Optimally accurate and valid estimates of cancer risk in genetically- defined subgroups requires population-based research. For rare genetic traits, very large studies are needed. Such is the case in hereditary nonpolyposis colorectal cancer syndrome (HNPCC), caused by mutations in mismatch repair (MMR) genes. A potentially more efficient approach is genotyped-proband design (GPD), in which probands are genotyped and the phenotypes of their relatives are investigated. However, to date most information comes from registry-based studies, where ascertainment bias makes interpretation difficult. Development in testing technology will lead to more identified mutation carriers, producing a clinical imperative to estimate risk, despite these problems. We reviewed the available results, and concluded that male mutation carriers have a lifetime colorectal cancer risk of 74% or more; female mutation carriers have a lower risk which is still many times higher than the risk in the general population. Risk is highest between the ages of 40 and 60, but considerable even before age 40. Lifetime endometrial cancer risk is 42% or more; the highest incidence is between age 40 and 60, and diagnosis before the age of 35 is rare. MMR mutation carriers are at elevated risk for ovarian, gastric, urologic tract, small bowel, hepatobiliary tract cancer, and for brain tumors. The risk of these cancer types is much lower than the risk for colorectal and endometrial cancer, but accurate, especially age-related estimates of risk are not available. Prevention strategies depend on estimates of age-specific risk. Clearly, multicenter studies to obtain such estimates are needed.