OBJECTIVES: Accurate estimates of cancer epidemiology are fundamental to quantifying the economic burden of cancer as well as supporting a variety of researches in public health and commercial activities. However, the complete prevalence and incidence of late-stage cancers is difficult to obtain as most surveillance programs report cases at initial diagnosis, so recurrent cases, by definition are not captured. The objective of this study is to present a simple mathematical approach to estimate the epidemiology of late-stage cancers. METHODS: We developed an Excel-based tool to estimate the epidemiology of late-stage cancers with minimal historic information. Data needed include annual national or local cancer-related mortality rates, late-stage cancer survival rates and population size. Our approach starts with the patients who died from the specific cancer and tracks back to estimate the incidence and prevalence. The approach assumes most deaths attributed to the cancer are late-stage disease. We tested our approach by estimating the incidence and prevalence of metastatic breast cancer and metastatic melanoma with historic mortality and survival data from the National Cancer Institute Surveillance, Epidemiology and End Results (SEER) Program. RESULTS: We estimated that the 2007 US incidence of stage IV breast cancer and melanoma were approximately 32.4 per 100,000 women and 2.7 per 100,000 persons, respectively. These results corresponded to a total of 49,505 patients (10,426 newly diagnosed and 39,079 recurrent cases) for stage IV breast cancer and total of 8,279 patients (3,690 newly diagnosed and 4,589 recurrent cases) for late-stage melanoma. Results are also available by age and gender groups. CONCLUSIONS: Comparison of results using this epidemiology tool with estimates from databases and chart review studies demonstrated that our approach is reasonably accurate in its estimation. This approach could be adopted for uncommon cancers or regions with scarce data.