Lung Cancer Screening Results: Easily Misunderstood

Abstract

The recent report on lung cancer screening by the International Early Lung Cancer Action Project (IELCAP) group, published in The New England Journal of Medicine,1International Early Lung Cancer Action Program Investigators Survival of patients with stage I lung cancer detected on CT screening.N Engl J Med. 2006; 355: 1763-1771Crossref PubMed Scopus (1469) Google Scholar and the media hype that followed2Kolata G Study sees gain on lung cancer. New York Times.Available at: www.nytimes.com/2006/10/26/health/26lung.html?ex=1164171600&en=dc4176562e22ba10&ei=5070Date: October 26, 2006Google Scholar, 3Knox R Study: scan may be best test for lung cancer.Available at: www.npr.org/templates/story/story.php?storyId=6383404Google Scholar, 4Associated press Chest scans for smokers may save lives. November 14, 2006.Available at: www.msnbc.msn.com/id/15414518/Google Scholar unfortunately say more about what is misunderstood by screening for lung cancer than about the potential for computed tomographic (CT) screening to reduce death due to lung cancer. The IELCAP investigators reported a 10-year disease-specific survival of 80% for participants with lung cancer detected by CT.1International Early Lung Cancer Action Program Investigators Survival of patients with stage I lung cancer detected on CT screening.N Engl J Med. 2006; 355: 1763-1771Crossref PubMed Scopus (1469) Google Scholar One is tempted to compare this to the current 5-year survival of 15% for nonscreen-detected lung cancer5Jemal A Siegel R Ward E et al.Cancer statistics, 2006.CA Cancer J Clin. 2006; 56: 106-130Crossref PubMed Scopus (5509) Google Scholar and conclude that clearly CT screening is proved to reduce cancer deaths. As desirable as that conclusion may be, it cannot legitimately be made. No scientifically valid conclusions can be made by comparing survival in a screen-detected group to a nonscreen-detected group. Screening introduces inherent bias, and thus apparent survival may be dramatically improved even if screening is of no benefit and the number of lives saved remains unchanged.6Bach PB Kelley MJ Tate RC McCrory DC Screening for lung cancer: a review of the current literature.Chest. 2003; 123: 72S-82SCrossref PubMed Google Scholar To understand these issues, the reader must discriminate among true survival (the interval between biological onset of disease and death, ie, an often incalculable value), apparent survival (how long people live after diagnosis), and mortality (how many people die of the disease). If CT screening is effective, it will result in fewer deaths due to lung cancer among people screened compared with those at similar risk for the disease who do not undergo screening. Given the inability to determine a change in true survival for lung cancer, mortality reduction is the key to show that screening is effective. Living comparatively longer after screen-detected diagnosis does not on its own establish that outcome is improved and death due to lung cancer has been reduced. Demonstrating survival benefit with screening compared to no screening is anticipated because of the inherent biases (lead time, length time, and overdiagnosis) of screening. Lead-time bias is defined as apparent improved survival by earlier diagnosis with screening even when the outcome (death due to lung cancer) remains unchanged. This effect occurs because there is simply more time between diagnosis and death than there would have been without screening. Length-time bias has the effect of improving apparent survival by selecting cancers that by their slower growing nature have a good prognosis compared to cancers found by symptoms in the absence of screening. Screening may miss the most aggressive cancers—patients present with symptoms between annual screening examinations or, alternatively, participants might die of lung cancer without ever being recognized as having the disease. Overdiagnosis bias is an extreme form of length-time bias and improves apparent overall survival by the identification of cancer that, in the absence of screening, would have gone undetected because the cancer would neither cause symptoms nor death due to its indolent nature. Observational studies such as the IELCAP can provide important information but cannot prove efficacy of screening through mortality reduction. Several prior single-arm (no control group) observational studies of CT screening studies have reported the ability of CT screening to detect a high proportion of surgical stage I cancers. Stage I (T1 or T2 and N0M0) cancers represented 70% to 93% of cancers detected by the baseline CT screen (prevalence) and 61% to 100% of those detected on subsequent scans (incidence) in these studies.7Swensen SJ Jett JR Hartman TE et al.CT screening for lung cancer: five-year prospective experience.Radiology. 