that lung cancer “causes” cigarette smoking, perhaps because a precancerous condition sets up a process which leads to a craving for tobacco; 3) that cigarette smoking and lung cancer both have a common cause, usually specified as a special constitutional make-up, perhaps genetic in origin, which predisposes certain individuals to lung cancer and also makes them cigarette smokers.

The second hypothesis was advanced by Fisher 36 , apparently for the sake of logical completeness, and it is not clear whether it is intended to be regarded as a serious possibility. Since we know of no evidence to support the view that the bronchogenic carcinoma diagnosed after age 50 began before age 18, the median age at which cigarette smokers begin smoking, we shall not discuss it further.

The Constitutional Hypothesis

The first hypothesis may be referred to as the causal hypothesis and the third as the constitutional hypothesis. Nothing short of a series of independently conducted, controlled, experiments on human subjects, continued for 30 to 60 years, could provide a clear-cut and unequivocal choice between them. We nevertheless argue that evidence, in addition to that associating an increased mortality from lung cancer with cigarette smoking, is entirely consistent with the causal hypothesis but inconsistent, in many respects, with the constitutional hypothesis, so that even in the absence of controlled experimentation on human beings the weight of the evidence is for the one and against the other.

The difficulties with the constitutional hypothesis include the following considerations: (a) changes in lung-cancer mortality over the last half century; (b) the carcinogenicity of tobacco tars for experimental animals; (c) the existence of a large effect from pipe and cigar tobacco on cancer of the buccal cavity and larynx but not on cancer of the lung; (d) the reduced lung-cancer mortality among discontinued cigarette smokers. No one of these considerations is perhaps sufficient by itself to counter the constitutional hypothesis ad hoc modification of which can accommodate each additional piece of evidence. A point is reached, however, when a continuously modified hypothesis becomes difficult to entertain seriously.

Changes in Mortality

Mortality from lung cancer has increased continuously in the last 50 years, and considerably more for males than females. Such an increase can be explained either as the result of an environmental change (to which males are more exposed or more sensitive than females, if both are equally exposed) or as the result of a sex-linked mutation. The constitutional hypothesis must be modified in the light of this increase, since an unchanging constitutional make-up cannot by itself explain an increase in mortality. Proponents of the constitutional hypothesis have not indicated the type of modification they would consider. Three suggest themselves to us: 1) differences in constitutional make-up are genetic in origin, but rather than predisposing one to lung cancer, they make one sensitive to some new environmental agent (other than tobacco), which does induce lung cancer; 2) differences in constitutional make-up are not genetic but are the result of differential exposure to some new environmental agent, which both predisposes to lung cancer and creates a craving for cigarette smoke; 3) the mutation has led to a greater susceptibility to lung cancer and a preference for cigarette smoke.

In the first two situations the effect of the postulated constitutional make-up would be mediated through an environmental agent. The modified hypothesis thus requires the existence of an environmental agent other than tobacco, exposure to which would be at least as highly correlated with lung-cancer mortality as exposure to cigarettes, and which also would be highly correlated with cigarette consumption. No such agent has yet been found or even suggested. In view of the magnitude of the increase in mortality from lung cancer, the third situation would require a mutation rate exceeding anything previously observed.

Experimental Carcinogenesis With Tobacco Tar

Condensed tobacco smoke contains substances that are carcinogenic for mouse and rabbit skin. It does not necessarily follow that these substances are also carcinogenic for human lungs nor does it follow that they are not. However, the constitutional hypothesis asserts they are not; and that it is simply a coincidence that these materials which are carcinogenic for experimental animals are also associated with a higher lung-cancer mortality in man.

Types of Tobacco and Cancer Site

A greatly increased lung-cancer risk is associated with increased cigarette consumption but not with increased consumption of pipe and cigar tobacco. Studies on cancer of the buccal cavity and larynx, however, have demonstrated a considerably higher risk among smokers, irrespective of the form or tobacco used. Only two ways of modifying the constitutional hypothesis to take account of this evidence occur to us: 1) There are two different constitutional make-ups, one of which predisposes to cigarettes but not to pipe and cigar consumption and to cancer of the lung, and the other predisposes to cancer of the buccal cavity and larynx but not of the lung and to tobacco consumption in any form. 2) Constitutional make-up predisposes to cigarette consumption and lung cancer only, but tobacco smoke, whether from cigarettes, cigars, or pipes, is carcinogenic for the mucosa of the buccal cavity and the larynx but not for the bronchial epithelium.

Mortality Among Discontinued Smokers

Mortality from lung cancer among discontinued cigarette smokers is less than that among those continuing to smoke 9 , 10 ; the magnitude of the reduction depending on amount previously smoked and the length of the discontinuance. The hypothetical constitutional factor which predisposes to lung cancer and cigarette smoking cannot therefore be a constant characteristic of an individual over his lifetime but must decrease in force at some time in life, thus resulting in the cessation of cigarette smoking and a concomitant, but not causally related, reduction in the lung-cancer risk. Furthermore, since cigarette smoking is rarely begun after age 35 50 , it must be inferred that the constitutional factor cannot increase in force with the passage of time, even though it may decrease.

In summary, the constitutional hypothesis does not provide a satisfactory explanation of all the evidence. It is natural, therefore, to inquire about the positive findings which support it. Even those who regard this hypothesis with favor would agree, we believe, that supporting evidence is quite scanty.

There are a number of characteristics in which cigarette smokers are known to differ from nonsmokers and presumably more will be discovered. Thus, cigarette smokers consume more alcohol, more black coffee, change jobs more often, engage more in athletics, and are more likely to have had at least one parent with hypertension or coronary artery disease 82 . Discontinued cigarette smokers are weaned at a later age than those continuing to smoke 83 . Recently, Fisher 83 reported that 51 monozygotic twins resembled each other more in their smoking habits than 33 dizygotic twins, thus suggesting a genetic determinant.

Two somewhat obvious, but necessary, comments on results of this type are in order: 1) The demonstration that a characteristic is related to smoking status does not by itself create a presumption that it is a common cause. It must also be shown to be related to the development of lung cancer among subgroups of individuals with the same smoking status. Alcohol and coffee fail to meet this test, while none of the other characteristics related to smoking status have been investigated from this point of view. 2) There is a quantitative question. Cigarette smokers have a ninefold greater risk of developing lung cancer than nonsmokers, while over-two-pack-a-day smokers have at least a 60-fold greater risk. Any characteristic proposed as a measure of the postulated cause common to both smoking status and lung-cancer risk must therefore be at least nine-fold more prevalent among cigarette smokers than among nonsmokers and at least 60-fold more prevalent among two-pack-a-day smokers. No such characteristic has yet been produced despite diligent search.

These comments on the quantitative aspects of association apply also to the relationship of certain characteristics with lung cancer. Thus, a possible genetic basis to lung cancer has been suggested to some by the association between gastric cancer and blood group. The difference, in risk of developing gastric cancer, between blood groups A and O, however, is 20 percent, while the only study of lung cancer and blood groups 84 with which we are familiar shows a difference of 27 percent (and is not quite significant at the P = 0.01 level. 1 Such differences are suggestive for further work, but cannot be considered as casting much light on differences of magnitude, ninefold to 60-fold.

Measures of Differences

The comments in the last two paragraphs have utilized a relative measure of differences in lung-cancer risk. Since Berkson 32 has argued that a relative measure is inappropriate in the investigation of smoking and mortality, we now discuss the use of relative and absolute measures of differences in risk. When an agent has an apparent effect on several diseases, the ranking of the diseases by the magnitude of the effect will depend on whether an absolute or a relative measure is used. Thus in Dorn's study 8 of American veterans there were 187 lung-cancer deaths among cigarette smokers compared with an expectation of 20 deaths, based on the rates for nonsmokers. This yields a mortality ratio of 9.35 as a relative measure and an excess of 167 deaths as an absolute measure. For cardiovascular diseases there were 1,780 deaths among cigarette smokers compared to an expectation of 1,165. This gives a relative measure of 1.53 and an absolute measure of 615 deaths. Relatively, cigarettes have much larger effect on lung cancer than on cardiovascular disease, while the reverse is true if an absolute measure is used.

Both the absolute and the relative measures serve a purpose. The relative measure is helpful in 1) appraising the possible noncausal nature of an agent having an apparent effect; 2) appraising the importance of an agent with respect to other possible agents inducing the same effect; and 3) properly reflecting the effects of disease misclassification or further refinement of classification. The absolute measure would be important in appraising the public health significance of an effect known to be causal.

If an agent, A, with no causal effect upon the risk of a disease, nevertheless, because of a positive correlation with some other causal agent, B, shows an apparent risk, r, for those exposed to A, relative to those not so exposed, then the prevalence of B, among those exposed to A, relative to the prevalence among those not so exposed, must be greater than r.

If two uncorrelated agents, A and B, each increase the risk of a disease, and if the risk of the disease in the absence of either agent is small (in a sense to be defined), then the apparent relative risk for A, r, is less than the risk for A in the absence of B.

If a causal agent A increases the risk for disease I and has no effect on the risk for disease II, then the relative risk of developing disease I, alone, is greater than the relative risk of developing disease I and II combined, while the absolute measure is unaffected.

The Causal Hypothesis

When the sexes are compared it is found that lung cancer has been increasing more rapidly in men relatively to women … But it is notorious, and conspicuous in the memory of the most of us, that over the last 50 years the increase of smoking among women has been great, and that among men (even if positive) certainly small. The theory that increasing smoking is ‘the cause’ of the change in apparent incidence of lung cancer is not even tenable in the face of this contrast.

It would thus appear that cigarette smoking is one of the causes of all ills and contributes to the over-all death rate, remembering that this rate includes such causes as accident, homicide, etc. It seems quite clear that cigarette smoking is a symptom, not a cause. It is possible – even though this is a conjecture – that they type of person who is careful of his health is less likely to be a cigarette smoker and that the cigarette smoker is likely to be the person who generally takes greater health risks.

Berkson 32 also has pointed to the multiple findings in both the Hammond-Horn and the Doll-Hill results and concluded that the observed associations may have some other explanation than a causal one. He suggests three: 1) “The observed associations are ‘spurious’ …. 2) The observed associations have a constitutional basis. Persons who are nonsmokers, or relatively light smokers, are the kind of people who are biologically self-protective, and biologically this is correlated with robustness in meeting mortal stress from disease generally. 3) Smoking increases the ‘rate of living’ (Pearl), and smokers at a given age are, biologically, at an age older than their chronologic age.”

One might ask why the finding of an association with a number of diseases, rather than just one, is necessarily contradictory and must be regarded as supporting the constitutional hypothesis. Arkin 34 supplied no answer, while the relevant statements of Berkson 32 on this point were:

When an investigation set up to test the theory, suggested by evidence previously obtained, that smoking causes lung cancer, turns out to indicate that smoking causes or provokes a whole gamut of diseases, inevitably it raises the suspicion that something is amiss. It is not logical to take such a set of results [e.g., an association of smoking with a ‘wide variety of diseases’] as confirming the theory that tobacco smoke contains carcinogenic substances which, by contact with the pulmonary tissues, initiate cancerous changes at the site of contact.

The apparent multiple effects of tobacco do raise a question with respect to the mode of action, however, and since this question is related to another alleged contradiction – the apparent lack of an inhalation effect – we shall discuss them together. What mode of action, it has been asked, can one postulate to explain these diverse effects? Two remarks are in order: 1) The evidence that tobacco is a causal agent in the development of other diseases seems weaker than the evidence for lung cancer simply because the effects are smaller. While we could not exclude the possibility that cigarettes play a causal role in, for instance, the development of arteriosclerotic-coronary heart disease, the possibility that a common third factor will be discovered, which explains a 70 percent elevation in risk from coronary heart disease among cigarette smokers, is less remote than the possibility that the ninefold risk for lung cancer will be so explained. 2) Accepting, for the sake of discussion, the causal role of cigarettes for any disease showing an elevated mortality ratio, no mater how small, the presence of other causes will be manifested in a lowered mortality ratio. Thus, even if cigarette consumption causes an elevation of 70 percent in mortality from coronary heart disease, other causes of great importance must also be present, as is manifested by the high mortality from this disease among nonsmokers. The existence of a small number of nonsmokers who develop lung cancer is a definite indication, by the same token, that cigarettes are not an absolutely necessary condition and that there are other causes of lung cancer.

If tobacco smoke does have multiple effects, each of these effects must be studied separately because of the complex nature of the agent. To postulate in advance that a single mode of action will be found to characterize them all is an unwarranted oversimplification. It is generally accepted, for example, that tobacco smoke causes thromboangiitis obliterans in susceptible humans by interfering with the peripheral circulation, and that it causes tumors when painted on the backs of susceptible mice because of the presence of carcinogenics in the tars. The a priori postulation of a single mode of action for these two effects is no substitute for detailed study of each.

As to the possible mode of action of tobacco smoke in inducing lung cancer, the evidence at this writing suggests direct action of substances in tobacco smoke on susceptible tissues with which they are in contact. Aside from background knowledge derived from experimental carcinogensis which suggests this explanation, the following evidence favors it: 1) Cigarette smoke, which is usually drawn into the lungs is associated with mortality from lung cancer, while smoke from pipes and cigars, which is usually not inhaled, if not. 2) For sites with which smoke is in direct contact, whether or not inhaled, particularly buccal cavity and larynx, the type of tobacco used makes less difference in incidence. 3) In experimental carcinogenesis, which uses tobacco tars, tumors have appeared at the site of application, and their incidence has not yet seriously dependent on the type of tobacco used. 4) The relative risk of lung cancer is higher among cigarette smokers who inhale than among those smoking the same number of cigarettes per day, but who do not inhale.

Several critics 36 , 38 , 39 have stressed the failure of Doll and Hill 62 , in their preliminary report, to find a difference in risk between inhalers and noninhalers, but this finding was contradicted in three other studies 4 , 68 , 69 . Further work on this point is desirable, but would be more convincing if a more objective measure were found of the amount of smoke to which human bronchial epithelium is exposed in the course of smoking a cigarette.

Why, it is sometimes asked, do most heavy cigarette smokers fail to develop lung cancer if cigarettes are in fact a causal agent? We have no answer to this question. But neither can we say why most of the Lübeck babies who were exposed to massive doses of virulent tubercle bacilli failed to develop tuberculosis. This is not a reason, however, for doubting the causal role of the bacilli in the development of the disease.

One cannot discuss the mode of action of tobacco without becoming aware of the necessity of vastly expanded research in the field. The idea that the subject of tobacco and mortality is a closed one requiring no further study is not one we share. As in other fields of science, new findings lead to new questions, and new experimental techniques will continue to cast further light on old ones. This does not imply that judgment must be suspended until all the evidence is in, or that there are hierarchies of evidence, only some types of which are acceptable. The doctrine that one must never assess what has already been learned until the last possible piece of evidence would be a novel one for science.

It would be desirable to have a set of findings on the subject of smoking and lung cancer so clear-cut and unequivocal that they were self-interpreting. The findings now available on tobacco, as in most other fields of science, particularly biologic science, do not meet this ideal. Nevertheless, if the findings had been made on a new agent, to which hundreds of millions of adults were not already addicted, and on one which did not support a large industry, skilled in the arts of mass persuasion, the evidence for the hazardous nature of the agent would be generally regarded as beyond dispute. In the light of all the evidence on tobacco, and after careful consideration of all the criticisms of this evidence that have been made, we find ourselves unable to agree with the proposition that cigarette smoking is a harmless habit with no important effects on health or longevity. The concern shown by medical and public health authorities with the increasing diffusion to ever younger groups of an agent that is a health hazard seems to us to be well founded.

* Cornfield J et al. Smoking and lung cancer: recent evidence and a discussion of some questions. JNCI 1959;22:173–203. Reprinted with permission.

1 Our attention has been called to a summary of three additional studies, which report no association between ABO blood groups and lung cancer, by Roberts JAF. Blood groups and susceptibility to disease. Brit. J. Prev. & Social Med. 11: 107–125, 1957.

