BACKGROUND: Net survival, the survival that might occur if cancer was the only cause of death, is a major epidemiological indicator. Recent findings have shown that the classical methods used for the estimation of net survival from cancer registry data, referred as to “relative-survival methods,” provided biased estimates.
OBJECTIVES: The aim of this study was to provide, for the first time, long-term net survival rates for colorectal cancer by using a population-based digestive cancer registry.
DESIGN: This study is a population-based cancer registry analysis. The recently proposed unbiased nonparametric Pohar-Perme estimator was used.
PATIENTS: Overall, 14,715 colorectal cancers diagnosed between 1976 and 2005 and registered in the population-based digestive cancer registry of Burgundy (France) were included.
MAIN OUTCOME MEASURES: The primary outcome measured was cancer net survival, ie, the survival that might occur if all risks of dying of other causes than cancer were removed
RESULTS: Ten-year net survival increased from 31% during the 1976 to 1985 period to 47% during the 1986 to 1995 period and then leveled out (48% during the 1996–2005 period). There was a major improvement in 10-year net survival after resection for cure and for stage I to III. It was striking for stage III cancers, for which 10-year net survival increased from 21% (1976–1985) to 49% (1996–2005). The corresponding net survivals were 70% and 87% for stage I and 49% and 65% for stage II. These trends can be related to the decrease in operative mortality, the increase in the proportion of patients resected for cure, and the improvement in stage at diagnosis. They were mainly seen between 1976 and 1995, explaining why survival leveled out after 1995.
LIMITATIONS: The study was limited by its retrospective and population-based nature.
CONCLUSIONS: Further improvements for colorectal cancer management can be expected from more effective treatments and from the implementation of organized cancer screening.