Surgical resection is the treatment of choice for non-advanced lung cancer, but is encumbered with an overall relative poor long time prognosis. The purpose of this study was to examine if long time survival for patients operated for non-small cell lung cancer have changed over a 15 years period. We retrospectively studied hospital records of the 351 patients operated, with the intention to cure, for a primary non-small cell carcinoma (NSCLC) in our department between 1 January 1988 and 31 December 2002. Preoperative clinical variables were noted together with variables allowing staging based on pathological examination. Absolute survival and survival relative to expected was studied for the whole group using uni- and multivariate Cox analyses. Early 30 days mortality was 2.0%. The 5-year absolute and relative survivals for all patients were 46.3% and 52.6%, respectively. After 10 years corresponding values were 32.9% and 44.6%. At the end of the study, the 15-year absolute survival was 27.8% with a relative survival of 46.2%. Univariate analysis revealed that age, gender, nodular stage, tumour size, p-stage, type of resection, time of operation and additional cardiovascular disease at the time of operation significantly influenced survival. Multivariate analysis for all patients revealed that low age, female gender, low nodular stage, and operation late in the study period were significant prognostic factors predicting improved survival. When including a population based age- and gender-adjusted median expected life time for every patient as a predictor for survival, only female gender and low nodular stage were additional significant and independent positive prognostic factors.