In this retrospective, single institution study on a large group of patients operated on for non-small lung cancer two time periods were compared and survival differences analysed. Patients with incomplete resection were excluded and the two time periods considered were not equal in length. More patients were operated on during the late era which extended until 2002, resulting in a rather short median follow-up time and a low number of patients at risk at 5 years for the late period. The authors are to be congratulated for an extremely low operative mortality. In their detailed analysis, they found a significant, recent improvement in survival for patients with pathological stage I and III disease, but not for stage II disease. When comparing recent results with historical controls many possible confounding factors have to be considered and the results interpreted with caution. Have surgeons become better recently? Probably, they got more experienced as the authors indicate themselves. Has staging become better? This is undoubtedly true because new diagnostic methods as high resolution computed tomography and positron emission tomography became available during the last decade. Intraoperative staging probably also became more accurate. So, the Will Rogers phenomenon or stage migration may have influenced the results. Even more strikingly is the fact that the number of pneumonectomies decreased from 6% to only 1% in the late time period, an extremely low figure. The number of segmentectomies increased from 11 to 25%, introduction of screening programs resulting in the detection of more early stage I cancers in the late period. Unfortunately, no specific details on local or distal recurrences are provided for this group of patients. On the other hand, the number of patients treated by induction therapy decreased from 12 to 10%, which is rather surprising, considering the combined modality therapies introduced over the last years. Perhaps more important than the comparison between the early and late period, are the overall survival figures reported by the authors. In the late period the 5-year survival rate for stage IA disease was 89.6%, an extremely high figure and much better than usually reported in recent series. Although the comparison between the early and late time period may not be entirely valid, the authors report impressive survival results in a large group of patients which may certainly be considered for the database project of the International Association for the Study of Lung Cancer, preparing the next revision of the TNM classification which is scheduled for 2007.