Google Scholar

Google Preview

We feel obliged to give proof of the rather obvious statement on the magnitudes of relative risk because it has been suggested that the use of a relative measurement is merely "instinctive" and lacking in rational justification. Let the disease rate for those exposed to the causal agent, B, be r 1 and for those not exposed, r 2 , each rate being unaffected by exposure or nonexposure to the noncausal agent, A. Let r 1 > r 2 . Of those exposed to A, let the proportion exposed to B be p 1 , and of those not exposed to A, let the proportion exposed to B be p 2 . Because of the assumed positive correlation between A and B, p 1 > p 2 . Then

R 1 = rate for those exposed to A = p 1 r 1 + (1 – p 1 ) r 2

The proof again is simple. Let r 11 denote the risk of the disease in the presence of both A and B, r 12 , the risk in the presence of A and absence of B, r 12 , the risk in the absence of A and presence of B, and r 22 the risk in the absence of both A and B. It is reasonable to assume r 22 = 0, but the less restrictive specification r 22 < r 12 r 21 / r 11 is sufficient for what follows. The proportion of the population exposed to B is denoted by p , and this, by hypothesis, is the same whether A is present or absent. Then

R 1 = rate for those exposed to A = pr 11 + (1 – p ) r 12

• lung cancer

Email alerts

Citing articles via, looking for your next opportunity.

• About International Journal of Epidemiology
• Recommend to your Library

Affiliations

• Online ISSN 1464-3685
• Copyright © 2024 International Epidemiological Association
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

• Cancer Home
• Clinical Trials
• Acknowledgements

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 14 January 2022

Deep learning-based algorithm for lung cancer detection on chest radiographs using the segmentation method

• Akitoshi Shimazaki 1 ,
• Daiju Ueda 1 , 2 ,
• Antoine Choppin 3 ,
• Akira Yamamoto 1 ,
• Takashi Honjo 1 ,
• Yuki Shimahara 3 &
• Yukio Miki 1  

Scientific Reports volume  12 , Article number:  727 ( 2022 ) Cite this article

22k Accesses

57 Citations

7 Altmetric

Metrics details

• Lung cancer
• Radiography

We developed and validated a deep learning (DL)-based model using the segmentation method and assessed its ability to detect lung cancer on chest radiographs. Chest radiographs for use as a training dataset and a test dataset were collected separately from January 2006 to June 2018 at our hospital. The training dataset was used to train and validate the DL-based model with five-fold cross-validation. The model sensitivity and mean false positive indications per image (mFPI) were assessed with the independent test dataset. The training dataset included 629 radiographs with 652 nodules/masses and the test dataset included 151 radiographs with 159 nodules/masses. The DL-based model had a sensitivity of 0.73 with 0.13 mFPI in the test dataset. Sensitivity was lower in lung cancers that overlapped with blind spots such as pulmonary apices, pulmonary hila, chest wall, heart, and sub-diaphragmatic space (0.50–0.64) compared with those in non-overlapped locations (0.87). The dice coefficient for the 159 malignant lesions was on average 0.52. The DL-based model was able to detect lung cancers on chest radiographs, with low mFPI.

Similar content being viewed by others

Evaluation of the feasibility of explainable computer-aided detection of cardiomegaly on chest radiographs using deep learning

Short-term Reproducibility of Pulmonary Nodule and Mass Detection in Chest Radiographs: Comparison among Radiologists and Four Different Computer-Aided Detections with Convolutional Neural Net

Better performance of deep learning pulmonary nodule detection using chest radiography with pixel level labels in reference to computed tomography: data quality matters

Introduction.

Lung cancer is the primary cause of cancer death worldwide, with 2.09 million new cases and 1.76 million people dying from lung cancer in 2018 1 . Four case-controlled studies from Japan reported in the early 2000s that the combined use of chest radiographs and sputum cytology in screening was effective for reducing lung cancer mortality 2 . In contrast, two randomized controlled trials conducted from 1980 to 1990 concluded that screening with chest radiographs was not effective in reducing mortality in lung cancer 3 , 4 . Although the efficacy of chest radiographs in lung cancer screening remains controversial, chest radiographs are more cost-effective, easier to access, and deliver lower radiation dose compared with low-dose computed tomography (CT). A further disadvantage of chest CT is excessive false positive (FP) results. It has been reported that 96% of nodules detected by low-dose CT screening are FPs, which commonly leads to unnecessary follow-up and invasive examinations 5 . Chest radiography is inferior to chest CT in terms of sensitivity but superior in terms of specificity. Taking these characteristics into consideration, the development of a computer-aided diagnosis (CAD) model for chest radiograph would have value by improving sensitivity while maintaining low FP results.

The recent application of convolutional neural networks (CNN), a field of deep learning (DL) 6 , 7 , has led to dramatic, state-of-the-art improvements in radiology 8 . DL-based models have also shown promise for nodule/mass detection on chest radiographs 9 , 10 , 11 , 12 , 13 , which have reported sensitivities in the range of 0.51–0.84 and mean number of FP indications per image (mFPI) of 0.02–0.34. In addition, radiologist performance for detecting nodules was better with these CAD models than without them 9 . In clinical practice, it is often challenging for radiologists to detect nodules and to differentiate between benign and malignant nodules. Normal anatomical structures often appear as if they are nodules, which is why radiologists must pay careful attention to the shape and marginal properties of nodules. As these problems are caused by the conditions rather than the ability of the radiologist, even skillful radiologists can misdiagnose 14 , 15 .

There are two main methods for detecting lesions using DL: detection and segmentation. The detection method is a region-level classification, whereas the segmentation method is a pixel-level classification. The segmentation method can provide more detailed information than the detection method. In clinical practice, classifying the size of a lesion at the pixel-level increases the likelihood of making a correct diagnosis. Pixel-level classification also makes it easier to follow up on changes in lesion size and shape, since the shape can be used as a reference during detection. It also makes it possible to consider not only the long and short diameters but also the area of the lesion when determining the effect of treatment 16 . However, to our knowledge, there are no studies using the segmentation method to detect pathologically proven lung cancer on chest radiographs.

The purpose of this study was to train and validate a DL-based model capable of detecting lung cancer on chest radiographs using the segmentation method, and to evaluate the characteristics of this DL-based model to improve sensitivity while maintaining low FP results.

The following points summarize the contributions of this article:

This study developed a deep learning-based model for detection and segmentation of lung cancer on chest radiographs.

Our dataset is high quality because all the nodules/masses were pathologically proven lung cancers, and these lesions were pixel-level annotated by two radiologists.

The segmentation method was more informative than the classification or detection methods, which is useful not only for the detection of lung cancer but also for follow-up and treatment efficacy.

Materials and methods

Study design.

We retrospectively collected consecutive chest radiographs from patients who had been pathologically diagnosed with lung cancer at our hospital. Radiologists annotated the lung cancer lesions on these chest radiographs. A DL-based model for detecting lung cancer on radiographs was trained and validated with the annotated radiographs. The model was then tested with an independent dataset for detecting lung cancers. The protocol for this study was comprehensively reviewed and approved by the Ethical Committee of Osaka City University Graduate School of Medicine (No. 4349). Because the radiographs had been acquired during daily clinical practice and informed consent for their use in research had been obtained from patients, the Ethical Committee of Osaka City University Graduate School of Medicine waived the need for further informed consent. All methods were performed in accordance with the relevant guidelines and regulations.

Eligibility and ground truth labelling

Two datasets were used to train and test the DL-based model, a training dataset and a test dataset. We retrospectively collected consecutive chest radiographs from patients pathologically diagnosed with lung cancer at our hospital. The training dataset was comprised of chest radiographs obtained between January 2006 and June 2017, and the test dataset contained those obtained between July 2017 and June 2018. The inclusion criteria were as follows: (a) pathologically proven lung cancer in a surgical specimen; (b) age > 40 years at the time of the preoperative chest radiograph; (c) chest CT performed within 1 month of the preoperative chest radiograph. If the patient had multiple chest radiographs that matched the above criteria, the latest radiograph was selected. Most of these chest radiographs were taken as per routine before hospitalization and were not intended to detect lung cancer. Chest radiographs on which radiologists could not identify the lesion, even with reference to CT, were excluded from analysis. For eligible radiographs, the lesions were annotated by two general radiologists (A.S. and D.U.), with 6 and 7 years of experience in chest radiography, using ITK-SNAP version 3.6.0 ( http://www.itksnap.org/ ) . These annotations were defined as ground truths. The radiologists had access to the chest CT and surgical reports and evaluated the lesion characteristics including size, location, and edge. If > 50% of the edge of the nodule was traceable, the nodule was considered to have a “traceable edge”; if not, it was termed an “untraceable edge”.

Model development

We adopted the CNN architecture using segmentation method. The segmentation method outputs more information than the detection method (which present a bounding box) or the classification method (which determine the malignancy from a single image). Maximal diameter of the tumor is particularly important in clinical practice. Since the largest diameter of the tumor often coincides with an oblique direction, not the horizontal nor the vertical direction, it is difficult to measure with detection methods which present a bounding box. Our CNN architecture was based on the encoder-decoder architecture to output segmentation 17 . The encoder-decoder architecture has a bottleneck structure, which reduces the resolution of the feature map and improves the model robustness to noise and overfitting 18 .

In addition, one characteristic of this DL-based model is that it used both a normal chest radiograph and a black-and-white inversion of a chest radiograph. This is an augmentation that makes use of the experience of radiologists 19 . It is known that black-and-white inversion makes it easier to confirm the presence of lung lesions overlapping blind spots. We considered that this augmentation could be effective for this model as well, so we applied a CNN architecture to each of the normal and inverted images and then an ensemble model using these two architectures 20 . Supplementary Fig. S1 online shows detailed information of the model.

Using chest radiographs from the training dataset, the model was trained and validated from scratch, utilizing five-fold cross-validation. The model when the value of the loss function was the smallest within 100 epochs using Adam (learning rate = 0.001, beta_1 = 0.9, beta_2 = 0.999, epsilon = 0.00000001, decay = 0.0) was adopted as the best-performing.

Model assessment

A detection performance test was performed on a per-lesion basis using the test dataset to evaluate whether the model could identify malignant lesions on radiographs. The model calculated the probability of malignancy in a lesion detected on chest radiographs as an integer between 0 and 255. If the center of output generated by the model was within the ground truth, it was considered true positive (TP). All other outputs were FPs. When two or more TPs were proposed by the model for one ground truth, they were considered as one TP. If there was no output from the model for one ground truth, it was one FN. Two radiologists (A.S. and D.U.) retrospectively referred to the radiograph and CT to evaluate what structures were detected by the FP output. The dice coefficient was also used to evaluate segmentation performance.

Statistical analysis

In the detection performance test, metrics were evaluated on a per-lesion basis. We used the free-response receiver-operating characteristic (FROC) curve to evaluate whether the bounding boxes proposed by the model accurately identified malignant cancers in radiographs 21 . The vertical axis of the FROC curve is sensitivity and the horizontal axis is mFPI. Sensitivity is the number of TPs that the model was able to identify divided by the number of ground truths. The mFPI is the number of FPs that the model mistakenly presented divided by the number of radiographs in the dataset. Thus, the FROC curve shows sensitivity as a function of the number of FPs shown on the image.

One of the authors (D.U.) performed all analyses, using R version 3.6.0 ( https://www.r-project.org/ ). The FROC curves were plotted by R software. All statistical inferences were performed with two-sided 5% significance level.

Figure 1 shows a flowchart of the eligibility criteria for the chest radiographs. For the training dataset, 629 radiographs with 652 nodules/masses were collected from 629 patients (age range 40–91 years, mean age 70 ± 9.0 years, 221 women). For the test dataset, 151 radiographs with 159 nodules/masses were collected from 151 patients (age range 43–84 years, mean age 70 ± 9.0 years, 57 women) (Table 1 ).

Flowchart of dataset selection.

The DL-based model had sensitivity of 0.73 with 0.13 mFPI in the test dataset (Table 2 ). The FROC curve is shown in Fig.  2 . The highest sensitivity the model attained was 1.00 for cancers with a diameter of 31–50 mm, and the second highest sensitivity was 0.85 for those with a diameter > 50 mm. For lung cancers that overlapped with blind spots such as the pulmonary apices, pulmonary hila, chest wall, heart, or sub-diaphragmatic space, sensitivity was 0.52, 0.64, 0.52, 0.56, and 0.50, respectively. The sensitivity of lesions with traceable edges on radiographs was 0.87, and that for untraceable edges was 0.21. Detailed results are shown in Table 2 .

The dice coefficient for all 159 lesions was on average 0.52 ± 0.37 (standard deviation, SD). For 116 lesions detected by the model, the dice coefficient was on average 0.71 ± 0.24 (SD). The dice coefficient for all 71 lesions overlapping blind spots was 0.34 ± 0.38 (SD). For 39 lesions detected by the model that overlapped with blind spots, the dice coefficient was 0.62 ± 0.29 (SD).

Of the 20 FPs, 19 could be identified as some kind of structure on the chest radiograph by radiologists (Table 3 ). In these 20 FPs, 13 overlapped with blind spots. There were 43 FNs, ranging in size from 9 to 72 mm (mean 21 ± 15 mm), 32 of which overlapped with blind spots (Table 4 ). There were four FNs > 50 mm, all of which overlapped with blind spots. Figure  3 shows representative cases of our model. Figure  4 shows overlapping of a FP output with normal anatomical structures and Fig.  5 shows a FN lung cancer that overlapped with a blind spot. Supplementary Fig. S2 online shows visualized images of the first and last layers. An ablation study to use black-and-white inversion images is shown in Supplementary Data online.

Free-response receiver-operating characteristic curve for the test dataset.

Two representative true positive cases. The images on the left are original images, and those on the right are images output by our model. ( a ) A 48-year-old woman with a nodule in the right lower lobe that was diagnosed as adenocarcinoma. The nodule was confused with rib and vessels (arrows). The model detected the nodule in the right middle lung field. ( b ) A 74-year-old woman with a nodule in the left lower lobe that was diagnosed as squamous cell carcinoma. The nodule overlapped with the heart (arrows). The lesion was identifiable by the model because its edges were traceable.

Example of one false positive case. The image on the left is an original image, and the image on the right is an image output by our model. An 81-year-old woman with a mass in the right lower lobe that was diagnosed as squamous cell carcinoma. The mass in the right middle lung field (arrows) was carcinoma. Our model detected this lesion, and also detected a slightly calcified nodule in the right lower lung field (arrowhead). This nodule was an old fracture of the right tenth rib, but was misidentified as a malignant lesion because its shape was obscured by overlap with the right eighth rib and breast.

Example of one false negative case. The image on the left is a gross image, and the image on the right is an enlarged image of the lesion. A 68-year-old man with a mass in the left lower lobe that was diagnosed as adenocarcinoma. This lesion overlapped with the heart and is only faintly visible (arrows). Our model failed to detect the mass.

In this study, we developed a model for detecting lung cancer on chest radiographs and evaluated its performance. Adding pixel-level classification of lesions in the proposed DL-based model resulted in sensitivity of 0.73 with 0.13 mFPI in the test dataset.

To our knowledge, ours is the first study to use the segmentation method to detect pathologically proven lung cancer on chest radiographs. We found several studies that used classification or detection methods to detect lung cancer on chest radiographs, but not the segmentation method. Since the segmentation method has more information about the detected lesions than the classification or detection methods, it has advantages not only in the detection of lung cancer but also in follow-up and treatment efficacy. We achieved performance as high as that in similar previous studies 9 , 10 , 11 , 12 , 13 using DL-based lung nodule detection models, with fewer training data. It is particularly noteworthy that the present method achieved low mFPI. In previous studies, sensitivity and mFPI were 0.51–0.84 and 0.02–0.34, respectively, and used 3,500–13,326 radiographs with nodules or masses as the training data, compared with the 629 radiographs used in the present study. Although comparisons to these studies are difficult because the test datasets were different, our accuracy was similar to that of the detection models employed in most of the previous studies. We performed pixel-level classification of the lesions based on the segmentation method and included for analysis only lesions that were pathologically proven to be malignant, based on examination of surgically resected specimens. All previous studies 9 , 10 , 11 , 12 , 13 have included potentially benign lesions, clinically malignant lesions, or pathologically malignant lesions by biopsy in their training data. Therefore, our model may be able to analyze the features of the malignant lesions in more detail. In regard with the CNN, we created this model based on Inception-ResNet-v2 17 , which combines the Inception structure and the Residual connection. In the Inception-ResNet block, convolutional filters of multiple sizes are combined with residual connections. The use of residual connections not only avoids the degradation problem caused by deep structures but also reduces the training time. In theory, the combination of these features further improves the recognition accuracy and learning efficiency 17 . By using this model with combining normal and black-white-inversion images, our results achieved comparable or better performance with fewer training data than previous studies. In regard with the robustness of the model, we consider this model to be relatively robust against imaging conditions or body shape because we consecutively collected the dataset and did not set any exclusion criteria based on imaging conditions or body shape.

