|
|
||||||||
Guest Access | Sign In via User Name/Password |
|||||||||

* From the Division of Medical Oncology (Dr. Strauss), Brown Medical School and Rhode Island Hospital, Providence, RI; Center for Thoracic Surgery (Dr. Dominioni), University of Insubria, Varese, Italy; the Division of Pulmonary Medicine (Dr. Jett), Mayo Clinic, Rochester, MN; Imaging Science and Information Systems Research Center (Dr. Freedman), Department of Oncology, Georgetown University, Washington, DC; and the Department of Thoracic Surgery (Dr. Grannis), City of Hope National Medical Center, Duarte, CA.
Cosponsors of the Como International Conference include Facoltà di Medicina e Chirurgia, University of Insubria, Varese, Italy; the Department of Medicine, Rhode Island Hospital and Brown Medical School, Providence, RI; and the American Cancer Society, Atlanta, GA. This position statement, however, has not received the endorsement of the American Cancer Society.
Correspondence to: Gary M. Strauss, MD, MPH, Division of Medical Oncology, Rhode Island Hospital, Brown Medical School, 593 Eddy St, Providence, RI 02903; e-mail: gstrauss{at}lifespan.org
| Abstract |
|---|
|
|
|---|
Purpose: The primary objective of the Como Conference was to consider whether there is sufficient scientific evidence to advise screening for lung cancer among asymptomatic individuals outside the context of a clinical trial. Methodological issues that are relevant to the proper interpretation of early detection trials were carefully considered. Advantages and problems associated with technological advances in CT scans and digital chest radiographs (CXRs) were fully explored. Economic issues relevant to screening were also considered.
Recommendations: It is recommended that physicians assume responsibility for informing high-risk individuals regarding options for screening for lung cancer. Targeted high-risk individuals include middle-aged or elderly men and women who are current or former cigarette smokers of > 20 to 30 pack-years without serious medical comorbidities. It is recommended that such persons be informed that symptomatic lung cancer is usually advanced and incurable, while surgery for early lung cancer offers a far better chance of cure. They should also be informed about advances in imaging technology, as they relate to CT scans and CXRs.
Conclusions: Whenever possible, high-risk individuals should be encouraged to enroll in ongoing trials. For subjects who, though eligible, do not have access to such trials, a process of shared decision-making between physicians and at-risk individuals is strongly recommended. After discussion of the existing state of knowledge, high-risk individuals should be made aware that it is reasonable for them to choose to undergo testing for lung cancer.
Key Words: chest roentgenogram cigarette smoking computed tomography consensus statement early detection lung cancer mortality overdiagnosis bias screening survival
| Introduction |
|---|
|
|
|---|
Currently, lung cancer is the most common cancer in the world. This is true both with regard to incidence (1.2 million new cases annually representing 12.3% of all cancers) and mortality (1.1 million annual deaths or 17.8% of total cancer mortality).1 Despite the global burden of disease, the absence of unequivocal evidence from previous randomized controlled trials23456 has prevented screening from being recommended by any public policy organization. Instead, strategies directed toward reducing lung cancer mortality have focused almost exclusively on tobacco control. Diagnostic procedures to detect lung cancer have traditionally been recommended only when symptoms develop. However, it has long been recognized that symptomatic lung cancer is usually advanced-stage disease. Moreover, advanced-stage lung cancer is almost always fatal.
| The 1998 Varese Conference |
|---|
|
|
|---|
At the completion of the Varese Conference, a consensus statement was generated.89 With regard to prevention, the statement reaffirmed the fact that "because cigarette smoking is the vastly predominant cause, lung cancer is almost entirely preventable." However, it also pointed out that "even after quitting, long term smokers remain at high-risk for prolonged periods."
The consensus statement also recognized the more favorable prognosis of lung cancer treated at an early stage. For example, it asserted that among those with lung cancer, "outcome is dramatically better when the disease is detected at an early stage and surgically treated. Unfortunately, at this time, the majority of lung cancers are diagnosed when the disease is overtly symptomatic, and in an advanced stage when prognosis is extremely poor."
While the consensus statement recognized the potential of early detection to improve outcome in lung cancer, a consensus was not reached that screening for lung cancer should be offered outside the context of an experimental trial. Because of ambiguities in the data, the consensus statement concluded that existing studies provide us "with an imperfect basis for health policy."
The Varese Conference helped to revitalize interest in screening for lung cancer.10 Based on recommendations from the Varese Conference, the American Cancer Society (ACS) modified its narrative about testing for early lung cancer detection.111213 While not recommending screening, the ACS endorsed the practice that individuals who are at high-risk for lung cancer should be informed about their risk, and that those who seek testing for early lung cancer detection should be informed about options for testing for early detection, and about the current state of knowledge on its risks, benefits, and harms, so that they could make an informed decision. The ACS also stressed the importance of quality control and appropriate follow-up.
