Introduction - This article presents some results from the first survival study based on the combined data from the cancer registries of the Francim network. Materials and Methods - Data from 205,562 cases registered between 1989 and 1997 were analyzed. Relative survival was estimated by modeling the excess cancer-specific mortality rate for 46 cancer sites. Results - The age-standardized five-year relative survival rates were as follows: 84% for breast cancer, 77% for prostate cancer, and, in men and women respectively, 55% and 57% for colorectal cancer, and 12% and 16% for lung cancer. An advanced age at diagnosis was generally associated with higher excess mortality than that observed in younger patients. This effect was often observed during the first year following diagnosis and much less thereafter. For certain sites, the prognosis was better in women, sometimes significantly so. Finally, for certain cancers, the most recently diagnosed cases had a better prognosis. Conclusion - The procedures developed through this first study by the Francim network allow for regular updates of the results. A new tool for epidemiological cancer surveillance is now available in France. (R.A.)