Genomic analysis using tissue samples is an essential approach in cancer genetics. However, technical and biological limits exist in this approach. Microsatellite instability (MSI) is frequently observed in human tumors. MSI assays are now prevalent and regarded as commonplace. However, several technical problems have been left unsolved in the conventional assay technique. Indeed, the reported frequencies of MSI differ widely in each malignancy. An example is pancreatic cancer. Using a unique fluorescent technique, we found that MSI is extremely infrequent in this malignancy, despite the relatively high frequencies in some reports. In a series of simulations, we have demonstrated that the extremely low frequency was derived neither from less sensitive assays nor from a scarcity of cancer cells in tissue samples. Furthermore, analyzing laser-capture microdissection (LCM)-processed cell populations of a microsatellite-unstable colorectal cancer cell line, HCT116, we have shown that MSI can be detected only when comparing two cell populations that have grown independently to a sufficiently large size. When MSI is not detected in analyses using tissue samples, LCM is not advisable. We therefore did not extend our study to LCM of tissue specimens. We conclude that microsatellite sequence alterations are not detectable in human pancreatic cancer.