The dice coefficient for 159 malignant lesions was on average 0.52. On the other hand, for the 116 lesions detected by the model, the dice coefficient was on average 0.71. These values provide a benchmark for the segmentation performance of lung cancer on chest radiograph. The 71 lesions which overlapped with blind spots tended to have a low dice coefficient with an average of 0.34, but for 39 lesions detected by the model that overlapped with blind spots, the average dice coefficient was 0.62. This means that lesions overlapping blind spots were not only difficult to detect, but also had low accuracy in segmentation. On the other hand, the segmentation accuracy was relatively high for lesions that were detected by the model even if they overlapped with the blind spots.

Two interesting tendencies were found after retrospectively examining the characteristics of FP outputs. First, 95% (19/20) FPs could be visually recognized on chest radiographs as nodule/mass-like structures. The model identified some nodule-like structures (FPs), which overlapped with vascular shadows and ribs. This is also the case for radiologists in daily practice. Second, nodules with calcification overlapped with normal anatomical structures tended to be misdiagnosed by the model (FPs). Five FPs were non-malignant calcified lung nodules on CT and also overlapped with the heart, clavicle or ribs. As the model was trained only on malignant nodules without calcification in the training dataset, calcified nodules should not be identified in theory. Most calcified nodules are actually not identified by the model, however, this was not the case for calcified nodules that overlapped with normal anatomical structures. In other word, there is a possibility that the model could misidentify the lesion as a malignant if the features of calcification that should signal a benign lesion are masked by normal anatomical structures.

When we investigated FNs, we found that nodules in blind spots and metastatic nodules tended to be FNs. With regard to blind spots, our model showed a decrease in sensitivity for lesions that overlapped with normal anatomical structures. It was difficult for the model to identify lung cancers that overlapped with blind spots even when the tumor size was large (Fig.  5 ). In all FNs larger than 50 mm, there was wide overlap with normal anatomical structures, for the possible reason that it becomes difficult for the model to detect subtle density differences in lesions that overlapped with large structures such as the heart. With regard to metastatic nodules, 33% (14/43) metastatic lung cancers were FNs. These metastatic nodules ranged in size from 10 to 20 mm (mean 14 ± 3.8 mm) and were difficult to visually identify on radiographs, even with reference to CT. In fact, the radiologists had overlooked most of the small metastatic nodules at first and could only identify them retrospectively, with knowledge of the type of lung cancer and their locations.

There are some limitations of this study. The model was developed using a dataset collected from a single hospital. Although our model achieved high sensitivity with low FPs, the number of FPs may be higher in a screening cohort and the impact of this should be considered. Furthermore, an observer’s performance study is needed to evaluate the clinical utility of the model. In this study, we included only chest radiographs containing malignant nodules/masses. The fact that we used only pathologically proven lung cancers and pixel-level annotations by two radiologists in our dataset is a strength of our study, on the other hand, it may reduce the detection rate of benign nodules/masses. This is often not a problem in clinical practice. Technically, all areas other than the malignant nodules/masses could be trained as normal areas. However, normal images should be mixed in and tested to evaluate the model for detailed examination in clinical practice.

In conclusion, a DL-based model developed using the segmentation method showed high performance in the detection of lung cancer on chest radiographs. Compared with CT, chest radiographs have advantages in terms of accessibility, cost effectiveness, and low radiation dose. However, the known effectiveness of the model for lung cancer detection is limited. We believe that a CAD model with higher performance can support clinical detection and interpretation of malignant lesions on chest radiographs and offers additive value in lung cancer detection.

Bray, F. et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. 68 , 394–424. https://doi.org/10.3322/caac.21492 (2018).

Article   PubMed   Google Scholar  

Sagawa, M. et al. The efficacy of lung cancer screening conducted in 1990s: Four case–control studies in Japan. Lung Cancer 41 , 29–36. https://doi.org/10.1016/s0169-5002(03)00197-1 (2003).

Fontana, R. S. et al. Lung cancer screening: The Mayo program. J. Occup. Med. 28 , 746–750. https://doi.org/10.1097/00043764-198608000-00038 (1986).

Article   CAS   PubMed   Google Scholar  

Kubik, A. et al. Lack of benefit from semi-annual screening for cancer of the lung: Follow-up report of a randomized controlled trial on a population of high-risk males in Czechoslovakia. Int. J. Cancer 45 , 26–33. https://doi.org/10.1002/ijc.2910450107 (1990).

Raghu, V. K. et al. Feasibility of lung cancer prediction from low-dose CT scan and smoking factors using causal models. Thorax 74 , 643–649. https://doi.org/10.1136/thoraxjnl-2018-212638 (2019).

Hinton, G. Deep learning—a technology with the potential to transform health care. JAMA 320 , 1101–1102. https://doi.org/10.1001/jama.2018.11100 (2018).

LeCun, Y., Bengio, Y. & Hinton, G. Deep learning. Nature 521 , 436–444. https://doi.org/10.1038/nature14539 (2015).

Article   ADS   CAS   PubMed   Google Scholar  

Ueda, D., Shimazaki, A. & Miki, Y. Technical and clinical overview of deep learning in radiology. Jpn. J. Radiol. 37 , 15–33. https://doi.org/10.1007/s11604-018-0795-3 (2019).

Nam, J. G. et al. Development and validation of deep learning–based automatic detection algorithm for malignant pulmonary nodules on chest radiographs. Radiology 290 , 218–228. https://doi.org/10.1148/radiol.2018180237 (2019).

Park, S. et al. Deep learning-based detection system for multiclass lesions on chest radiographs: Comparison with observer readings. Eur. Radiol. 30 , 1359–1368. https://doi.org/10.1007/s00330-019-06532-x (2020).

Yoo, H., Kim, K. H., Singh, R., Digumarthy, S. R. & Kalra, M. K. Validation of a deep learning algorithm for the detection of malignant pulmonary nodules in chest radiographs. JAMA Netw. Open 3 , e2017135. https://doi.org/10.1001/jamanetworkopen.2020.17135 (2020).

Article   PubMed   PubMed Central   Google Scholar  

Sim, Y. et al. Deep convolutional neural network–based software improves radiologist detection of malignant lung nodules on chest radiographs. Radiology 294 , 199–209. https://doi.org/10.1148/radiol.2019182465 (2020).

Hwang, E. J. et al. Development and validation of a deep learning-based automated detection algorithm for major thoracic diseases on chest radiographs. JAMA Netw. Open 2 , e191095. https://doi.org/10.1001/jamanetworkopen.2019.1095 (2019).

Manser, R. et al. Screening for lung cancer. Cochrane Database Syst Rev CD001991. https://doi.org/10.1002/14651858.CD001991.pub3 (2013).

Berlin, L. Radiologic errors, past, present and future. Diagnosis (Berl) 1 , 79–84. https://doi.org/10.1515/dx-2013-0012 (2014).

Article   Google Scholar  

From the RECIST committee. Schwartz, L.H. et al. RECIST 1.1-Update and clarification. Eur. J. Cancer. 62 , 132–137. https://doi.org/10.1016/j.ejca.2016.03.081 (2016).

Szegedy, C., Ioffe, S., Vanhoucke, V. & Alemi, A. A. Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning . In Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence , 4278–4284 (AAAI Press, San Francisco, California, USA, 2017).

Matějka P. et al. Neural Network Bottleneck Features for Language Identification. In Proceedings of Odyssey 2014. vol. 2014. International Speech Communication Association , 299–304 (2014).

Sheline, M. E. et al. The diagnosis of pulmonary nodules: Comparison between standard and inverse digitized images and conventional chest radiographs. Am. J. Roentgenol. 152 (2), 261–263. https://doi.org/10.2214/ajr.152.2.261 (1989).

Article   CAS   Google Scholar  

Wang, G., Hao, J., Ma, J. & Jiang, H. A comparative assessment of ensemble learning for credit scoring. Expert. Syst. Appl. 38 (1), 223–230 (2011).

Bunch, P., Hamilton, J., Sanderson, G. & Simmons, A. A free response approach to the measurement and characterization of radiographic observer performance. Proc. SPIE 127 , 124–135. https://doi.org/10.1117/12.955926 (1977).

Article   ADS   Google Scholar  

Download references

Acknowledgements

We are grateful to LPIXEL Inc. for joining this study.

Author information

Authors and affiliations.

Department of Diagnostic and Interventional Radiology, Graduate School of Medicine, Osaka City University, Osaka, Japan

Akitoshi Shimazaki, Daiju Ueda, Akira Yamamoto, Takashi Honjo & Yukio Miki

Smart Life Science Lab, Center for Health Science Innovation, Osaka City University, Osaka, Japan

LPIXEL Inc, Tokyo, Japan

Antoine Choppin & Yuki Shimahara

You can also search for this author in PubMed   Google Scholar

Contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by A.S., D.U., A.Y. and T.H. Model development was performed by A.C. and Y.S. The first draft of the manuscript was written by A.S. and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Daiju Ueda .

Ethics declarations

Competing interests.

Akitoshi Shimazaki has no relevant relationships to disclose. Daiju Ueda has no relevant relationships to disclose. Antoine Choppin is an employee of LPIXEL Inc. Akira Yamamoto has no relevant relationships to disclose. Takashi Honjo has no relevant relationships to disclose. Yuki Shimahara is the CEO of LPIXEL Inc. Yukio Miki has no relevant relationships to disclose.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Supplementary information., rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Shimazaki, A., Ueda, D., Choppin, A. et al. Deep learning-based algorithm for lung cancer detection on chest radiographs using the segmentation method. Sci Rep 12 , 727 (2022). https://doi.org/10.1038/s41598-021-04667-w

Download citation

Received : 04 August 2021

Accepted : 29 December 2021

Published : 14 January 2022

DOI : https://doi.org/10.1038/s41598-021-04667-w

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

Deep volcanic residual u-net for nodal metastasis (nmet) identification from lung cancer.

• M. Ramkumar
• K. Kalirajan

Biomedical Engineering Letters (2024)

A Comprehensive Study and Research Perception towards Secured Data Sharing for Lung Cancer Detection with Blockchain Technology

• Hari Krishna Kalidindi
• N. Srinivasu

Annals of Data Science (2024)

ELCD-NSC2: a novel early lung cancer detection and non-small cell classification framework

• Hadeer A. Helaly
• Mahmoud Badawy
• Amira Y. Haikal

Neural Computing and Applications (2024)

Fairness of artificial intelligence in healthcare: review and recommendations

• Taichi Kakinuma
• Shinji Naganawa

Japanese Journal of Radiology (2024)

Lung tumor analysis using a thrice novelty block classification approach

• S. L. Soniya
• T. Ajith Bosco Raj

Signal, Image and Video Processing (2023)

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing: Cancer newsletter — what matters in cancer research, free to your inbox weekly.

• Open access
• Published: 29 July 2024

Pericytes recruited by CCL28 promote vascular normalization after anti-angiogenesis therapy through RA/RXRA/ANGPT1 pathway in lung adenocarcinoma

• Ying Chen 1 , 2 , 3 ,
• Zhiyong Zhang 1 , 2 , 3 ,
• Fan Pan 3 , 6 ,
• Pengfei Li 1 , 2 , 3 ,
• Weiping Yao 3 , 6 ,
• Yuxi Chen 1 , 2 , 3 ,
• Lei Xiong 5 ,
• Tingting Wang 1 , 2 , 3 ,
• Yan Li 4 &
• Guichun Huang 6 , 7  

Journal of Experimental & Clinical Cancer Research volume  43 , Article number:  210 ( 2024 ) Cite this article

290 Accesses

Metrics details

It has been proposed that anti-angiogenesis therapy could induce tumor "vascular normalization" and further enhance the efficacy of chemotherapy, radiotherapy, target therapy, and immunotherapy for nearly twenty years. However, the detailed molecular mechanism of this phenomenon is still obscure.

Overexpression and knockout of CCL28 in human lung adenocarcinoma cell line A549 and murine lung adenocarcinoma cell line LLC, respectively, were utilized to establish mouse models. Single-cell sequencing was performed to analyze the proportion of different cell clusters and metabolic changes in the tumor microenvironment (TME). Immunofluorescence and multiplex immunohistochemistry were conducted in murine tumor tissues and clinical biopsy samples to assess the percentage of pericytes coverage. Primary pericytes were isolated from lung adenocarcinoma tumor tissues using magnetic-activated cell sorting (MACS). These pericytes were then treated with recombinant human CCL28 protein, followed by transwell migration assays and RNA sequencing analysis. Changes in the secretome and metabolome were examined, and verification of retinoic acid metabolism alterations in pericytes was conducted using quantitative real-time PCR, western blotting, and LC–MS technology. Chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR) was employed to validate the transcriptional regulatory ability and affinity of RXRα to specific sites at the ANGPT1 promoter.

Our study showed that after undergoing anti-angiogenesis treatment, the tumor exhibited a state of ischemia and hypoxia, leading to an upregulation in the expression of CCL28 in hypoxic lung adenocarcinoma cells by the hypoxia-sensitive transcription factor CEBPB. Increased CCL28 could promote tumor vascular normalization through recruiting and metabolic reprogramming pericytes in the tumor microenvironment. Mechanistically, CCL28 modified the retinoic acid (RA) metabolism and increased ANGPT1 expression via RXRα in pericytes, thereby enhancing the stability of endothelial cells.

We reported the details of the molecular mechanisms of "vascular normalization" after anti-angiogenesis therapy for the first time. Our work might provide a prospective molecular marker for guiding the clinical arrangement of combination therapy between anti-angiogenesis treatment and other therapies.

Introduction

Lung cancer stands as the foremost cause of mortality among patients with malignant tumors, with lung adenocarcinoma being a predominant pathological subtype [ 1 , 2 ]. Angiogenesis is a crucial hallmark of lung adenocarcinoma, and significant strides have been made in applying anti-angiogenesis therapy for its treatment [ 3 ]. Extensive research data indicates that anti-angiogenic tumor therapy exhibits the potential to augment the effectiveness of diverse therapeutic modalities, including chemotherapy, radiotherapy, tyrosine kinase inhibitors (TKIs), and immunotherapy. Nevertheless, a subset of patients fails to derive benefits from the combined approach of anti-angiogenic therapy and other anti-tumor treatments. Unraveling the synergistic mechanisms underlying anti-angiogenic therapy holds substantial clinical implications for optimizing combination therapies.

Since the 2000s, there has been a fundamental shift in the understanding of anti-angiogenic therapy for tumors, transitioning from the initial concept of solely inhibiting angiogenesis to inducing vascular normalization in tumors [ 4 ]. Presently, there is a consensus that optimizing the effectiveness of additional anti-tumor treatments through anti-angiogenic therapy is imperative. This notion is central to anti-angiogenic drugs' capability to induce transient vascular normalization in tumor blood vessels [ 5 ]. Preliminary investigations undertaken by our research group propose that anti-angiogenic drugs, employing diverse mechanisms of action, can prompt vascular normalization in lung adenocarcinoma [ 6 , 7 , 8 , 9 ]. The balance between pro-angiogenesis and anti-angiogenesis factors was postulated as the key molecular base of vascular normalization [ 10 ]. However, uncertainties persist regarding the timing and the limited duration of vascular normalization post-treatment. Addressing this issue is critical, as the temporal aspects of vascular normalization and its short-lived nature need clarification [ 7 , 8 , 9 , 11 , 12 ]. The pressing clinical challenge involves establishing protocols to monitor and regulate vascular normalization in lung adenocarcinoma, offering valuable guidance for developing combined anti-angiogenic therapies with other anti-tumor treatment strategies. A comprehensive understanding of the molecular mechanisms underpinning anti-angiogenic therapy-induced vascular normalization in lung adenocarcinoma forms the foundational basis for resolving this intricate clinical dilemma.

Pericytes, one of the structural cells found in capillary vessels, envelop the basal membrane of endothelial cells and interact with vascular endothelial cells to ensure the stability and functionality of microvessels [ 13 , 14 ]. Commonly used markers for identifying pericytes include platelet-derived growth factor receptor β (PDGFRβ), chondroitin sulfate proteoglycan 4 (CSPG4, also known as NG2), and alpha-smooth muscle actin (α-SMA). Several reports suggest that platelet-derived growth factor B (PDGFB) can promote vascular normalization in colon cancer and malignant melanoma by recruiting pericytes [ 15 ]. However, PDGFB does not perform this function in pericytes in lung adenocarcinoma [ 16 ]. The mechanism of pericyte recruitment in lung adenocarcinoma following anti-angiogenesis therapy remains unclear.

