Background: The aim was to compare long-term survival rates for different types of cancer estimated by means of up to date period analysis with those from more traditional cohort analysis. Patients and methods: Data from the Eindhoven Cancer Registry were used. In total 140 137 newly diagnosed patients diagnosed between 1980 and 2002 and followed until 1 January 2005 were included. Five-, 10- and 20-year relative survival rates were calculated. Results: For total cancer in men and women, childhood cancer, rectal cancer, melanoma in women, breast cancer, prostate cancer and all leukaemias, much higher 10-year survival rates were found with period analyses (differences with cohort analyses were 5.1%, 3.6%, 7.4%, 5.6%, 6.5%, 4.0%, 5.1% and 10.5%, respectively). For laryngeal and bladder cancer the 10-year survival rates estimated with period analyses were about 7.5% lower compared with those estimated by means of cohort analyses. Conclusions: Period analysis, based on the most recent period of diagnosis, enabled us to show higher survival rates for total cancer, childhood cancer, rectal cancer, melanoma, breast cancer, prostate cancer and acute leukaemia, but also lower rates for laryngeal and bladder cancer. Period analysis should be the preferred tool for showing up-to-date survival rates to cancer patients and their physicians.