2005 Apr; 235 (Epub 2005 Feb 4.): 259-265Crossref PubMed Scopus (572) Google Scholar, 8Kaneko M Eguchi K Ohmatsu H et al.Peripheral lung cancer: screening and detection with low-dose spiral CT versus radiography.Radiology. 1996; 201: 798-802PubMed Google Scholar, 9Sone S Takashima S Li F et al.Mass screening for lung cancer with mobile spiral computed tomography scanner.Lancet. 1998; 351: 1242-1245Abstract Full Text Full Text PDF PubMed Scopus (877) Google Scholar, 10Henschke CI McCauley DI Yankelevitz DF et al.Early Lung Cancer Action Project: overall design and findings from baseline screening.Lancet. 1999; 354: 99-105Abstract Full Text Full Text PDF PubMed Scopus (2170) Google Scholar, 11Henschke CI Naidich DP Yankelevitz DF et al.Early Lung Cancer Action Project: initial findings on repeat screenings.Cancer. 2001; 92: 153-159Crossref PubMed Scopus (459) Google Scholar, 12Nawa T Nakagawa T Kusano S Kawasaki Y Sugawara Y Nakata H Lung cancer screening using low-dose spiral CT: results of baseline and 1-year follow-up studies.Chest. 2002; 122: 15-20Crossref PubMed Scopus (296) Google Scholar, 13McWilliams A Mayo J MacDonald S et al.Lung cancer screening: a different paradigm.Am J Respir Crit Care Med. 2003 Nov 15; 168 (Epub 2003 Jul 25.): 1167-1173Crossref PubMed Scopus (110) Google Scholar Sobue et al14Sobue T Moriyama N Kaneko M et al.Screening for lung cancer with low-dose helical computed tomography: anti-lung cancer association project.J Clin Oncol. 2002; 20: 911-920Crossref PubMed Scopus (378) Google Scholar reported a 5-year overall survival of 71% for screen-detected cancers and a 5-year disease-specific survival of 85% for lung cancers. Similarly, the IELCAP group reported a 92% actuarial 10-year survival for stage I cancers resected within 1 month of diagnosis (62% of all cancers detected).1International Early Lung Cancer Action Program Investigators Survival of patients with stage I lung cancer detected on CT screening.N Engl J Med. 2006; 355: 1763-1771Crossref PubMed Scopus (1469) Google Scholar This is the largest study of its kind, but the results say the same thing, albeit with perhaps a louder voice, as the previous studies: CT screening has the ability to detect a large percentage of cancers in early stage, yielding an improvement in apparent survival. However, such a conclusion is not to be equated with prolongation of the interval between biological disease onset (whether detected or not) and death due to cancer or other cause, ie, true survival. The improved survival reported by the IELCAP group may simply be a reflection of bias from screening. An observational screening study of sufficient duration and follow-up may serve as a surrogate to a control trial in compensating for the issue of lead-time bias. In the IELCAP study, a first annual study was performed in 84% of the participants and a second annual scan in only 34%11Henschke CI Naidich DP Yankelevitz DF et al.Early Lung Cancer Action Project: initial findings on repeat screenings.Cancer. 2001; 92: 153-159Crossref PubMed Scopus (459) Google Scholar—these participants are included in the current IELCAP report.1International Early Lung Cancer Action Program Investigators Survival of patients with stage I lung cancer detected on CT screening.N Engl J Med. 2006; 355: 1763-1771Crossref PubMed Scopus (1469) Google Scholar The median baseline screen date overall for the combined study was 2001,1International Early Lung Cancer Action Program Investigators Survival of patients with stage I lung cancer detected on CT screening.N Engl J Med. 2006; 355: 1763-1771Crossref PubMed Scopus (1469) Google Scholar thus giving a median follow-up of not more than 5 years. Given the limited follow-up, lead-time bias remains a concern. In the Mayo CT screening study reported in 2005, 95% of the participants were imaged at 4 years and 80% at 5 years of follow-up.7Swensen SJ Jett JR Hartman TE et al.CT screening for lung cancer: five-year prospective experience.Radiology. 2005 Apr; 235 (Epub 2005 Feb 4.): 259-265Crossref PubMed Scopus (572) Google Scholar At 5-years of follow-up, 20 (91%) of 22 participants with stage I lung cancers on the baseline scan were alive, yet mortality in this study was not suggested to be reduced in comparison to historic controls. The fact that CT screening is most sensitive at detecting peripheral nodular cancers lends support for length-time bias as an issue because these cancers are more likely to be adenocarcinomas with a better prognosis. Computed tomographic screening is less sensitive for central cancers, such as squamous cell and small cell cancers, the latter having shorter doubling times and a poorer prognosis. In the IELCAP study, 71% of the stage I cancers detected were adenocarcinomas.1International Early Lung Cancer Action Program Investigators Survival of patients with stage I lung cancer detected on CT screening.N Engl J Med. 2006; 355: 1763-1771Crossref PubMed Scopus (1469) Google Scholar Observational studies cannot address the impact of overdiagnosis bias (detection of cancers that would not result in death), and multiple studies support that this is a real problem in lung cancer screening.15Marcus PM Bergstralh EJ Zweig MH Harris A Offord KP Fontana RS Extended lung cancer incidence follow-up in the Mayo Lung Project and overdiagnosis.J Natl Cancer Inst. 2006; 98: 748-756Crossref PubMed Scopus (175) Google Scholar, 16Hasegawa M Sone S Takashima S et al.Growth rate of small lung cancers detected on mass CT screening.Br J Radiol. 