Knowledge of the cancer profile is an important step in planning rational cancer control programs andevaluation of their impact. Due to rapid changes in cancer incidence in China, national surveys may beinsufficiently timely to provide adequate descriptions of the national burden. To evaluate the utility of cancerregistries in describing the national cancer profile, this study compared two methods of estimating nationalcancer-specific incidence and mortality in China 2005, based on the Third National Death Survey (methodI) as compared with registry material (method II). A total of 2.6 million cancer cases and 1.8 million cancerdeaths were estimated by method I, as compared to 2.8 million cancer cases and 1.9 million cancer deaths usingmethod II. The higher level of burden using the latter method in part may be due to a sizable differential in themagnitude of incidence rates across registries for certain cancer sites. Most cancer registries have been locatedin relatively more developed urban areas, or rural areas associated with higher risk for certain cancers. Thereare substantial differences in the cancer profile between urban and rural communities in China, and there maybe concerns regarding the national representativeness of the data aggregated from this set of cancer registries.Timely and reliable estimation of cancer can only be realized if accurate information is available from cancerregistries covering representative samples of the country.