Correspondence
Screening for Lung Cancer
N Engl J Med 2001; 344:935-936March 22, 2001
- Article
To the Editor:
The review by Patz et al. (Nov. 30 issue)1 on screening for lung cancer omits mention of a major historical change over the past three decades: the most common cell type has shifted from squamous-cell to adenocarcinoma — a shift that appears to result from the physicochemical changes in late-20th-century cigarette smoke (e.g., increased levels of tobacco-specific nitrosamines).2 For example, 21 of the 27 (78 percent) peripheral stage I carcinomas detected by computed tomography (CT) among the 1000 smokers in the recent Early Lung Cancer Action Project screening study were adenocarcinoma, 3 (11 percent) were adenosquamous, and 1 (4 percent) was squamous.3
In the lung, primary adenocarcinoma grows more slowly than squamous carcinoma; data on doubling time indicate that, on average, adenocarcinoma grows from 1 cm in diameter to 3 cm in 35 months and that squamous carcinoma grows that much in 16 months.4 The shift to a more slowly growing cell type lengthens the window of opportunity for presymptomatic diagnosis.
The surgical literature amply documents a five-year survival among persons with resected stage IA (<3 cm) non–small-cell lung cancer of 65 to 85 percent; among those with unresected stage I cancer, the five-year survival is approximately 5 percent.5 If “overdiagnosis” exists here, the effect appears to be modest. For resected adenocarcinomas 2 cm or smaller, a five-year survival after resection of 75 percent has been reported.6
Until mortality data from modern screening studies become available, we can only speculate on possible outcomes. However, the change in the predominant cell type to adenocarcinoma appears to be favorable for screening by imaging examinations.
6 ReferencesJohn H.M. Austin, M.D.
Steven D. Stellman, Ph.D., M.P.H.
Gregory D.N. Pearson, Ph.D., M.D.
Columbia Presbyterian Center, New York, NY 100321
Patz EF Jr ,Goodman PC ,Bepler G . Screening for lung cancer. N Engl J Med 2000;343:1627-1633
Full Text | Web of Science | Medline2
Stellman SD ,Muscat JE ,Hoffmann D ,Wynder EL . Impact of filter cigarette smoking on lung cancer histology. Prev Med 1997;26:451-456
CrossRef | Web of Science | Medline3
Henschke CI ,McCauley DI ,Yankelevitz DF , et al. Early Lung Cancer Action Project: overall design and findings from baseline screening. Lancet 1999;354:99-105
CrossRef | Web of Science | Medline4
Usuda K ,Saito Y ,Sagawa M , et al. Tumor doubling time and prognostic assessment of patients with primary lung cancer. Cancer 1994;74:2239-2244
CrossRef | Web of Science | Medline5
Flehinger BJ ,Kimmel M ,Melamed MR . The effect of surgical treatment on survival from early lung cancer: implications for screening. Chest 1992;101:1013-1018
CrossRef | Web of Science | Medline6
Noguchi M ,Morikawa A ,Kawasaki M , et al. Small adenocarcinoma of the lung: histologic characteristics and prognosis. Cancer 1995;75:2844-2852
CrossRef | Web of Science | Medline
To the Editor:
We think of CT screening for lung cancer as CT-based early (presymptomatic) diagnosis of the disease, with CT also involved in the definitive diagnosis after an initial positive result is obtained and in the assessment of the rate of growth of the detected lesion.
The ultimate goal of early diagnosis is a reduction in the case fatality rate associated with lung cancer, attained by early intervention instead of late intervention. The “disease-specific mortality” is a concept that is much too muddled. The intermediate goal, we agree, is to achieve a “stage shift” toward earlier-stage disease at diagnosis, but we are astounded by the authors' argument that “it remains to be seen whether a stage shift will result when low-dose CT is used for screening.” Our base-line diagnoses were stage I in 85 percent of the cases,1 and nothing less is expected from our repeated-screening experience.
Patz et al. state that “the true clinical significance of the small tumors found by screening is unknown” and advocate randomized, controlled trials. The clinical significance of those tumors, with and without early intervention, can be addressed by nonrandomized, quasiexperimental studies according to the paradigms set forth by Flehinger et al.2 and Sobue et al.3 If required and if the ethics permit, randomized, controlled trials on early intervention could be done. The kind of randomized, controlled trial Patz et al. advocate would never address either the frequency or overdiagnosis (of nonfatal “cancer”) according to subtype of small tumor or the curability rate of any type of cancer of the lung.
As for true, aggressive cancers, Patz et al. question the desirability of detection at sizes detectable by CT only. We wonder: Would they recommend delaying resection until the lesion becomes detectable by traditional-type radiography?
3 ReferencesOlli S. Miettinen, M.D., Ph.D.
David F. Yankelevitz, M.D.