2000; 73: 1252-1259PubMed Google Scholar Only a randomized controlled trial can determine whether mortality reduction is achieved in the participants who underwent screening.6Bach PB Kelley MJ Tate RC McCrory DC Screening for lung cancer: a review of the current literature.Chest. 2003; 123: 72S-82SCrossref PubMed Google Scholar The randomized controlled screening trials conducted at Johns Hopkins, Memorial Sloan-Kettering, and Mayo Clinic that began in the 1970s showed that screening, even with chest x-ray and sputum cytology, resulted in the identification of more cancers, more early-stage cancers, and an apparent improvement in survival.17Frost JK Ball Jr, WC Levin ML et al.Early lung cancer detection: results of the initial (prevalence) radiologic and cytologic screening in the Johns Hopkins study.Am Rev Respir Dis. 1984; 130: 549-554PubMed Google Scholar, 18Flehinger BJ Melamed MR Zaman MB Heelan RT Perchick WB Martini N Early lung cancer detection: results of the initial (prevalence) radiologic and cytologic screening in the Memorial Sloan-Kettering study.Am Rev Respir Dis. 1984; 130: 555-560PubMed Google Scholar, 19Fontana RS Sanderson DR Taylor WF et al.Early lung cancer detection: results of the initial (prevalence) radiologic and cytologic screening in the Mayo Clinic study.Am Rev Respir Dis. 1984; 130: 561-565PubMed Google Scholar, 20Fontana RS Sanderson DR Woolner LB Taylor WF Miller WE Muhm JR Lung cancer screening: the Mayo program.J Occup Med. 1986; 28: 746-750Crossref PubMed Scopus (462) Google Scholar However, despite these achievements, more frequent screening did not decrease the number of deaths due to lung cancer compared to the control groups who had less frequent screening. These randomized studies also showed a 5-year survival of 70% for screen-detected stage I lung cancers treated with surgical resection compared to 10% for clinical stage I lesions not treated with surgery. 21Flehinger BJ Kimmel M Melamed MR The effect of surgical treatment on survival from early lung cancer: implications for screening.Chest. 1992; 101: 1013-1018Crossref PubMed Scopus (301) Google Scholar Given the superiority of cancer detection with CT vs chest x-ray, it is not surprising (rather expected) that CT screening would show further improved survival for stage I cancers vs that of chest x-ray screening. A meta-analysis of randomized controlled trials showed that, despite an improvement in apparent survival, there was actually a significant increase in mortality associated with more frequent chest x-ray screening due to the resulting interventions leading to more deaths.22Manser R Wright G Hart D Byrnes G Campbell DA Surgery for early stage non-small cell lung cancer.Cochrane Database Syst Rev. 2005; 1 (CD004699)PubMed Google Scholar Demonstrating mortality reduction would be the assurance that screening resulted in more good than harm. The editorial by Unger23Unger M A pause, progress, and reassessment in lung cancer screening [editorial].N Engl J Med. 2006; 355: 1822-1824Crossref PubMed Scopus (32) Google Scholar that accompanied the IELCAP report appropriately stopped short of recommending a move to screen patients with CT. We sincerely hope that CT screening will result in a reduction in deaths due to lung cancer. However, mere survival numbers from observational studies are not conclusive. The fact is that mortality reduction needs to be demonstrated for lung cancer screening to be proved effective. Good science necessitates that this be demonstrated; public health policy and advocacy groups should accept nothing less. Mortality reduction is needed before making the move to incorporating CT screening as usual care for our patients at risk for lung cancer. We anticipate that several ongoing trials, including the National Lung Screening Trial24Gohagan JK Marcus PM Fagerstrom RM et al.Lung Screening Study Research Group. Final results of the Lung Screening Study, a randomized feasibility study of spiral CT versus chest X-ray screening for lung cancer.Lung Cancer. 2005; 47: 9-15Abstract Full Text Full Text PDF PubMed Scopus (265) Google Scholar (randomized controlled trial comparing CT screening to chest x-ray), will answer this question.