The clinical differentiation between hereditary nonpolyposis colorectal cancer (HNPCC) and attenuated familial adenomatous polyposis (AFAP) is very difficult. The 62-yr-old proband presented with duodenal adenocarcinoma. His history of subtotal colectomy for colon cancer, the rarity of duodenal adenocarcinoma in the general population, and his family history of colon cancer made us suspect that he might have FAP. We investigated this family by obtaining medical records and performing gene analysis. The proband had only 10 adenomatous colon polyps when he underwent subtotal colectomy for the cancer, so classic FAP was excluded. His family history included rectal cancer in his brother at 69 yr of age, colon cancer in his mother at 75 yr, and colon cancer in one maternal cousin at 42 yr. Three months after we started to study this family, the proband's 32-yr-old son presented with rectal cancer. His family fulfilled the Amsterdam criteria for HNPCC, but AFAP could not be excluded. Upon gene testing, the proband was negative for APC gene germline mutation, which made AFAP highly unlikely. Moreover, high microsatellite instability (MSI) was detected in his adenomas and cancer tissues. The fulfillment of Amsterdam criteria, the exclusion of FAP and AFAP, and the high MSI established the diagnosis of HNPCC in this family. We also summarize the differences between FAP, AFAP, and HNPCC; extend the graphic description of the MSI mechanism; and propose a diagnostic strategy for HNPCC.