Claudia I. Henschke, Ph.D., M.D.
Weill Cornell Medical College of Cornell University, New York, NY 100211
Henschke CI ,McCauley DI ,Yankelevitz DF , et al. Early Lung Cancer Action Project: overall design and findings from baseline screening. Lancet 1999;354:99-105
CrossRef | Web of Science | Medline2
Flehinger BJ ,Kimmel M ,Melamed MR . The effect of surgical treatment on survival from early lung cancer: implications for screening. Chest 1992;101:1013-1018
CrossRef | Web of Science | Medline3
Sobue T ,Suzuki T ,Matsuda M ,Kuroishi T ,Ikeda S ,Naruke T . Survival for clinical stage I lung cancer not surgically treated: comparison between screen-detected and symptom-detected cases. Cancer 1992;69:685-692
CrossRef | Web of Science | Medline
To the Editor:
The issue posed by the emergence of spiral CT as a potential screening tool for early lung cancer is not about changing screening recommendations now but rather about the design of the requisite validation trial. Is a long and expensive randomized trial required to establish whether this type of screening reduces cancer-related mortality? Could more efficient trial designs be substituted, especially considering the increasing rate of innovation in imaging technology? A definitive randomized trial could cost hundreds of millions of dollars and take a decade to complete. Recently, the merits of alternatives to randomized trial designs have been discussed.1-3 Nevertheless, many governments, regulatory bodies, and reimbursement agencies still require evidence from randomized trials before changing existing health care policy.
Currently, it appears that for lung-cancer screening, several studies with different trial designs are likely to proceed simultaneously with an unprecedented amount of sharing of data-management and image-processing tools, an approach that may eventually permit pooled analysis of the different cohorts. If this collective trial experience is used to define the relative contributions of the various potential biases, this analysis may provide an objective basis for streamlining the clinical-trial process, an outcome that may be beneficial regardless of the observed value of spiral CT.
3 ReferencesJames L. Mulshine, M.D.
National Cancer Institute, Bethesda, MD 20892Robert W. Day, M.D.
Fred Hutchinson Cancer Research Center, Seattle, WA 981091
Sackett DL ,Wennberg JE . Choosing the best research design for each question. BMJ 1997;315:1636-1636
CrossRef | Web of Science | Medline2
Concato J ,Shah N ,Horwitz RI . Randomized, controlled trials, observational studies, and the hierarchy of research designs. N Engl J Med 2000;342:1887-1892
Full Text | Web of Science | Medline3
Benson K ,Hartz AJ . A comparison of observational studies and randomized, controlled trials. N Engl J Med 2000;342:1878-1886
Full Text | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: Austin et al. speculate that the shift to adenocarcinoma and potentially more slowly growing tumors would increase the opportunity for making a diagnosis in a presymptomatic phase. This is an interesting but unproven hypothesis. One might just as easily argue that a more slowly growing cell type would permit later intervention and resection of the tumor without harmful effect.
Miettinen and colleagues suggest that our proposal to use a decrease in the rate of death from lung cancer as the ultimate goal of a diagnostic screening test is a “concept that is much too muddled.” To us this seems much less confusing and not subject to the same biases as case fatality rates. We agree with Miettinen et al. that a stage shift represents an intermediate goal in the assessment of CT screening for lung cancer, but we strongly disagree that they have thus far demonstrated a true stage shift. Reporting the relative distribution of stages (e.g., 85 percent with stage I disease) fails to confirm a stage shift, since it does not account for both essential components that must be demonstrated. First, there must be an increase in the number of patients with early-stage disease (which they have shown), and second, there must be a concomitant decrease in the number of patients with advanced-stage disease (which they have not shown).
Mulshine and Day discuss the most controversial issue concerning lung screening: trial design. We did not address this complicated subject since it was beyond the scope of the review article. Arguments have been made for both nonrandomized and randomized trials, but it behooves us to understand why we should accept any “quasiexperimental studies,” as suggested by Miettinen et al., when we have the opportunity to perform appropriately constructed hypothesis-driven trials.
We agree with the statement by Austin et al. that “until mortality data from modern screening studies become available, we can only speculate on possible outcomes.” At this time, no screening studies are complete. We need studies that will provide the scientific evidence with which to guide public policy.
Edward F. Patz, Jr., M.D.
Philip C. Goodman, M.D.
Duke University Medical Center, Durham, NC 27710Gerold Bepler, M.D., Ph.D.
Roswell Park Cancer Institute, Buffalo, NY 14263- Citing Articles (1)
Citing Articles
1
Ming-Yue Li, Tak W. Lee, Anthony P.C. Yim, George G. Chen. (2006) Function of Pparγ and Its Ligands in Lung Cancer. Critical Reviews in Clinical Laboratory Sciences 43:2, 183-202
CrossRef
