摘要：Purpose: To investigate the long-term outcome of using X-ray and the 3-day "pooled" sputum cytology in lung cancer screening using the Mayo Lung Project data. Methods: In this study, the statistical method developed by the authors using the Mayo Lung Project data was applied. Each participant eventually falls into one of the four categories depending on whether they would be diagnosed with lung cancer and whether symptoms would have appeared before death. The derived probability of each outcome was presented. Bayesian inference was made for the percentages of true-early-detection, no-early-detection, over-diagnosis, and symptom-free-life. The authors present estimates of these proportions for male heavy smokers by using Markov Chain Monte Carlo samples in the simulation. Human lifetime was treated as a random variable. Results: The probability that a male heavy smoker will live a lung cancer-free life is about 80%; the probability for developing lung cancer for a male heavy smoker is about 20%, much higher than in the general population (6.95%). The probability of over-diagnosis is low, about 6%–9% among the early detected cases. The probability of true-early-detection is greater than 90%. Finally, the probability of no-early-detection increases as the screening interval increases; among the diagnosed cases, it is about 15% if the screens are 6 months apart, and it is close to 50% if the screens are 2 years apart. Conclusion: This research provides a framework and a systematic approach for policy makers regarding how to evaluate long-term outcomes of chest X-ray and pooled sputum cytology in lung cancer screening.
关键词：true early detection; no early detection; over diagnosis; symptom-free life; pooled sputum cytology
