This study applies a method for the estimation and temporal projection of cancer incidence. The estimates are disaggregated to the Local Health Units; they refer to Tuscany for the year 2000 and to the incidence of cancer at three specific sites (lung, breast, prostate). The mortality data derive from the Regional Mortality Registry and the incidence data from the Tuscan Cancer Registry, which includes two provinces: Florence and Prato. The choice of method is discussed and a validation of the estimates obtained, using the incidence data observed in the province of Prato. The estimates and the temporal projections are required for accurate health planning, and they present the means to overcome health planning based on the historical data of medical care only. We suggest that appropriate communication of these estimates should address their limits, due to the uncertainty of the temporal prediction on the one hand and the possibility (only partially verifiable) of the violation of the assumptions on which any method for the estimation of cancer incidence is based.