Human CC motif chemokine ligands 28 (CCL28) can recruit various cell types, including lymphatic endothelial cells [ 13 ], T cells, and plasma cells, through its receptors CCR3 or CCR10. Apart from its chemotactic properties, CCL28 has been implicated in promoting tumor development in ovarian cancer [ 17 ], gastric cancer [ 18 ], liver cancer, and lung adenocarcinoma [ 19 ]. Under hypoxic conditions, tumor cells mobilize various cell types using different chemokines to induce angiogenesis in response to nutritional and oxygen requirements. Notably, hypoxia induces the expression of CCL28 in ovarian cancer, which, in turn, enhances angiogenesis by recruiting regulatory T cells (Tregs) [ 17 ]. In contrast, CCL28 has been identified as a negative regulator of tumor growth and bone invasion in oral squamous cell carcinoma [ 20 ]. These diverse findings suggest that CCL28 may play distinct roles in various tumors. While one study discovered that CCL28 increased the proliferation, migration, and secretion of IL-6 and HGF in oral fibroblasts [ 21 ], the functional role of CCL28 in pericytes in lung adenocarcinoma has not been elucidated. Furthermore, previous studies have reported that CCL28 was involved in angiogenesis in various diseases, including tumors, skin wound healing [ 22 ], and rheumatoid arthritis [ 23 ]. However, little is known about the mechanism of CCL28 in tumor vascular normalization.

Our previous report highlighted that CCL28 can moderately enhance the angiogenesis in lung adenocarcinoma [ 19 ]. Interestingly, we found a normalized vasculature in CCL28 highly expressed tumors. The current study delves into the previously undisclosed connection between pericytes and CCL28 in lung adenocarcinoma. Both in vivo and in vitro experiments collectively illustrate that CCL28 derived from tumor cells is pivotal in advancing vascular normalization by mobilizing and reprogramming pericytes.

Materials and methods

Cancer cell lines.

The human lung adenocarcinoma cell line (A549, SPC-A1, and H1975) and mouse Lewis lung cancer cell line (LLC) were purchased from the Shanghai Institute of Biochemistry and Cell Biology (SIBCB) and maintained in our lab. Lung cancer cells were cultured in RPMI-1640, or DMEM medium (Gibco, LifeTech, USA) supplemented with 10% fetal bovine serum (FBS), 1% antibiotics (100U/ml penicillin and 100 μg/ml streptomycin) at 37℃ in a humidified 5% CO 2 atmosphere.

Isolation and identification of pericytes

Pericytes were isolated and cultured as we previously described [ 24 ]. Fresh lung cancer samples were collected from patients in Jinling Hospital (Nanjing, China). Tissue samples were cut into small blocks of approximately 1 ~ 2 mm in diameter, digested with trypsin and 0.5% collagenase, and filtered through the cell strainer to obtain single-cell suspension. PDGFRβ + cells were isolated by PE-PDGFRβ antibody and anti-PE magnetic beads (Miltenyi Biotec,130–123-772). The separated cells were cultured in F12K medium (Gibco, LifeTech, USA) containing 10% fetal bovine serum (Gibco, Life Tech, USA) with 100 U/mL penicillin and 100 μg/mL streptomycin. After two or three passages, pericytes were identified by morphology and immunofluorescence staining for α-Smooth Muscle Actin (α-SMA, CST, #19,245), platelet-derived growth factor receptor β (PDGFRβ, Abcam, ab69506) and chondroitin sulfate proteoglycan 4 (NG2, Abcam, ab129051). Patients with incomplete data were excluded to evaluate the clinical effect. Written informed consent was obtained from all subjects before collecting the samples. All the methods followed the institutional guidelines and were approved by the Ethical Review Committee of Jinling Hospital, Nanjing, China(2022DZGZR-QH-005).

Hypoxic culture model and RNA sequence assay

As previously reported, the hypoxic cell culture model was established with hypoxic chambers in our lab [ 19 ]. Briefly, lung adenocarcinoma cells were cultured under two different oxygen concentrations, 1% and 20%, respectively. The model was testified by the expression changes of HIF-1α and its regulated genes, such as GLUT1 and VEGFA (Fig. 1 C). Pericytes were cultured under different stimulation. After culturing for 24 h, the cells were collected, and the total amount of RNA or protein was extracted for western blot or quantitative PCR. RNA sequence assay was applied to detect the gene expression differences of pericytes cultured with or without the stimulation of CCL28 (MCE, HY-P7250) under hypoxia.

CCL28 expression is upregulated after anti-angiogenesis therapy by hypoxia-sensitive transcription factor CEBPB in lung adenocarcinoma

RNA extraction and quantitative PCR

Total RNA was extracted from different cell lines using Trizol reagent (Invitrogen, USA). Subsequently, reverse transcription and quantitative PCR were performed using SYBR Green with an ABI StepOne Plus System (Applied Biosystems, Life Tech, USA). Relative gene expression was calculated by the ΔΔCt method based on glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or β-actin levels. All reactions were run in triplicate.

Luciferase reporter assay

Wild-type (pGL3 wt, 5’-TGATTATGCAATGG-3') and mutant (pGL3 mut, 5’-ACTAATACGTTACC-3') promoters of the CCL28 gene were constructed into pGL3 firefly luciferase reporter plasmid vector purchased from Nanjing Realgene Bio-Technology Company (Nanjing China). The pRL-TK vector expressing renilla luciferase was used as an internal control. The reporter plasmid was co-transfected with pRL-TK vector and CEBPB expressing pcDNA3.1 vector or control vector into A549 cell. Dual-Luciferase Reporter Assay System (Promega, USA) was used to detect the luciferase activity after 48-h incubation.

Chromatin Immunoprecipitation (ChIP)

Briefly, pericytes treated with or without recombinant human CCL28 or CCR3 (R&D Systems, MAB155-100) neutralizing antibodies were collected and fixed by adding a cross-linking agent, formaldehyde, to stabilize the interactions between chromatin proteins and DNA. Subsequently, chromatin immunoprecipitation was conducted to enrich DNA fragments bound by chromatin proteins. After cross-link reversal and DNA purification, real-time quantitative PCR (qPCR) technology is employed to measure enriched DNA, determining the relative abundance at specific gene loci.

Liquid Chromatography-Mass Spectrometry (LC–MS)

Liquid chromatography-mass spectrometry was applied to measure retinoic acid (RA, HY-14649) in pericytes. Briefly, after treatment with or without CCL28, pericytes were subjected to enzymatic digestion, washed three times with PBS, flash-frozen in liquid nitrogen for one minute, and then stored at -80 °C. When performing the detection, begin by thawing the sample at 4 °C. Next, add 0.5 mL of methanol solution, followed by 10 min of sonication and 30 min of shaking for extraction. Subsequently, place the centrifuge tube in a low-temperature centrifuge and centrifuge for 10 min at 4 °C and 12,000 rpm. Collect 800μL of the supernatant, evaporate it, and then add 100μL of methanol solution. Finally, retain 80 μL of the supernatant for subsequent liquid chromatography-mass spectrometry analysis. A standard curve is generated to calculate the sample's content using a retinoic acid standard (yuanye Bio-Technology).

Single-cell RNA sequencing

Lewis lung cancer cells overexpressing CCL28 were inoculated in C57BL/6 mice subcutaneously (1 × 10 6 cells per mouse). When the tumor grew to about 100 mm 3 , fresh tumor samples were collected and dissociated into single cells. The tissue was washed with PBS, then chopped and incubated in an enzyme digestion solution for 40 min at 37 °C. After digestion, the solution was filtered through a 40 μm filter and centrifuged. Red blood cells were removed using a red blood cell lysis solution, followed by cell counting. The cell suspension was filtered using FACS tubes, centrifuged again, and washed twice. Finally, after microscopic examination and cell counting, the single-cell suspension was used for scRNA sequencing.

The scRNA sequencing analysis was carried out according to a single cell analysis workflow with BD Rhapsody™ Systems (BD Biosciences, USA), and RNA sequencing and data analysis were completed on the platform of NovaSeq 6000 system (Illumina, USA). The entired scRNA sequencing workflow was provided in the Supplementary document 1.

Animal model

Six-week-old female BALB/c nude mice and C57BL/6 female mice were used for tumor model establishment. LLC (1 × 10 6 per mouse) with or without overexpressing CCL28 were subcutaneously injected in C57BL/6 mice ( n  = 6, each group). Mice inoculated with CCL28 knock-out LLC cells and wild-type LLC were randomly divided into two groups, respectively, and treated with or without RA by gavage daily (n = 6, each group). Mice were observed, and the tumors' length and width were measured daily. When tumors grew to a specific size, subcutaneous tumors were collected and photographed. The A549 (CCL28 overexpression or CCL28 knock-out) tumor model followed the same procedure. All animal experiments were carried out following the institutional guidelines and approved by the Ethical Review Committee of Comparative Medicine, Jinling Hospital, Nanjing, China (2022DZGKDWLS-0091).

Western blot and ELISA

Proteins were extracted from the cultured cells by lysis buffer, separated by SDS-PAGE, and transferred to PVDF membranes (Millipore, USA). The filters were blocked in Tris-buffered saline containing 0.2% Tween plus 5% non-fat milk and incubated with primary antibodies overnight at 4 °C. Secondary antibodies were used for visualization through chemiluminescence (ECL, Amersham Pharmacia Biotech, UK). Primary antibodies against RXRα (CST,1:1000), RARα (CST,1:1000), RDH13 (Proteintech,1:1000), DHRS11 (Proteintech,1:1000), CCR10 (Abmart,1:1000), β-actin (Servicebio, 1:1000) and neutralizing antibody against CCR3 (R&D Systems, USA) were applied in the present study. Anti-rabbit IgG, HRP-linked Antibody (CST, 1:1000) and anti-mouse IgG, HRP-linked Antibody (CST, 1:1000) were utilized as the secondary antibodies. CCL28 in serum was detected by an ELISA kit (Abcam, USA) according to the manufacturer's instructions.

Immunofluorescence and multiplex Immunohistochemistry (mIHC)

Tumor samples were collected and fixed in 10% formalin before processing and paraffin embedding. Immunofluorescence was conducted on 5 µm sections. For culture cell staining, lung adenocarcinoma-associated pericytes were washed with 1 × PBS and fixed with acetone for 10 min on ice. Tumor microvascular endothelial cells and cancer pericytes were stained by CD31 antibody (Abcam, ab281583) and rat anti-human/mouse monoclonal NG2 antibody (Abcam, ab129051). At the same time, tumor cells were stained by mouse anti-human pan-cytokeratin (Pan-ck) monoclonal antibody (Abcam, ab215838). Goat anti-mouse IgG antibody (labeled with Alex Fluor 488), goat anti-rabbit IgG antibody (labeled with Alex Fluor 555 or Alex Fluor 488), and goat anti-rat IgG antibody (labeled with Alex Fluor 555 or Alex Fluor 488) (Abcam) were applied as secondary antibody in immunofluorescence staining analysis.

For multiplex immunohistochemistry, the slides were incubated with the primary antibodies (anti-CD31, anti-NG2, anti-Pan-CK antibody, and anti-CCL28 (Proteintech, 18,214–1-AP), respectively) and horseradish peroxidase-conjugated secondary antibody, and tyramine signal amplification (TSA) was performed following the pre-optimized antibody concentration and the order of staining. Antibody stripping and antigen retrieval were performed after each round of TSA. DAPI (Sigma-Aldrich, USA) was used for nuclei staining. A whole slide scan of the multiplex tissue sections produced multispectral fluorescent images visualized in SlideViewer software. A specialized pathologist chose representative regions of interest (ROI), and multiple fields of view were acquired at 20 × power for further analysis. The mean fluorescence intensity of CCL28 in each ROI was calculated by the ImageJ program ( https://imagej.nih.gov/ij/ ).

CRISPR-cas9 knock out

Briefly, three designed sgRNAs were synthesized and co-transfected with Cas9 nuclease/sgRNA in A549 lung adenocarcinoma cells. After 72 h, the cells were seeded in a 96-well plate for single-cell cloning by limiting dilution analysis (LDA). After 7 to 15 days, about ten single-cell clones were selected for PCR and ELISA. Mutant sites were further confirmed by DNA sequencing.

Knockdown of RDH13 and DHRS11

For RDH13 and DHRS11 gene silence, pericytes were infected with negative control siRNAs, RDH13-targeting, or DHRS11-targeting siRNAs (genepharma). Briefly, pericytes were seeded in 24-well plates, and the cells were transfected with siRNAs using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s instructions.

Bioinformatic analysis

Transcription factor binding sites prediction in promotor of CCL28 gene was conducted in PROMO 3.0, virtual laboratory TFSEARCH (ver.1.3) and JASPAR ( https://alggen.lsi.upc.edu/rerecer/menu_recerca.html , https://diyhpl.us/~bryan/irc/protocol-online/protocol-cache/TFSEARCH.html , https://jaspar.elixir.no/ ). RNA sequencing dates of lung adenocarcinoma tumor samples or lung adenocarcinoma cell lines were extracted on the cBioPortal platform ( https://www.cbioportal.org/ ). The relationship between expression levels of CEBPB and CCL28, RXRα, and ANNGPT1 were calculated online, as well as the predictive value of CEBPB for the survival of lung adenocarcinoma patients.

Statistical analyses

Data were presented as mean ± SEM. Student's unpaired two-tailed tests were used for comparisons between two groups. Pearson or Spearman correlation was applied to analyze the relationship between the expression scores of two genes. Statistical analyses were performed on GraphPad Prism 8.0. Multiple comparisons were analyzed by one-way ANOVA using the LSD test. Statistical significance was confirmed when p  < 0.05.

CCL28 expression is up-regulated after anti-angiogenesis therapy by hypoxia-sensitive transcription factor CEBPB in lung adenocarcinoma

The GEO database ( https://www.ncbi.nlm.nih.gov/gds/ ) was retrieved to elucidate the impact of anti-angiogenesis therapy on the gene expression of cancer patients. Molecular expression profile data (GSE61676) of lung adenocarcinoma patients undergoing anti-angiogenesis treatment was reanalyzed. Twenty-three stage IV lung adenocarcinoma patients, treated with a combination of anti-VEGF monoclonal antibody (Bevacizumab) and TKI, were selected for analysis. The changes in mRNA profile (Affymetrix Human Exon 1.0 ST Arrays) were examined before and 24 h after Bevacizumab treatment. The findings revealed a significant increase in the hypoxia index (GLUT1, one of the primary target genes regulated by hypoxia-inducing factor HIF-1) and up-regulation of CCL28 expression in lung adenocarcinoma patients 24 h after VEGF monoclonal antibody treatment compared to baseline levels (Fig. 1 A, 1B). After treatment with bevacizumab, CCL28 expression showed a log2 fold change of 1.258 compared to baseline, with a corresponding p-value of 0.0152, as depicted in Fig. 1 A. After anti-angiogenesis therapy, tumor cells often experience a state of ischemia and hypoxia. To simulate this hypoxic state of tumor cells, nine cell lines of 4 different tumor types were culture in hypoxia and normoxia conditions, including two hepatoma cell lines (HEPG2 and SMMC), three lung adenocarcinoma cell lines (A549, SPC-A1, and H1975), two breast carcinoma cell lines (MCF7 and MDA-MB-231), and two colorectal cancer cell lines (HCT116 and SW480). The expression of GLUT1 and VEGFA, which was driven by the hypoxia-induced factor (HIF), was detected by qRT-PCR. VEGFA and GLUT1 were up-regulated in all nine cell lines in the hypoxic chamber, indicating that the hypoxic microenvironment was established (Fig. 1 C). We previously reported that high expression of CCL28 was induced in hypoxic lung adenocarcinoma cell lines, A549 and SPC-A1. To further confirm our study, we examined the expression of CCL28 under hypoxia conditions in the nine cell lines. CCL28 was up-regulated in all three lung adenocarcinoma cell lines, whereas others were not (Fig. 1 D). Then, we utilized PROMO 3.0 to predict the transcription factors binding sites to the CCL28 promoter region and intersected them with previously reported hypoxia-related transcription factors, revealing six transcription factors involved (Fig. 1 E). CEBPB and Sp1 can potentially regulate CCL28, while only CEBPB was highly expressed in hypoxic tumor cells (as listed in Supplementary Table 1). We then used a public integrative database (cBioPortal) to analyze the correlation of the expression of CEBPB with that of CCL28 in 44 lung adenocarcinoma cell lines (as listed in Supplementary Table 2). As expected, CEBPB expression strongly correlated with CCL28 (Fig. 1 F). The CEBPB binding motif was shown in Fig. 1 G. Thus, we speculated that hypoxia-induced high expression of CCL28 was mediated by CEBPB. To verify our prediction, we generated luciferase reporter plasmids harboring either wild-type (WT) or mutated (MUT) CEBPB binding sequencing within the promotor element of the CCL28 gene. Luciferase reporter analysis indicated that CEBPB could directly regulate the expression of the CCL28 gene in A549 cells (Fig. 1 H). In addition, disease-free survival and overall survival of lung adenocarcinoma patients with high CEBPB expression were significantly decreased ( p  = 0.0065 and p  = 0.032, respectively) (Fig. 1 I).

