Title: A mathematical estimation of cancer incidence using data from population-based cancer registries Authors:

Accurate cancer incidence data is needed to plan, monitor and evaluate national cancer control programs. In Japan, however, such information is not available due to low reporting rates to cancer registries. In order to have more accurate incidence estimation, we developed a new method. We obtained a nonlinear regression model between observed incidence/mortality (IM) ratios and death certificate notification (DCN) rates in various cancer registries. This model gives us the ”true IM ratio”, which is at zero point of the DCN rate in the regression curve. This is an ideal registration state without any missing cases. By multiplying it to cancer mortality number from the National Vital Statistics, corrected cancer incidence can be estimated. The results of applying this method for Japanese cancer incidence estimation in 1997 were the true IM ratios 2.074 for men and 2.587 for women. Cancer incidences in Japan were estimated to be 346,000 for men and 280,000 for women. This method gives us the best guess regarding the National cancer incidence in a country having several cancer registries with various qualities. However, from the viewpoint of cancer control, every registry should establish a system to collect cancer incidence data completely in a region.