Review of Screening Principles and Update on Current Status
W e appreciated Dr. Black's review article 1 that indicated that com- puted tomography (CT) screening for lung cancer is a ''hot topic.'' Indeed, 2 recent publications, one by the International Early Lung Cancer Action Program Investigators (I-ELCAP) 2 and 1 by Bach et al, 3 reported discordant results. After a brief summarization, the I-ELCAP study reported a 10-year survival rate of 88% for patients with stage I disease. The study diagnosed lung cancer in 484 subjects among 31,567 asymptomatic individuals who underwent baseline screening. Of these 484 persons, 412 had clinical stage I lung cancer and 302 underwent surgical resection. The comparator cohort were 8 patients with clinical stage I lung cancer who received no treatment at all and died of their disease within 5 years. 2 Shortly after the I-ELCAP study stated the benefit of repeat CT scan in improving survival, a second study contradicted these findings, indicating that there was no mortality benefit with CT scanning, despite early detection. 3 In this second study, 3246 asymptomatic smokers were screened. The CT scans detected lung cancer in 144 of them, a number that is 3 times more than would be expected without screening. However, the number of individuals who died of lung cancer, 38, was approximately the same as it would have been without screening. A statistical model to estimate the expected number of cancers and the expected death rate without screening was used.
The main difference between the 2 studies is the question of what counts as evidence of effectiveness: a decline in the death rate from lung cancer or an increase in survival time after being diagnosed?. Thus, estimating survival, as the I-ELCAP study did, can be misleading and potentially confounded by lead-time and overdiagnosis biases of unknown magnitude, making meaningful interpretation of results difficult. For the I-ELCAP study, the endpoint was lung cancer-specific rather than overall survival. The critical measure of outcome for screening was long-term mortality among the entire screened population, not the survival time of diagnosed cases. Furthermore, when examining the comparator cohort, it might be reasonable to hypothesize that there are patients who were ineligible for any treatments and the authors did not state whether these untreated patients died from lung cancer or from competing causes.