Tumor-derived CCL28 promotes vascular normalization and pericyte recruitment in the tumor microenvironment

To investigate the effects of CCL28 on tumor growth and vascular normalization, we established CCL28 overexpression and knock-out lung adenocarcinoma cells by lentivirus vectors and Cas9 nuclease/sgRNA (Supplementary Fig. 1), respectively. As we previously reported, in vivo studies indicated that CCL28 could promote tumor growth in A549 human lung adenocarcinoma (Fig. 2 A). Microvessels and pericytes were subsequently assessed by immunofluorescence staining using antibodies against CD31 (red) and NG2 (green) in tumor tissue in different groups. The percentage of pericyte coverage and microvessel density (MVD) in A549-CCL28 tumor tissue significantly increased compared to A549-NC, and this effect was reversed after CCL28 was knocked out (Fig. 2 B). These results suggested that CCL28 could promote vascular normalization and mobilize pericytes.

Tumor-derived CCL28 recruits pericytes to promote vascular normalization in the tumor microenvironment

Further, we isolated primary pericytes from lung cancer tissues to explore how CCL28 recruits pericytes in vitro. The characteristics of pericytes were identified by immunofluorescence staining α-SMA, PDGFRβ, and NG2 (Fig. 2 C). Cell viability remained unaffected by different concentrations of recombinant CCL28 in pericytes and A549 (Supplementary Fig. 2A,2B), but enhanced cell migration of A549 (Supplementary Fig. 2C). The transwell migration assay was performed to confirm the effect of CCL28 on the migration of pericytes by using different concentrations of human recombinant CCL28 protein (Fig. 2 D, left panel). The number of cells migrated to the lower chamber was calculated (Fig. 2 D, right panel), and we found that the optimal concentration of CCL28 for enhancing the migration efficiency of pericytes is 250 ng/ml. Thus, we selected a concentration of 250 ng/ml to conduct subsequent migration experiments. To further verify the recruitment effect of CCL28 on pericytes, cells were inoculated around matrigel mixed with or without CCL28 recombinant protein. The graphs visually illustrate cell migration distance and quantity (Fig. 2 E, left panel). CCL28 significantly increased the migration distance of pericytes compared to the control group (Fig. 2 E, right panel). Subsequently, we assessed the expression of two CCL28 receptors, CCR3 and CCR10, in pericytes with or without CCL28. Western blot experiments revealed that the expression of CCR3 was significantly higher than that of CCR10 in pericytes, suggesting that CCR3 likely plays a predominant role in mediating CCL28 signal transduction within pericytes (Fig. 2 F). Blocking CCR3 with neutralizing antibodies diminished the chemotactic effect of CCL28 on pericytes (Fig. 2 G). Further, we validated the relationship between CCL28 and pericytes coverage in biopsy tissue using multiplex immunofluorescence. CCL28 was mainly expressed in lung adenocarcinoma cells due to the fluorescence overlapping between CCL28 and PAN-CK (Fig. 2 H). We found a significant positive correlation between CCL28 expression levels and the percent of pericytes coverage (Fig. 2 I). The above findings indicate that CCL28 could recruit pericytes through the receptor CCR3, thereby promoting vascular normalization. Because the action of CCL28 leads to vascular normalization, making the vascular network healthier and more regular, it may also increase the density and functionality of tumor blood vessels, providing more nutrients and oxygen to promote tumor growth.

Tumor-derived CCL28 promotes expression of angiopoietin-1 via CCR3 in pericytes

Pericytes interact with vascular endothelial cells to promote the maturation of neovasculature [ 25 ]. However, the molecular mechanisms underlying pericytes-induced vascular normalization in lung adenocarcinoma after anti-angiogenesis therapy remain unclear. To clarify the exact role and potential mechanism of CCL28 in vascular normalization, we mimicked the hypoxic microenvironment in lung adenocarcinoma, cultured vascular pericytes, and conducted secretome analysis after CCL28 treatment. We detected augmented production of angiopoietin-1 in pericytes stimulated by CCL28 (Fig. 3 A). Compared to the control group, the expression of ANGPT1 changed by 5.415-fold under CCL28 stimulation, with a p-value of 0.033. RNA-seq results were confirmed by quantitative PCR in different concentrations of recombinant CCL28 protein under hypoxia conditions (Fig. 3 A right panel). Angiopoietin-1, a protein typically associated with angiogenesis and vascular normalization, maintains vascular stability and integrity by interacting with endothelial cells. Consistent with prior studies [ 24 , 26 ], western blot experiments showed that recombinant human angiopoietin-1 could up-regulate endothelial nitric oxide synthase (eNOS) expression in a dose-dependent manner in human umbilical vein endothelial cell (HUVEC) (Fig. 3 B). Likewise, recombinant human angiopoietin-1promoted cell survival by activating the PI3K-AKT signaling pathway in a Tie2-dependent manner in endothelial cells (Fig. 3 C). To further explore the regulatory mechanism of ANGPT1 expression, we identified a significant increase in the transcription factor RXRa and RARα after CCL28 stimulation in transcriptome data, verified by qPCR (Fig. 3 D and E ). However, protein level detected by western blot, CCL28 up-regulated RXRα and CCR3 but not RARα (Fig. 3 F). Also, the effect of CCL28 on RXRα up-regulation was reversed by adding a CCR3 neutralizing antibody (Fig. 3 G). Next, we confirmed the relation between transcription factor RXRα and expression of ANGPT1. Correlation analysis showed that RXRα expression positively correlates with ANGPT1 expression (Fig. 3 H). The binding motif of the RARα and RXRα complex (Fig. 3 I) and the transcription factor binding sites in the promotor area of the CCL28 gene (Fig. 3 J) were shown. CHIP-qPCR was performed to investigate the binding of transcription factor RXRα to the promoter region of the target gene ANGPT1 . Compared to the NC group, CCL28 enhanced the degree of enrichment, and this effect could be dampened after the CCR3 neutralization, suggesting a direct interaction between RXRα and the ANGPT1 promoter. The electrophoresis image displayed the specificity of the amplified DNA fragments (Fig. 3 K).

Tumor-derived CCL28 promotes the expression of angiopoietin-1 via CCR3 in pericytes

Furthermore, the mRNA level of ANGPT1 was increased by the addition of CCL28 but attenuated by the blockade of CCR3 (Fig. 3 L). RXRα is a subtype of nuclear receptor closely associated with vitamin A metabolism and retinoic acid signaling. These results led us to speculate whether retinoic acid metabolism was also involved in regulating vascular normalization by CCL28. Thus, we measured the retinoic acid content in pericytes using LC–MS technology. Interestingly, the retinoic acid production was notably increased (almost 2.5-fold change) compared to the control group in pericytes stimulated by CCL28 (Fig. 3 M). Additionally, the exogenous addition of retinoic acid in pericytes could promote the expression of ANGPT1 in a dose-dependent manner (Fig. 3 N). These results indicate that CCL28 can activate the nuclear transcription factor RXRα through the CCR3 receptor, promoting the expression of ANGPT1 in pericytes. In interaction with endothelial cells, pericytes-deriore, the mRNA level of ANGPT1 was increased by the addition of CCL28 but attenuated by the blockade ofon.

CCL28 activates retinoic acid signaling in pericytes through CCR3

We further analyzed the sequencing data to understand better the mechanism responsible for retinoic acid accumulation in pericytes after CCL28 stimulation. Enrichment analysis of gene pathways showed that RA signaling pathways were activated after CCL28 stimulation (Fig. 4 A). Two key enzymes involved in retinoic acid metabolism have changed, including RDH13 and DHRS11 (Fig.4B). RDH13 plays a role in converting retinol to retinaldehyde, a critical rate-limiting enzyme step in the retinoid metabolic pathway. However, DHRS11 reduces retinoic acid metabolites to related forms of retinol. In this way, the balance between RDH13 and DHRS11 can affect the level of retinoic acid in cells. The metabolic balance model diagram of retinoic acid shows the transformation process of the three substances (Fig. 4 C). In this metabolic process, RDH13 was up-regulated, but DHRS11 was decreased in pericytes after treatment of CCL28 (Fig. 4 B). These results were confirmed by qPCR (Fig. 4 D) and western blot (Fig. 4 E). The correlation analysis exhibited a positive relationship between RDH13 and CCL28 in lung adenocarcinoma (Fig. 4 F). Expression changes of RDH13 and DHRS11 in pericytes stimulated by CCL28 could be reversed by blocking CCR3 (Fig. 4 G). However, treating pericytes directly with retinoic acid did not result in changes in key molecules involved in RA metabolism (Supplementary Fig. 3 A). To further investigate whether the expression of two key enzymes in the retinoic acid synthesis process is correlated with the expression of ANGPT1, we knocked down the expression of RDH13 and DHRS11 using siRNA and then examined the expression of ANGPT1. After knockdown of RDH13 (Fig. 4 H), RXRα and ANGPT1 were significantly reduced (Fig. 4 I, J). However, DHRS11 did not exhibit this effect (Supplementary Fig. 3B, 3C). To further elucidate the relationship between CCL28 and its downstream molecules, immunofluorescence staining of key molecules—RDH13, DHRS11, and ANGPT1—was performed on tumor biopsy tissues (Fig. 4 K, left panel), followed by correlation analysis. In these samples, we observed a positive correlation between CCL28 expression and the levels of ANGPT1 and RDH13, while a negative correlation was found with DHRS11 (Fig. 4 K, right panel).All these results illuminated that CCL28 promoted the retinoic acid synthesis process by disturbing the balance between RDHs and DHRS, but the changes are temporary and dynamic.

Retinoic acid signaling is activated by CCL28 in pericytes through CCR3 A , Volcano plot of changes in metabolic pathways after CCL28 stimulation. B Volcano plot of the enrichment of gene expression after CCL28 stimulation. C Diagram of the metabolic conversion process in the retinoic acid metabolic signaling pathway. D and E Expression of RDH13 and DHRS11 detected by qPCR and western blot with or without exogenous supplement of CCL28. F Correlation of expression of CCL28 with RDH13 in lung adenocarcinoma. G The protein level of DHRS11 and RDH13 stimulated with or without CCL28 and CCR3 neutralizing antibody in pericytes (left) and gray value was calculated(right). H Knockdown efficiency of RDH13 was confirmed by qPCR. I and J Relative expression of RXRα and ANGPT1 after knockdown of RDH13 with or without stimulation of CCL28. K Representative immunofluorescence images of PAN-CK, NG2, CCL28 with DHRS11 or RDH13 or Angiopoietin-1 on biopsy tissues from lung cancer patients (left panel). Scale bar = 100 μm. The correlation between the expression of CCL28 and the levels of DHRS11, RDH13, and angiopoietin-1 (right panel). Data with error bars are shown as mean ± SEM. Each symbol represents data from a replicate. Each panel is a representative experiment of at least three independent biological replicates. *, **, *** represent p  < 0.05, p  < 0.01 and p  < 0.001, respectively. Abbreviation: MFI, Mean fluorescence intensity

Both CCL28 and retinoic acid could promote vascular normalization in vivo

Further, we investigated the role of CCL28 in tumor growth and vascular normalization in immunocompetent mice. We overexpressed CCL28 in LLC cells, which was validated by ELISA. Compared to the control group, CCL28 overexpression showed a significant increase in CCL28 concentration in both the cell supernatants and serum in mice implanted with tumors (Supplementary Fig. 4A, 4B). As expected, in vivo studies indicated that CCL28 could promote tumor growth in Lewis lung adenocarcinoma (LLC) (Fig. 5 A). Compared to the control group, NG2 + pericytes accumulation and angiogenesis enhanced in the LLC-CCL28 group (Fig. 5 B).

To more clearly explore the function of CCL28 on stromal cells and immune cells in the tumor microenvironment, we performed single-cell sequencing analysis on tumor tissues of LLC-NC and LLC-CCL28. Single-cell sequencing analysis provided a panoramic study of the tumor microenvironment, and the cell population in the tumor microenvironment underwent significant alterations. In total, 6380 cells were obtained after quality filtering in the 2 conditions. Specifically, there were 3829 cells in the CCL28 overexpression group, and 2551 cells in the control group. The t-SNE diagrams showed that these cells were divided into 19 cell clusters with a total of 10 cell types (Fig. 5 C and Supplementary Fig. 4D, 4E). Of all the cell types, three types showed the most significant changes in cell proportions: fibroblasts, pericytes, and macrophages, with fold changes of 2.39, 1.95, and 1.45, respectively. The percentage of pericytes was 1.09% in LLC-NC and rose to the rate of 2.14% in LLC-CCL28 (Fig. 5 D). In the CCL28 overexpressing group, we identified 82 pericytes, whereas the control group had 28 pericytes. This significant difference suggested that CCL28 may play a pivotal role in influencing pericyte populations within the tumor microenvironment. Moreover, a previous study has reported that CCL28 could recruit Tregs and promote angiogenesis in ovarian carcinoma [ 17 ]. However, the percentage of T cells was reduced in the CCL28 up-regulated LLC, indicating that CCL28 might play different roles in different cancer types. Moreover, the number of DC cells remains unchanged, but their proportion decreased (Fig. 5 D).

Subsequently, we focused on the analysis of pericytes. Cluster 14, which was CSPG4 + and ACTA2 + , was recognized as pericytes. The marker genes t-SNE plot for Cluster 14 was presented in Supplementary Fig. 5. Interestingly, metabolic pathway enrichment analysis revealed that the retinol metabolism was activated in 13 clusters, including DC, macrophages, fibroblasts, and pericytes, suggesting that the synthesis and metabolism of retinoic acid play essential physiological roles in tumor microenvironment (Fig. 5 E). In the cellular communication network, pericytes interact closely with various cell types. The heat map indicated that pericytes exhibited the strongest interaction with themselves, followed by cluster 3 and cluster 6, representing tumor and endothelial cells. Cluster 3 was characterized as Ki-67-positive and Top2a-positive, and cluster 6 was positive for Pecam (Fig. 5 F). Our results in Fig. 3 M also suggested CCL28 can promote retinoic acid accumulation in pericytes in vitro. We further investigated the effects of retinoic acid by daily gavage on tumor growth and vascular normalization in vivo. Knocking out CCL28 (LLC-CCL28-KO) or supplementing retinoic acid (LLC-NC + RA) could suppress tumor growth in LLC models (Fig. 5 G and Supplementary Fig. 4C). However, the suppression effect is more pronounced after knocking out CCL28 than the control group (LLC-NC), and CCL28 knockout with supplementing RA could further synergistically inhibit tumor growth (Fig. 5 G). Multiplex immunofluorescence was used to detect the normalization of tumor blood vessels. We found that compared to the control group, knocking out CCL28 resulted in a significant reduction in the proportion of NG2 + cells and decreased coverage of pericytes. As expected, supplementation of retinoic acid was able to promote a certain degree of restoration of pericytes coverage both in wild-type and CCL28 knockout tumors (Fig. 5 H). Then, we quantified the vascular density in different groups and found that, compared to the control group, CCL28 knockout resulted in significantly decreased vascular density, while retinoic acid had only a slight effect (Fig. 5 H). Additionally, we evaluated the effect of retinoic acid on tumor cell viability using a CCK8 assay (Supplementary Fig. 4F). The results indicated that retinoic acid did not significantly affect tumor cell viability at pharmacological concentrations (1–2 μg/ml, equivalent to 3.33–6.66 μM). However, cytotoxic effects were observed at higher concentrations, which exceed pharmacologically relevant levels in the body necessary to support essential cellular functions. To confirm the influence on tumor hypoxia, we immunofluorescently labeled mouse tumor tissues using CA9 (Carbonic Anhydrase 9), which is considered an indicator of hypoxia. Our findings indicated that compared to the control group, tumor hypoxia significantly increased in tumors with CCL28 knockout. However, retinoic acid does not appear to have a significant effect on hypoxia (Supplementary Fig. 4G). These results indicated that RA inhibited tumor growth and enhanced vascular normalization, rather than angiogenesis.