The Varese Conference was an important catalytic event that helped to reawaken interest in lung cancer screening. It helped to establish that the existing scientific evidence about screening was not a sound basis for the prevailing view that lung cancer screening was ineffective. Nonetheless, although a substantial body of evidence suggests that testing for early lung cancer has the potential to reduce lung cancer deaths,1014 screening for lung cancer continues to be viewed in the mainstream as experimental. Accordingly, testing for the early detection of lung cancer is not widely practiced outside the context of a clinical trial.
Five years after the Varese Conference, the Como Conference was held to consider whether there existed a sufficient scientific basis to go beyond the 1998 Varese consensus statement. The particular emphasis was on considering what might be offered to those high-risk individuals who are not participating in an existing clinical trial.
| Como Conference: Position Statement |
|---|
|
|
|---|
Both clinicians and individuals who are at high risk for lung cancer presently do not have specific or consistent guidance about the balance of benefit and harm that is associated with testing for the early detection of lung cancer. While the results of early randomized trials were not definitive due to known methodological limitations, and observational studies also have known limitations, a better prognosis and more successful treatment have been consistently evident when lung cancer is diagnosed and treated at an early stage. Moreover, new imaging technologies have been developed over the last 10 years that can detect smaller cancers. For these reasons, it is reasonable that individuals who are at high risk for lung cancer should be informed about their risk, and what is known and unknown about the potential benefits, limitations, and harms associated with testing for early lung cancer detection.89111213
| Como Conference Recommendations |
|---|
|
|
|---|
Medical and public health organizations should work together to develop educational materials that facilitate shared decision making between at-risk individuals and their medical providers. Such materials should inform doctors and patients about the risk of lung cancer in current and former smokers, and, particularly, should dispel the misconception that stopping smoking begins a process of reducing risk to the level of a never-smoker.
Any testing for lung cancer, if it is performed, should take place in settings with experience in the interpretation of imaging procedures for the detection of small lung cancers, and there should be ready access to multidisciplinary teams who work in a coordinated manner for further evaluation and follow-up.12 Efforts should be made to minimize anxiety and the performance of unnecessary invasive procedures.
| Appendix |
|---|
|
|
|---|
The epidemiologic and clinical rationale for screening for lung cancer is strong. Lung cancer is the most common and most deadly malignancy in the world.1 When diagnosed on the basis of signs or symptoms, lung cancer is usually incurable and lethal. Indeed, in the United States 60% of lung cancer patients die within 1 year of diagnosis.18 Because screening for lung cancer is not widely practiced, the vast majority of patients are symptomatic at diagnosis. Accordingly, the 5-year survival rate in persons with lung cancer is 11% worldwide and 14% in the United States.119 Among patients with lung cancer, 85% will actually die from their disease.20
In dramatic contrast, lung cancer is highly curable when complete resection is performed for stage I non-small cell lung cancer. Among patients with stage I disease, cure rates increase with decreasing tumor size. For example, cure rates are 69% when the tumor size is < 1.5 cm compared to 43% when the tumor size is > 4.5 cm.21
The question of whether evidence exists from randomized trials that screening for lung cancer is superior to no screening is controversial. It is true that the findings of three National Cancer Institute (NCI)-sponsored randomized trials42223 in the United States and a randomized trial from Czechoslovakia24 that was conducted in the 1970s and 1980s have been interpreted as indicating that CXR screening was ineffective. This is because of a failure to demonstrate significant reductions in lung cancer mortality in populations that were randomized to CXR screening. On the other hand, these same trials have consistently demonstrated significant and rather dramatic improvements in long-term survival in experimental populations.
In the Mayo Lung Project,25 the 5-year survival rate was more than twofold higher in the group randomized to CXR screening (33% vs 15%, respectively).While the Memorial Sloan-Kettering Project26 and the Johns Hopkins Lung Project27 failed to demonstrate an advantage for the addition of sputum cytology investigation to an annual CXR alone, long-term survival was far superior to the contemporary data.28 In the Memorial Sloan-Kettering study,26 the 5-year survival rate was 35% in both groups.
Screening has been interpreted as being ineffective, because the mortality/survival discrepancy has been interpreted as indicating that CXR screening led to the overdiagnosis of lung cancer.2930313233 However, this has been a point of debate. A direct analysis of the data has suggested that the overdiagnosis was minimal and does not account for the survival/mortality discrepancies in these trials.3435
Indeed, long-term survival was only achieved among those persons undergoing surgical resection in these trials.3536 In the Mayo Lung Project,14 there were no long-term survivors among 185 lung cancer patients who did not undergo resection. In dramatic contrast, 50% of 181 resected patients were cured. Sobue et al37 have reported similar findings from Japan. There is, accordingly, no direct evidence that many nonlethal or "lanthanic"3839 lung cancers exist.
