Recent studies have shown that the survival of patients with lung cancer is improved if the tumour is resected before it becomes larger than 3 cm in diameter and before it spreads to lymph nodes. While this suggests a positive benefit from early detection, recent mass-screening studies have claimed that the benefit obtained from this procedure is illusory because it relates to a lead-time bias. This study reports the results of surgical resection of 143 primary lung cancers. The data confirmed that the predicted 5-year survival was greatest (74%) following resection of lesions that were less than 3 cm in diameter without node involvement. Analysis showed that the age of these patients was 63 +/- 8 years, the same as in patients with larger tumours and more extensive node involvement. This suggests that tumours progress rapidly from a stage at which resection is beneficial to stages at which it is not. Although it is desirable that tests predict the presence of small tumours, the high requirements for sensitivity and specificity at current prevalence rates for lung cancer make this goal impractical.