The above data indicate that CCL28-regulated retinoic acid plays a crucial role in vascular normalization, suggesting its potential as a therapeutic agent in anti-angiogenic tumor treatment. To investigate whether the combination of retinoic acid and bevacizumab has a synergistic effect, we established an A549 mouse model and administered intravenous injections of bevacizumab, coupled with oral gavage of retinoic acid. We observed that retinoic acid and bevacizumab could attenuate tumor growth in mice, respectively. Moreover, simultaneous administration of retinoic acid and bevacizumab had a synergistic effect on tumor growth (Fig. 5 I). These data indicate that retinoic acid can be used in combination with bevacizumab for tumor treatment, providing a new direction for clinical combination therapy.

CCL28 is involved in bevacizumab-mediated vascular normalization

Moreover, we investigated the synergistic effects of CCL28 knock-out and VEGF blocking on the tumor growth and vascular normalization of lung adenocarcinoma. Subcutaneously implanted lung adenocarcinoma cells with CCL28 knock-out grew much slower than wild-type tumor cells, while combination of VEGF blocker could stop the growth of the tumors (Fig. 6 A). In addition, significant promotion of vascular normalization is observed after bevacizumab treatment. However, this effect disappeared after knocking out CCL28, regardless of whether bevacizumab was added (Fig. 6 B). Knocking out CCL28 could inhibit tumor angiogenesis, and the effect was more pronounced when bevacizumab was combined (Fig. 6 B). These results indicated that CCL28 could participate in bevacizumab-mediated vascular normalization, and CCL28 might be a potential target for anti-angiogenesis therapy in lung adenocarcinoma.

Further, we evaluated the expression levels of CCL28 within the tumor tissue (Fig. 6 C, upper panel). The mean fluorescence intensity of CCL28 was significantly upregulated when bevacizumab was used compared to the control group (Fig. 6 C, lower panel), consistent with the results we found in the clinical samples. Also, the expression level of CCL28 in lung adenocarcinoma correlates with therapeutic efficacy. In lung adenocarcinoma patients, based on their treatment outcomes with bevacizumab, they were categorized into two groups: poor responders and good responders. PAN-CK, NG2, CD31, and CCL28 were stained in tumor tissues to analyze their expression levels with the therapeutic efficacy of bevacizumab treatment. Patients who responded well to bevacizumab showed significantly increased CCL28 expression, and high expression of CCL28 in tumor cells,, was associated with enhanced vascular maturity and favorable treatment outcomes (Fig. 6 D, 6E). The conclusion was evidenced by a notable decrease in malignant pericardial and pleural effusion and a significant reduction in metastatic lymph nodes after bevacizumab-based treatment (Fig. 6 F). Then, we analyzed the expression levels of DHRS11 and RDH13 in two groups of patients. Consistently, the therapeutic efficacy of bevacizumab appeared to be positively correlated with RDH13 expression and negatively correlated with DHRS11 expression (Fig. 6 G, H). These results suggest that the favorable response to anti-angiogenic therapy in patients may be attributed to the activation of CCL28, thereby deepening our understanding of treatment response variations.

Anti-angiogenesis therapy stands as a pivotal approach for metastatic lung adenocarcinoma [ 3 , 27 ], targeting the well-established VEGF/VEGFR pathway with numerous drugs developed over the last four decades [ 28 , 29 ]. However, its clinical impact has proven more complicated than initially anticipated. Recognizing the vasculature normalization effects of anti-angiogenic drugs has led to their integration into combination regimens with chemotherapy, radiotherapy, immunotherapy, and targeted therapy [ 4 ]. For example, bevacizumab, through the modulation of angiogenesis and improvement of the tumor microenvironment, lowers tumor vascular density and edema, thereby enhancing the efficiency of oxygen supply and making other treatment modalities more effective. Consequently, anti-angiogenesis drugs are employed as modulators for tumor vasculature and the microenvironment within combination regimens.

Anti-angiogenic therapy inhibits abnormal blood vessel formation, fostering vascular normalization through mechanisms like reducing vessel density, remodeling the extracellular matrix, regulating inflammation, balancing growth factors, and modulating the immune system. The tipped balance between pro-angiogenesis and anti-angiogenesis factors contributes to maintaining tumor vascular growth [ 30 ]. Anti-angiogenic therapy disrupts this tipped balance by inhibiting VEGF, enhancing the relative action of angiopoietin-1(ANGPT1), which aids in regulating and inducing vascular normalization. This rebalancing contributes to vascular normalization, ensuring newly formed blood vessels exhibit a more organized structure and function more normally. The action of angiopoietin-1 helps consolidate and stabilize the newly formed vessels, making them a more effective transportation system.

However, it is also widely accepted that agents targeting VEGF/VEGFR may ultimately elevate hypoxia levels within tumors [ 4 ]. This hypoxic microenvironment triggers alternative pro-angiogenesis molecular pathways in tumor or stromal cells within the tumor microenvironment, including FGF-2, HGF, DLL4/Notch, CCL28, and others [ 19 , 21 , 22 , 31 ]. Furthermore, hypoxia prompts a cascade of biological responses, leading tumor cells to adapt their metabolic pathways to low-oxygen conditions. The dysregulated metabolism observed in rapidly proliferating tumor cells is a hallmark of malignancy [ 32 ], contributing to the activation of various metabolism-related genes, including several hypoxia-related transcription factors like hypoxia-inducible factor (HIF), nuclear factor kappa-B, CREB, AP-1, p53, Sp1/3, Egr-1, and CEBPB [ 33 ].

In this study, we identified CCL28 as another crucial molecular target for vascular normalization in lung adenocarcinoma. It was observed that CCL28 could be up-regulated under hypoxic conditions. However, the molecular mechanism underlying hypoxia-induced transcriptional activation of the CCL28 gene remains unclear. Our investigation revealed that the transcription factors CEBEPB could regulate the expression of CCL28. CCL28, belonging to the subfamily of small cytokine CC chemokines, binds to chemokine receptors CCR3 and CCR10 [ 19 ]. It has been reported that CCL28 exhibits chemotactic activity for various immune cells and plays a role in the physiology of extracutaneous epithelial tissues [ 34 ]. Several studies have delved into the functions of CCL28 in the tumor microenvironment [ 17 , 18 , 20 ]. Therefore, we hypothesized that CCL28 might play a pivotal role in modulating the tumor microenvironment in lung adenocarcinoma. Here, we noted an augmentation in pericyte accumulation within the tumor microenvironment associated with the overexpression of CCL28. However, the mechanisms of pericyte recruitment in lung adenocarcinoma remain unknown. Thus, we examined the expression of CCR3/CCR10 on various tumor stromal cells, revealing widespread expression of CCR3 on endothelial cells, cancer-associated fibroblasts, and pericytes in lung adenocarcinoma. Two chemotactic experiments substantiated the effects of CCL28 on recruiting pericytes through the receptor CCR3.

Abnormal differentiation of stromal cells is an essential characteristic of malignant tumors and is correlated with abnormal metabolism. Many tumor stromal cells were differentiated from anti-tumor type to pro-tumor type, such as tumor-associated macrophage, cancer-associated fibroblasts, tumor-associated neutrophils, etc. [ 35 , 36 , 37 ]. Reprogramming the metabolism and modulating tumor stromal cell differentiation is a promising cancer treatment strategy. Interestingly, the present study found that retinoic acid metabolism was activated in a series of stromal cells in lung cancer.

Importantly, chemotactic factors play a crucial role in cell migration, influencing not only the cell's movement [ 38 ] but also closely interacting with cellular metabolism. They regulate energy production, metabolic pathways, and antioxidant responses, ensuring that cells have sufficient energy and resources during migration. Chemotactic factors influence cellular metabolism by regulating intracellular signaling pathways, such as PI3K/AKT, MAPK, AMPK, and mTOR [ 39 , 40 ]. The activation or inhibition of these signaling pathways can modulate the activity of intracellular metabolic enzymes, affecting processes such as glucose metabolism, lipid synthesis, and amino acid utilization. In the present study, we found chemokine CCL28 could influence the cellular levels of retinoic acid by modulating its synthesis.

Retinoic acid (RA), also known as vitamin A acid, and its related analogs participate in regulating the gene networks involved in cell growth, differentiation, homeostasis, and apoptosis. RA is an active metabolite of retinol. Retinol (Vitamin A) is a fat-soluble essential micronutrient that plays a crucial role in embryonic development, organ formation, immune system function, and vision [ 26 ]. Retinol can be metabolized into retinal by the retinol dehydrogenases (RDHs), and retinal can also be metabolized into retinol by the dehydrogenase/reductase SDR family (DHRS) [ 41 ]. Retinal can be further irreversibly metabolized to RA by the retinaldehyde dehydrogenases (ALDHs). The synthesis of RA depends on the balance between RDHs and DHRS. The present study found that CCL28 could disturb the balance between RDH13 and DHRS11. After being treated with CCL28, RDH13 in pericytes was significantly upregulated in a dose-dependent manner. However, the effect curve of CCL28 on DHRS11 expression in pericytes is not linear and might be bell-shaped, like the effects of VEGFA on endothelial cells. In addition, there might be an alternative pathway to maintain the balance of DHRS11 and RDH13 in pericytes. RA binds to RARα, promoting the formation of a heterodimer with retinoid X receptor alpha (RARα/RXRα) in the cell nucleus. This heterodimer binds to retinoic acid response elements (RAREs) in the promoter region of target genes [ 42 ], including ANGPT1 .

The angiopoietin family, including ANGPT1, ANGPT2, ANGPT3, and ANGPT4, is crucial in vascular development and normalization. They regulate endothelial cells' survival, proliferation, and migration by interacting with Tie receptors, thus vital to vascular normalization. ANGPT1 can activate the TIE2 receptor on endothelial cells, maintaining endothelial cell stability, enhancing tight connections between endothelial cells, and reducing microvascular permeability through a series of signaling pathways [ 26 ]. These signaling pathways include tyrosine kinase-related protein DOKR (also known as DOK2), endothelial nitric oxide synthase (eNOS), SH2 domain-containing phosphatase (SHP2), growth factor receptor-binding protein 2 (GRB2), and PI3K-Akt [ 24 ]. Various regulatory mechanisms influence the expression of the ANGPT1 . These include the hypoxia-inducible factor-1α (HIF-1α) signaling pathway under low oxygen conditions, interaction with vascular endothelial growth factor (VEGF), involvement of anti-inflammatory factors and growth factors, cell–cell interactions, as well as the nuclear factor-κB (NF-κB) signaling pathway and hormonal regulation. In the present study, we identified another regulatory mechanism controlling the expression of ANGPT1.

Vitamin A and its metabolites, particularly retinoic acid, exert regulatory effects on the vascular system through various pathways [ 43 ]. These include maintaining endothelial cell function, regulating angiogenesis, inhibiting vascular smooth muscle cell proliferation, suppressing inflammatory responses, and promoting cell differentiation. This comprehensive regulatory process contributes to maintaining blood vessels' stable structure and function, playing a crucial role in embryonic development, tissue repair, angiogenesis, and vascular health during inflammation and diseases. Interestingly, vitamin A and its metabolites have been used as a differentiation modulator of malignant cells for cancer treatment. The present study proposed a new strategy to modulate the differentiation of cancer stromal by vitamin A and its metabolites.

Conclusions

In conclusion, we elucidated that a specific chemokine CCL28 induced after anti-angiogenesis therapy can alter tumor stromal cell metabolism and reshape the tumor microenvironment. In summary, we identified a mechanism through which CCL28 promotes vascular normalization via a CCR3-pericytes-RA-RXRα-ANGPT1-dependent pathway (Fig. 7 ). With an in-depth exploration of the interplay between chemokine and ANGPT1, our work may provide valuable insights into the regulatory network of vascular normalization, offering new avenues for modulating tumor microenvironment and overcoming resistance to anti-angiogenesis therapy in lung adenocarcinoma treatment.

A schematic diagram of tumor microenvironment modulation effects of CCL28

Availability of data and materials

The data that support the findings of this study are available from the corresponding author upon reasonable request.Competing interests.

The authors declare that they have no competing interests.

Abbreviations

Platelet-derived growth factor receptor β

Chondroitin sulfate proteoglycan 4

Human CC motif chemokine ligands 28

Human umbilical vein endothelial cell

Dehydrogenase/reductase 11

Retinol dehydrogenase 13

Bevacizumab

Zhang Y, Vaccarella S, Morgan E, Li M, Etxeberria J, Chokunonga E, et al. Global variations in lung cancer incidence by histological subtype in 2020: a population-based study. Lancet Oncol. 2023;24:1206–18.

Article   PubMed   Google Scholar  

Xia C, Dong X, Li H, Cao M, Sun D, He S, et al. Cancer statistics in China and United States, 2022: profiles, trends, and determinants. Chin Med J (Engl). 2022;135:584–90.

Reck M, Popat S, Grohé C, Corral J, Novello S, Gottfried M, et al. Anti-angiogenic agents for NSCLC following first-line immunotherapy: rationale, recent updates, and future perspectives. Lung Cancer. 2023;179:107173.

Article   PubMed   CAS   Google Scholar  

Augustin HG, Koh GY. Antiangiogenesis: vessel regression, vessel normalization, or both? Cancer Res. 2022;82:15–7.

Zhang N, Yin R, Zhou P, Liu X, Fan P, Qian L, et al. DLL1 orchestrates CD8(+) T cells to induce long-term vascular normalization and tumor regression. Proc Natl Acad Sci U S A. 2021;118(22):e2020057118.

Huang G, Chen L. Recombinant human endostatin improves anti-tumor efficacy of paclitaxel by normalizing tumor vasculature in Lewis lung carcinoma. J Cancer Res Clin Oncol. 2010;136:1201–11.

Shi S, Chen L, Huang G. Antiangiogenic therapy improves the antitumor effect of adoptive cell immunotherapy by normalizing tumor vasculature. Med Oncol. 2013;30:698.

Shi S, Tao L, Song H, Chen L, Huang G. Synergistic antitumor effect of combining metronomic chemotherapy with adoptive cell immunotherapy in nude mice. APMIS. 2014;122:380–91.

Tao L, Huang G, Shi S, Chen L. Bevacizumab improves the antitumor efficacy of adoptive cytokine-induced killer cells therapy in non-small cell lung cancer models. Med Oncol. 2014;31:777.

Jain RK. Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science. 2005;307:58–62.

Shi S, Wang R, Chen Y, Song H, Chen L, Huang G. Combining antiangiogenic therapy with adoptive cell immunotherapy exerts better antitumor effects in non-small cell lung cancer models. PLoS One. 2013;8:e65757.

Article   PubMed   PubMed Central   CAS   Google Scholar  

Ramjiawan RR, Griffioen AW, Duda DG. Anti-angiogenesis for cancer revisited: is there a role for combinations with immunotherapy? Angiogenesis. 2017;20:185–204.

Article   PubMed   PubMed Central   Google Scholar  

Hartmann DA, Berthiaume AA, Grant RI, Harrill SA, Koski T, Tieu T, et al. Brain capillary pericytes exert a substantial but slow influence on blood flow. Nat Neurosci. 2021;24:633–45.

Figueiredo AM, Villacampa P, Dieguez-Hurtado R, Jose Lozano J, Kobialka P, Cortazar AR, et al. Phosphoinositide 3-kinase-regulated pericyte maturation governs vascular remodeling. Circulation. 2020;142:688–704.

Zhang Y, Cedervall J, Hamidi A, Herre M, Viitaniemi K, D’Amico G, et al. Platelet-specific PDGFB ablation impairs tumor vessel integrity and promotes metastasis. Cancer Res. 2020;80:3345–58.

Hosaka K, Yang Y, Seki T, Nakamura M, Andersson P, Rouhi P, et al. Tumour PDGF-BB expression levels determine dual effects of anti-PDGF drugs on vascular remodelling and metastasis. Nat Commun. 2013;4:2129.