Furthermore, a recent analysis40 has indicated that overdiagnosis secondary to comorbid disease is minimal in lung cancer. Read et al40 examined the records of 11,558 patients with lung, breast, prostate, and colon cancers, and concluded that concurrent comorbidity had the greatest prognostic impact among groups with the most favorable survival (ie, those patients with localized prostate cancer and breast cancer) and the least impact among groups with the poorest survival rates, which included most patients with lung cancer.40
An important question is whether a CXR or a CT scan should be the preferred method of screening. There is no question that the CT scan represents a technological advance and has much greater sensitivity than the conventional CXR. The conventional CXR is known to be scarcely sensitive to the presence of tumors that are < 2 cm in diameter.1041
A CT scan is clearly capable of detecting lesions than are too small to be reliably detected on a conventional CXR. In four nonrandomized studies,42434445 which included two in the United States4344 and two in Japan,4245 52% to 85% of cancers detected by CT scan were small stage IA lesions. In each study, many CT scan-detected cancers were not visible on a CXR. The ability to detect small early-stage cancers has the potential to translate into substantial improvements in survival.
Indeed, impressively high survival was reported from the Anti-Lung Cancer Association Project in Japan.45 Among 36 lung cancers that were detected on initial or repeat screening, the overall 5-year survival rate was 71% (95% confidence interval, 52 to 90%).45
However, unresolved issues remain with regard to CT scan screening. Problems with specificity have led to a high rate of false-positive findings in these observational studies.43444546 A higher false-positive rate with CT scans compared to CXRs was recently reported in the NCI-sponsored Lung Screening Study.47 In this randomized trial, which included 3,318 current or former smokers, false-positive findings were observed in 18.6% and 9.4%, respectively, of patients in the CT scan and CXR arms.
While a CT scan is more sensitive than a CXR, the detection of nodules on a CT scan that are not due to cancer presents problems with regard to anxiety and the possibility of performing unnecessary invasive procedures. In the Anti-Lung Cancer Association study,45 only 51% of patients (36 of 71 patients) undergoing biopsy because of nodules detected on CT scans were found to have lung cancer. To cope with this problem, the International Early Lung Cancer Action Project trial48 used a protocol that was explicitly designed to minimize such operations for benign disease. This algorithm is updated periodically based on the total experience, and its objective is to determine what further diagnostic studies are recommended when a small nodule is detected on CT scan screening.48 In the Mayo spiral CT scan screening trial,49 18% of all thoracic surgical procedures performed were for benign disease.
While the evidence supports the idea that overdiagnosis was not responsible for spurious survival advantages in older NCI-sponsored randomized trials focusing on CXR screening,34 no comparable evidence yet exists with regard to CT scan screening.5051 In the Lung Screening Study,47 there were 30 cancers detected by CT scan compared to 7 cancers detected by CXR on baseline screening. Higher lung cancer detection rates probably reflect the increased efficiency of CT scanning in the detection of small lung cancers, with its associated longer lead time. However, it is conceivable that some of the detected cancers might never have become life-threatening during the life of the patient.
Unresolved issues with regard to false-positive findings and overdiagnosis have implications with regard to the appropriateness and cost-effectiveness of CT scan screening.49525354 Accordingly, a definitive answer to the question of whether a CT scan is superior to a conventional CXR must await the results of ongoing randomized trials, particularly the National Lung Screening Study in the United States.55
It should also be noted that technical advances have likely improved on the effectiveness of CXRs, as employed in the older NCI-sponsored trials. Such enhancements include digital radiography, energy subtraction, and computer-aided detection.56 These newer CXR technologies have not been incorporated into ongoing randomized trials, including the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial57 and the National Lung Screening Study.55
Based on the existing evidence, high-risk individuals can be informed that if they choose to undergo screening for lung cancer, either a CXR or a CT scan is a reasonable option. It should be considered that in many parts of the world CT scanning is not widely available, and in the United States it may not be affordable by economically disadvantaged individuals. If a high-risk individual chooses to undergo screening and CT scanning is not available or affordable, annual CXR screening should be recommended.
| Footnotes |
|---|
Received for publication October 21, 2004. Accepted for publication December 14, 2004.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
F. W. Grannis Jr, W. M. Alberts, and D. Addrizzo-Harris There Are Major Problems With the American College of Chest Physicians Second Lung Cancer Guidelines Chest, April 1, 2008; 133(4): 1049 - 1051. [Full Text] [PDF] |
||||
![]() |
A Imperatori, R N Harrison, D N Leitch, F Rovera, G Lepore, G Dionigi, P Sutton, and L Dominioni Lung cancer in Teesside (UK) and Varese (Italy): a comparison of management and survival Thorax, March 1, 2006; 61(3): 232 - 239. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. M. Reich Assessing the Efficacy of Lung Cancer Screening Radiology, February 1, 2006; 238(2): 398 - 401. [Full Text] [PDF] |
||||
Read all eLetters
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |