Estimating cancer risks from a retrospective cohort of patients with MMR gene mutations.

lative risks of less common cancers (ie, noncolorectal and nonendometrial) in Lynch syndrome. Using the Kaplan-Meier method, risks were estimated by age, sex, and mutated genes in a large German and Dutch cohort of 2,118 patients who were proven carriers of MLH1, MSH2, and MSH6 mutations. The authors stated that ascertainment biasshould not have severely affected their estimates because most familieswereidentifiedbyclusteringorearlyonsetofcolorectalorendometrial cancer. However, selection of patients from the German and Dutch national Lynch syndrome registries was based on Amsterdam criteria and Bethesda guidelines, which include tumors from the whole Lynchsyndrome spectrum. We are therefore concerned that a potential bias in cancer-risk estimates may have been introduced by the fact that certain rarer tumors contributed to the recruitment of families. Themostappropriatemethodstoestimatecancerrisksarethose