Facciabene A, Peng X, Hagemann IS, Balint K, Barchetti A, Wang LP, et al. Tumour hypoxia promotes tolerance and angiogenesis via CCL28 and T(reg) cells. Nature. 2011;475:226–30.

Ji L, Qian W, Gui L, Ji Z, Yin P, Lin GN, et al. Blockade of beta-catenin-induced CCL28 suppresses gastric cancer progression via inhibition of treg cell infiltration. Cancer Res. 2020;80:2004–16.

Huang G, Tao L, Shen S, Chen L. Hypoxia induced CCL28 promotes angiogenesis in lung adenocarcinoma by targeting CCR3 on endothelial cells. Sci Rep. 2016;6:27152.

Park J, Zhang X, Lee SK, Song N-Y, Son SH, Kim KR, et al. CCL28-induced RARβ expression inhibits oral squamous cell carcinoma bone invasion. J Clin Investig. 2019;129:5381–99.

Incio J, Ligibel JA, McManus DT, Suboj P, Jung K, Kawaguchi K, et al. Obesity promotes resistance to anti-VEGF therapy in breast cancer by up-regulating IL-6 and potentially FGF-2. Sci Transl Med. 2018;10:10.

Article   Google Scholar  

Cascone T, Xu L, Lin HY, Liu W, Tran HT, Liu Y, et al. The HGF/c-MET pathway is a driver and biomarker of VEGFR-inhibitor resistance and vascular remodeling in non-small cell lung cancer. Clin Cancer Res. 2017;23:5489–501.

Chen Z, Kim SJ, Essani AB, Volin MV, Vila OM, Swedler W, et al. Characterising the expression and function of CCL28 and its corresponding receptor, CCR10, in RA pathogenesis. Ann Rheum Dis. 2015;74:1898–906.

Huang H, Bhat A, Woodnutt G, Lappe R. Targeting the ANGPT-TIE2 pathway in malignancy. Nat Rev Cancer. 2010;10:575–85.

Teichert M, Milde L, Holm A, Stanicek L, Gengenbacher N, Savant S, et al. Pericyte-expressed Tie2 controls angiogenesis and vessel maturation. Nat Commun. 2017;8:16106.

Khan KA, Wu FT, Cruz-Munoz W, Kerbel RS. Ang2 inhibitors and Tie2 activators: potential therapeutics in perioperative treatment of early stage cancer. EMBO Mol Med. 2021;13:e08253.

Wang M, Herbst RS, Boshoff C. Toward personalized treatment approaches for non-small-cell lung cancer. Nat Med. 2021;27:1345–56.

Ellis LM, Hicklin DJ. VEGF-targeted therapy: mechanisms of anti-tumour activity. Nat Rev Cancer. 2008;8:579–91.

Geindreau M, Ghiringhelli F, Bruchard M. Vascular endothelial growth factor, a key modulator of the anti-tumor immune response. Int J Mol Sci. 2021;22:22.

Yang T, Xiao H, Liu X, Wang Z, Zhang Q, Wei N, et al. Vascular normalization: a new window opened for cancer therapies. Front Oncol. 2021;11:719836.

Li J-L, Sainson RCA, Oon CE, Turley H, Leek R, Sheldon H, et al. DLL4-Notch signaling mediates tumor resistance to anti-VEGF therapy in vivo. Can Res. 2011;71:6073–83.

Article   CAS   Google Scholar  

Hanahan D. Hallmarks of cancer: new dimensions. Cancer Discov. 2022;12:31–46.

Cummins EP, Taylor CT. Hypoxia-responsive transcription factors. Pflugers Arch. 2005;450:363–71.

Mohan T, Deng L, Wang B-Z. CCL28 chemokine: an anchoring point bridging innate and adaptive immunity. Int Immunopharmacol. 2017;51:165–70.

Zhang X, Ji L, Li MO. Control of tumor-associated macrophage responses by nutrient acquisition and metabolism. Immunity. 2023;56:14–31.

Zhang H, Yue X, Chen Z, Liu C, Wu W, Zhang N, et al. Define cancer-associated fibroblasts (CAFs) in the tumor microenvironment: new opportunities in cancer immunotherapy and advances in clinical trials. Mol Cancer. 2023;22:159.

Que H, Fu Q, Lan T, Tian X, Wei X. Tumor-associated neutrophils and neutrophil-targeted cancer therapies. Biochim Biophys Acta Rev Cancer. 2022;1877:188762.

Stoll G, Pol J, Soumelis V, Zitvogel L, Kroemer G. Impact of chemotactic factors and receptors on the cancer immune infiltrate: a bioinformatics study revealing homogeneity and heterogeneity among patient cohorts. Oncoimmunology. 2018;7:e1484980.

Xu H, Lin S, Zhou Z, Li D, Zhang X, Yu M, et al. New genetic and epigenetic insights into the chemokine system: the latest discoveries aiding progression toward precision medicine. Cell Mol Immunol. 2023;20:739–76.

Lazennec G, Rajarathnam K, Richmond A. CXCR2 chemokine receptor - a master regulator in cancer and physiology. Trends Mol Med. 2024;30:37–55.

Adams MK, Belyaeva OV, Wu L, Kedishvili NY. The retinaldehyde reductase activity of DHRS3 is reciprocally activated by retinol dehydrogenase 10 to control retinoid homeostasis. J Biol Chem. 2014;289:14868–80.

Ghyselinck NB, Duester G. Retinoic acid signaling pathways. Development. 2019;146:146.

DiKun KM, Gudas LJ. Vitamin A and retinoid signaling in the kidneys. Pharmacol Ther. 2023;248:108481.

Download references

Acknowledgements

Not applicable.

This work was supported by the National Natural Science Foundation of China (Grant No. 82273326), Jiangsu Provincial Youth Medical Key Talents Project (Grant No. QNRC2016887), the Bethune Charitable Foundation (Grant No. KY202301-17) and Postgraduate Research & Practice Innovation Program of Jiangsu Province (Grant No. KYCX22_0176).

Author information

Authors and affiliations.

The State Key Laboratory of Pharmaceutical Biotechnology, Chemistry and Biomedicine Innovation Center (ChemBIC), Division of Immunology, Medical School, Nanjing University, Nanjing, 210093, China

Ying Chen, Zhiyong Zhang, Pengfei Li, Yuxi Chen & Tingting Wang

Jiangsu Key Laboratory of Molecular Medicine, Division of Immunology, Medical School, Nanjing University, Nanjing, 210093, China

Medical Schoolof, Nanjing University, Nanjing, Jiangsu, 210093, China

Ying Chen, Zhiyong Zhang, Fan Pan, Pengfei Li, Weiping Yao, Yuxi Chen & Tingting Wang

Department of Respiratory Critical Care Medicine, Affiliated Hospital of Medical School, Nanjing Drum Tower Hospital, Nanjing University, Nanjing, Jiangsu, 210008, China

Department of Cardio-Thoracic Surgery, Affiliated Hospital of Medical School, Jinling Hospital, Nanjing University, Nanjing, Jiangsu, 210008, China

Department of Medical Oncology, Affiliated Hospital of Medical School, Jinling Hospital, Nanjing University, Nanjing, Jiangsu, 210008, China

Fan Pan, Weiping Yao & Guichun Huang

Department of Oncology, Medical School, Zhongda Hospital, Southeast University, Nanjing, 210009, China

Guichun Huang

You can also search for this author in PubMed   Google Scholar

Contributions

TTW, YL and GCH: Provided funding support, conceptualization, methodological guidance, and manuscript revision for this study.. YC: conceived of the study, performed the experiments and wrote the paper. WPY: participated in animal experiments. ZYZ and FP: performed the statistical analysis. PFL, YXC and LX: Participated in experiment design and sample collection. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Tingting Wang , Yan Li or Guichun Huang .

Ethics declarations

Ethics approval and consent to participate.

All animal experiments were carried out following the institutional guidelines and approved by the Ethical Review Committee of Comparative Medicine, Jinling Hospital, Nanjing, China. Written informed consent was obtained from all patients before collecting the samples. All the methods were carried out following the institutional guidelines and approved by the Ethical Review Committee of Jinling Hospital, Nanjing, China.

Consent for publication

All authors have consented for publication.

Competing interests

Additional information, publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Supplementary material 1., supplementary material 2., supplementary material 3., supplementary material 4., rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Chen, Y., Zhang, Z., Pan, F. et al. Pericytes recruited by CCL28 promote vascular normalization after anti-angiogenesis therapy through RA/RXRA/ANGPT1 pathway in lung adenocarcinoma. J Exp Clin Cancer Res 43 , 210 (2024). https://doi.org/10.1186/s13046-024-03135-3

Download citation

Received : 21 March 2024

Accepted : 22 July 2024

Published : 29 July 2024

DOI : https://doi.org/10.1186/s13046-024-03135-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Anti-angiogenesis therapy
• Vascular normalization
• Lung adenocarcinoma
• Retinoic acid

Journal of Experimental & Clinical Cancer Research

ISSN: 1756-9966

• Submission enquiries: Access here and click Contact Us
• General enquiries: [email protected]

• Alzheimer's disease & dementia
• Arthritis & Rheumatism
• Attention deficit disorders
• Autism spectrum disorders
• Biomedical technology
• Diseases, Conditions, Syndromes
• Endocrinology & Metabolism
• Gastroenterology
• Gerontology & Geriatrics
• Health informatics
• Inflammatory disorders
• Medical economics
• Medical research
• Medications
• Neuroscience
• Obstetrics & gynaecology
• Oncology & Cancer
• Ophthalmology
• Overweight & Obesity
• Parkinson's & Movement disorders
• Psychology & Psychiatry
• Radiology & Imaging
• Sleep disorders
• Sports medicine & Kinesiology
• Vaccination
• Breast cancer
• Cardiovascular disease
• Chronic obstructive pulmonary disease
• Colon cancer
• Coronary artery disease
• Heart attack
• Heart disease
• High blood pressure
• Kidney disease
• Lung cancer
• Multiple sclerosis
• Myocardial infarction
• Ovarian cancer
• Post traumatic stress disorder
• Rheumatoid arthritis
• Schizophrenia
• Skin cancer
• Type 2 diabetes
• Full List »

share this!

July 30, 2024

This article has been reviewed according to Science X's editorial process and policies . Editors have highlighted the following attributes while ensuring the content's credibility:

fact-checked

peer-reviewed publication

Recent study reveals key immune cells as critical factors in lung cancer prognosis

by Terasaki Institute for Biomedical Innovation

An extensive analytical study performed at the Terasaki Institute and published in Frontiers in Immunology highlights the crucial role of tissue-resident memory T cells and how they influence the immune environment of patients with non-small cell lung cancer and their overall prognosis.

Non-small cell lung cancer accounts for ~85% of lung tumors and is a leading cause of death in adults. Tissue-resident memory T cells, a specialized subset of immune cells residing in peripheral tissues , have been suspected of impacting cancer progression . However, it's still not fully understood how tissue-resident memory T cells affect the tumor immune microenvironment and tumor progression in various non-small cell lung cancer patient populations.

In this comprehensive study, multiple independent datasets from lung cancer patient samples were analyzed. In addition, a machine learning model was developed and validated to predict patient survival, refining an 18-gene risk score that effectively categorizes lung cancer patients into low-risk and high-risk groups. In cancer research , the 18-gene risk score is used to predict disease progression or recurrence chances, which helps create personalized treatment plans. The scores are usually divided into low and high risk, with specific thresholds setting these categories.

In this study, patients with high-risk scores exhibited significantly lower overall survival rates than their low-risk counterparts. Distinct Tissue Resident Memory T cell biomarkers were identified that correlate positively with other immune cells within the tumor environment. Moreover, these biomarkers were strongly associated with immune checkpoint and stimulatory genes, directly influencing patient prognosis.

"The study's findings highlight the critical impact of Tissue Resident Memory T cell abundance on immune responses and patient outcomes in lung cancer," said Dr. Xiling Shen, Chief Scientific Officer at Terasaki Institute for Biomedical Innovation. "Our findings not only validate these cells as a prognostic marker but also underscore their potential in guiding personalized treatment strategies, particularly in immunotherapy."

This research, independently validated by the Cancer Genome Atlas Program and multiple lung cancer patient datasets, provides a deeper understanding of the complex interplay between Tissue Resident Memory T cells and the tumor. It represents a significant step towards advancing precision medicine in lung cancer treatment.

Explore further

Feedback to editors

Strengthening global regulatory capacity for equitable access to vaccines in public health emergencies

Aug 3, 2024

Can targeted therapies be applied to patients with Alzheimer's?

Controlling thickness in fruit fly hearts reveals new pathway for heart disease

Aug 2, 2024

Researchers discover source of deadly fungal infections in bone marrow transplant patients

Scientists discover a population of macrophages that participate in alveolar regeneration

White matter may aid recovery from spinal cord injuries: Study

Scientists discover mast cells trap and use living neutrophils during allergic reactions

Brain activity associated with specific words is mirrored between speaker and listener during a conversation, data show

Popularity of Ozempic and Wegovy among privately insured patients may worsen disparities, suggests study

Researchers highlight the genetic complexity of schizophrenia

Related stories.

Gene signatures from tissue-resident T cells as a predictive tool for melanoma patients

May 2, 2024

Study uncovers key immune cells for combating aggressive Merkel cell carcinoma

Jul 26, 2024

Study finds potential biomarkers for lung cancer immunotherapy resistance

Jul 12, 2024

Using machine learning to identify patients with cancer that would benefit from immunotherapy

Apr 16, 2024

Genetically engineered dendritic cells enhance the power of immunotherapy against lung cancer

Mar 26, 2024

Breaking bad blood: How rogue neutrophils help lung cancer spread

May 15, 2024

Recommended for you

Researchers develop promising therapy treatment that can kill glioblastoma cells in newly-discovered brain pathway

Researchers devise novel solution to preventing relapse after CAR T-cell therapy

Circular RNAs—the new frontier in cancer research

Drug developed for pancreatic cancer shows promise against most aggressive form of medulloblastoma

Let us know if there is a problem with our content.

Use this form if you have come across a typo, inaccuracy or would like to send an edit request for the content on this page. For general inquiries, please use our contact form . For general feedback, use the public comments section below (please adhere to guidelines ).

Please select the most appropriate category to facilitate processing of your request

Thank you for taking time to provide your feedback to the editors.

Your feedback is important to us. However, we do not guarantee individual replies due to the high volume of messages.

E-mail the story

Your email address is used only to let the recipient know who sent the email. Neither your address nor the recipient's address will be used for any other purpose. The information you enter will appear in your e-mail message and is not retained by Medical Xpress in any form.

Newsletter sign up

Get weekly and/or daily updates delivered to your inbox. You can unsubscribe at any time and we'll never share your details to third parties.

More information Privacy policy

Donate and enjoy an ad-free experience

We keep our content available to everyone. Consider supporting Science X's mission by getting a premium account.

E-mail newsletter

Information

• Author Services

Initiatives

You are accessing a machine-readable page. In order to be human-readable, please install an RSS reader.

All articles published by MDPI are made immediately available worldwide under an open access license. No special permission is required to reuse all or part of the article published by MDPI, including figures and tables. For articles published under an open access Creative Common CC BY license, any part of the article may be reused without permission provided that the original article is clearly cited. For more information, please refer to https://www.mdpi.com/openaccess .

Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for future research directions and describes possible research applications.

Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive positive feedback from the reviewers.

Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world. Editors select a small number of articles recently published in the journal that they believe will be particularly interesting to readers, or important in the respective research area. The aim is to provide a snapshot of some of the most exciting work published in the various research areas of the journal.

Original Submission Date Received: .

• Active Journals
• Find a Journal
• Proceedings Series
• For Authors
• For Reviewers
• For Editors
• For Librarians
• For Publishers
• For Societies
• For Conference Organizers
• Open Access Policy
• Institutional Open Access Program
• Special Issues Guidelines
• Editorial Process
• Research and Publication Ethics
• Article Processing Charges
• Testimonials
• Preprints.org
• SciProfiles
• Encyclopedia

Article Menu

• Subscribe SciFeed
• Recommended Articles
• Google Scholar
• on Google Scholar
• Table of Contents

Find support for a specific problem in the support section of our website.

Please let us know what you think of our products and services.

Visit our dedicated information section to learn more about MDPI.

JSmol Viewer

Real-world treatment patterns, health outcomes, and healthcare resource use in advanced common egfr-positive non-small cell lung cancer patients treated with osimertinib in alberta.

1. Introduction

2. materials and methods, 2.1. study population and data sources, 2.2. statistical analyses, 3.1. patient characteristics, 3.2. overall cohort: os and ttntd, 3.3. post-osimertinib cohort: os and ttntd, 3.4. healthcare resource utilization, 3.5. treatment patterns, 4. discussion, 4.1. overall cohort, 4.2. post-osimertinib cohort, 4.3. treatment patterns and healthcare resource utilization, 4.4. strengths and limitations, 5. conclusions, author contributions, institutional review board statement, informed consent statement, data availability statement, conflicts of interest.

• Canadian Cancer Statistics 2023. Available online: https://cdn.cancer.ca/-/media/files/research/cancer-statistics/2023-statistics/2023_PDF_EN.pdf (accessed on 7 March 2024).
• Travis, W. Pathology & Genetics Tumours of the Lung, Pleura, Thymus and Heart ; World Health Organization: Geneva, Switzerland, 2004. [ Google Scholar ]
• da Cunha Santos, G.; Shepherd, F.A.; Tsao, M.S. EGFR mutations and lung cancer. Annu. Rev. Pathol. 2011 , 6 , 49–69. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Graham, R.P.; Treece, A.L.; Lindeman, N.I.; Vasalos, P.; Shan, M.; Jennings, L.J.; Rimm, D.L. Worldwide Frequency of Commonly Detected EGFR Mutations. Arch. Pathol. Lab. Med. 2018 , 142 , 163–167. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Melosky, B.; Banerji, S.; Blais, N.; Chu, Q.; Juergens, R.; Leighl, N.B.; Liu, G.; Cheema, P. Canadian consensus: A new systemic treatment algorithm for advanced EGFR-mutated non-small-cell lung cancer. Curr. Oncol. 2020 , 27 , e146–e155. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Woodburn, J.R. The epidermal growth factor receptor and its inhibition in cancer therapy. Pharmacol. Ther. 1999 , 82 , 241–250. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Riess, J.W.; Gandara, D.R.; Frampton, G.M.; Madison, R.; Peled, N.; Bufill, J.A.; Dy, G.K.; Ou, S.I.; Stephens, P.J.; McPherson, J.D.; et al. Diverse EGFR Exon 20 Insertions and Co-Occurring Molecular Alterations Identified by Comprehensive Genomic Profiling of NSCLC. J. Thorac. Oncol. 2018 , 13 , 1560–1568. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Russo, A.; Franchina, T.; Ricciardi, G.; Battaglia, A.; Picciotto, M.; Adamo, V. Heterogeneous Responses to Epidermal Growth Factor Receptor (EGFR) Tyrosine Kinase Inhibitors (TKIs) in Patients with Uncommon EGFR Mutations: New Insights and Future Perspectives in this Complex Clinical Scenario. Int. J. Mol. Sci. 2019 , 20 , 1431. [ Google Scholar ] [ CrossRef ]
• Passaro, A.; Mok, T.; Peters, S.; Popat, S.; Ahn, M.J.; de Marinis, F. Recent Advances on the Role of EGFR Tyrosine Kinase Inhibitors in the Management of NSCLC With Uncommon, Non Exon 20 Insertions, EGFR Mutations. J. Thorac. Oncol. 2021 , 16 , 764–773. [ Google Scholar ] [ CrossRef ]
• Xu, J.; Jin, B.; Chu, T.; Dong, X.; Yang, H.; Zhang, Y.; Wu, D.; Lou, Y.; Zhang, X.; Wang, H.; et al. EGFR tyrosine kinase inhibitor (TKI) in patients with advanced non-small cell lung cancer (NSCLC) harboring uncommon EGFR mutations: A real-world study in China. Lung Cancer 2016 , 96 , 87–92. [ Google Scholar ] [ CrossRef ]
• Planchard, D.; Jänne, P.A.; Cheng, Y.; Yang, J.C.; Yanagitani, N.; Kim, S.W.; Sugawara, S.; Yu, Y.; Fan, Y.; Geater, S.L.; et al. Osimertinib with or without Chemotherapy in EGFR-Mutated Advanced NSCLC. N. Engl. J. Med. 2023 , 389 , 1935–1948. [ Google Scholar ] [ CrossRef ]
• Ramalingam, S.S.; Vansteenkiste, J.; Planchard, D.; Cho, B.C.; Gray, J.E.; Ohe, Y.; Zhou, C.; Reungwetwattana, T.; Cheng, Y.; Chewaskulyong, B.; et al. Overall Survival with Osimertinib in Untreated, EGFR-Mutated Advanced NSCLC. N. Engl. J. Med. 2020 , 382 , 41–50. [ Google Scholar ] [ CrossRef ]
• Mok, T.S.; Wu, Y.L.; Ahn, M.J.; Garassino, M.C.; Kim, H.R.; Ramalingam, S.S.; Shepherd, F.A.; He, Y.; Akamatsu, H.; Theelen, W.S.; et al. Osimertinib or Platinum-Pemetrexed in EGFR T790M-Positive Lung Cancer. N. Engl. J. Med. 2017 , 376 , 629–640. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Vyse, S.; Huang, P.H. Targeting EGFR exon 20 insertion mutations in non-small cell lung cancer. Signal Transduct. Target. Ther. 2019 , 4 , 5. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Passaro, A.; Wang, J.; Wang, Y.; Lee, S.H.; Melosky, B.; Shih, J.Y.; Wang, J.; Azuma, K.; Juan-Vidal, O.; Cobo, M.; et al. Amivantamab plus chemotherapy with and without lazertinib in EGFR-mutant advanced NSCLC after disease progression on osimertinib: Primary results from the phase III MARIPOSA-2 study. Ann. Oncol. 2024 , 35 , 77–90. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Jaiyesimi, I.A.; Leighl, N.B.; Ismaila, N.; Alluri, K.; Florez, N.; Gadgeel, S.; Masters, G.; Schenk, E.L.; Schneider, B.J.; Sequist, L.; et al. Therapy for Stage IV Non–Small Cell Lung Cancer With Driver Alterations: ASCO Living Guideline, Version 2023.3. J. Clin. Oncol. 2024 , 42 , e1–e22. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Koulouris, A.; Tsagkaris, C.; Corriero, A.C.; Metro, G.; Mountzios, G. Resistance to TKIs in EGFR-Mutated Non-Small Cell Lung Cancer: From Mechanisms to New Therapeutic Strategies. Cancers 2022 , 14 , 3337. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Hendriks, L.E.; Kerr, K.M.; Menis, J.; Mok, T.S.; Nestle, U.; Passaro, A.; Peters, S.; Planchard, D.; Smit, E.F.; Solomon, B.J.; et al. Non-oncogene-addicted metastatic non-small-cell lung cancer: ESMO Clinical Practice Guideline for diagnosis, treatment and follow-up. Ann. Oncol. 2023 , 34 , 358–376. [ Google Scholar ] [ CrossRef ]
• Fu, K.; Xie, F.; Wang, F.; Fu, L. Therapeutic strategies for EGFR-mutated non-small cell lung cancer patients with osimertinib resistance. J. Hematol. Oncol. 2022 , 15 , 173. [ Google Scholar ] [ CrossRef ]
• Rangachari, D.; Yamaguchi, N.; VanderLaan, P.A.; Folch, E.; Mahadevan, A.; Floyd, S.R.; Uhlmann, E.J.; Wong, E.T.; Dahlberg, S.E.; Huberman, M.S.; et al. Brain metastases in patients with EGFR-mutated or ALK-rearranged non-small-cell lung cancers. Lung Cancer 2015 , 88 , 108–111. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Lin, Y.T.; Tsai, T.H.; Wu, S.G.; Liu, Y.N.; Yu, C.J.; Shih, J.Y. Complex EGFR mutations with secondary T790M mutation confer shorter osimertinib progression-free survival and overall survival in advanced non-small cell lung cancer. Lung Cancer 2020 , 145 , 1–9. [ Google Scholar ] [ CrossRef ]
• Chmielecki, J.; Gray, J.E.; Cheng, Y.; Ohe, Y.; Imamura, F.; Cho, B.C.; Lin, M.C.; Majem, M.; Shah, R.; Rukazenkov, Y.; et al. Candidate mechanisms of acquired resistance to first-line osimertinib in EGFR-mutated advanced non-small cell lung cancer. Nat. Commun. 2023 , 14 , 1070. [ Google Scholar ] [ CrossRef ]
• Cho, B.C.; Felip, E.; Hayashi, H.; Thomas, M.; Lu, S.; Besse, B.; Sun, T.; Martinez, M.; Sethi, S.N.; Shreeve, S.M.; et al. MARIPOSA: Phase 3 study of first-line amivantamab + lazertinib versus osimertinib in EGFR-mutant non-small-cell lung cancer. Future Oncol. 2022 , 18 , 639–647. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Farah, E.; Carbonell, C.; Boyne, D.J.; Brenner, D.R.; Henning, J.-W.; Moldaver, D.; Shokar, S.; Cheung, W.Y. Treatment Patterns and Health Outcomes among Patients with HER2 IHC0/-Low Metastatic or Recurrent Breast Cancer. Cancers 2024 , 16 , 518. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Martins, D.; O’Sullivan, D.E.; Boyne, D.J.; Cheung, W.Y.; Allonby, O.; Habash, M.; Brenner, D.R.; Riemer, J.; McGee, J. Understanding Characteristics, Treatment Patterns, and Clinical Outcomes for Individuals with Advanced or Recurrent Endometrial Cancer in Alberta, Canada: A Retrospective, Population-Based Cohort Study. Curr. Oncol. 2023 , 30 , 2277–2289. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• O’Sullivan, D.E.; Cheung, W.Y.; Syed, I.A.; Moldaver, D.; Shanahan, M.K.; Bebb, D.G.; Sit, C.; Brenner, D.R.; Boyne, D.J. Real-World Treatment Patterns, Clinical Outcomes, and Health Care Resource Utilization in Extensive-Stage Small Cell Lung Cancer in Canada. Curr. Oncol. 2021 , 28 , 3091–3103. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Karim, S.; Lowther, J.; Gyulay, G.; O’Sullivan, D.; Wallis, C.J.D.; Yip, S.M.; Brenner, D.R.; Boyne, D.J.; Cheung, W.Y. A Real-World Evidence Study Using Alberta-Population-Based Data to Describe Treatment Patterns for Metastatic Castration-Sensitive Prostate Cancer Patients (AWARENESS). Curr. Oncol. 2023 , 30 , 8149–8158. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• O’Sullivan, D.E.; Boyne, D.J.; Gogna, P.; Brenner, D.R.; Cheung, W.Y. Understanding Real-World Treatment Patterns and Clinical Outcomes among Metastatic Melanoma Patients in Alberta, Canada. Curr. Oncol. 2023 , 30 , 4166–4176. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Yan, J.T.; Boyne, D.J.; Lo, E.; Farah, E.; O’Sullivan, D.E.; Cheung, W.Y. Real-world impact of patient-reported outcome measurement on overall survival, healthcare use and treatment discontinuation in cancer patients. J. Comp. Eff. Res. 2023 , 12 , e230061. [ Google Scholar ] [ CrossRef ] [ PubMed ]
• Boyne, D.J.; Ngan, E.; Carbonell, C.; Wani, R.J.; Cirone Morris, C.; Martinez, D.J.; Cheung, W.Y. Real-World Study to Assess Patterns of Treatment Practices and Clinical Outcomes in Metastatic Colorectal Cancer Patients with RAS Wild-Type Left-Sided Tumours in Canada. Curr. Oncol. 2023 , 30 , 8220–8232. [ Google Scholar ] [ CrossRef ]
• Gandara, D.R.; Li, T.; Lara, P.N.; Kelly, K.; Riess, J.W.; Redman, M.W.; Mack, P.C. Acquired resistance to targeted therapies against oncogene-driven non-small-cell lung cancer: Approach to subtyping progressive disease and clinical implications. Clin. Lung Cancer 2014 , 15 , 1–6. [ Google Scholar ] [ CrossRef ]

Click here to enlarge figure

Share and Cite

Cheung, W.Y.; Carbonell, C.; Navani, V.; Sangha, R.S.; Ewara, E.M.; Elia-Pacitti, J.; Iczkovitz, S.; Jarada, T.N.; Warkentin, M.T. Real-World Treatment Patterns, Health Outcomes, and Healthcare Resource Use in Advanced Common EGFR-Positive Non-Small Cell Lung Cancer Patients Treated with Osimertinib in Alberta. Curr. Oncol. 2024 , 31 , 4382-4396. https://doi.org/10.3390/curroncol31080327

Cheung WY, Carbonell C, Navani V, Sangha RS, Ewara EM, Elia-Pacitti J, Iczkovitz S, Jarada TN, Warkentin MT. Real-World Treatment Patterns, Health Outcomes, and Healthcare Resource Use in Advanced Common EGFR-Positive Non-Small Cell Lung Cancer Patients Treated with Osimertinib in Alberta. Current Oncology . 2024; 31(8):4382-4396. https://doi.org/10.3390/curroncol31080327

Cheung, Winson Y., Chantelle Carbonell, Vishal Navani, Randeep S. Sangha, Emmanuel M. Ewara, Julia Elia-Pacitti, Sandra Iczkovitz, Tamer N. Jarada, and Matthew T. Warkentin. 2024. "Real-World Treatment Patterns, Health Outcomes, and Healthcare Resource Use in Advanced Common EGFR-Positive Non-Small Cell Lung Cancer Patients Treated with Osimertinib in Alberta" Current Oncology 31, no. 8: 4382-4396. https://doi.org/10.3390/curroncol31080327

Article Metrics

Article access statistics, further information, mdpi initiatives, follow mdpi.

Subscribe to receive issue release notifications and newsletters from MDPI journals

Warning: The NCBI web site requires JavaScript to function. more...

An official website of the United States government

The .gov means it's official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings
• Browse Titles

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Cancer Control: Knowledge Into Action: WHO Guide for Effective Programmes: Module 4: Diagnosis and Treatment. Geneva: World Health Organization; 2008.

Cancer Control: Knowledge Into Action: WHO Guide for Effective Programmes: Module 4: Diagnosis and Treatment.

A plan for the diagnosis and treatment of cancer is a key component of any overall cancer control plan. Its main goal is to cure cancer patients or prolong their life considerably, ensuring a good quality of life. In order for a diagnosis and treatment programme to be effective, it must never be developed in isolation. It needs to be linked to an early detection programme so that cases are detected at an early stage, when treatment is more effective and there is a greater chance of cure. It also needs to be integrated with a palliative care programme, so that patients with advanced cancers, who can no longer benefit from treatment, will get adequate relief from their physical, psychosocial and spiritual suffering. Furthermore, programmes should include a awareness-raising component, to educate patients, family and community members about the cancer risk factors and the need for taking preventive measures to avoid developing cancer.

Where resources are limited, diagnosis and treatment services should initially target all patients presenting with curable cancers, such as breast, cervical and oral cancers that can be detected early. They could also include childhood acute lymphatic leukaemia, which has a high potential for cure although it cannot be detected early. Above all, services need to be provided in an equitable and sustainable manner. As and when more resources become available, the programme can be extended to include other curable cancers as well as cancers for which treatment can prolong survival considerably.

This module on diagnosis and treatment is intended to evolve in response to national needs and experience. WHO welcomes input from countries wishing to share their successes in diagnosis and treatment. WHO also welcomes requests from countries for information relevant to their specific needs. Evidence on the barriers to diagnosis and treatment in country contexts – and the lessons learned in overcoming them – would be especially welcome (contact at http://www.who.int/cancer ).

All rights reserved. Publications of the World Health Organization can be obtained from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: tni.ohw@sredrokoob ). Requests for permission to reproduce or translate WHO publications – whether for sale or for noncommercial distribution – should be addressed to WHO Press, at the above address (fax: +41 22 791 4806; e-mail: tni.ohw@snoissimrep ).

• Cite this Page Cancer Control: Knowledge Into Action: WHO Guide for Effective Programmes: Module 4: Diagnosis and Treatment. Geneva: World Health Organization; 2008. CONCLUSION.
• PDF version of this title (3.4M)

Other titles in this collection

• WHO Guidelines Approved by the Guidelines Review Committee

Recent Activity

• CONCLUSION - Cancer Control: Knowledge Into Action CONCLUSION - Cancer Control: Knowledge Into Action

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

Connect with NLM

National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894

Web Policies FOIA HHS Vulnerability Disclosure

Help Accessibility Careers