Patients with inflammatory bowel disease (IBD), particularly ulcerative colitis (UC), have an increased incidence of colorectal carcinoma. The underlying mechanism is unknown, but we postulated that microsatellite instability (MSI) might predispose the colonic mucosa of UC patients to mutations, thereby increasing their cancer risk. We also sought to determine the frequency of K-ras mutations, to determine whether MSI predisposed to K-ras mutations and to compare the molecular phenotype of biopsy and resection specimens in the same patient. We also sought to determine whether molecular alterations found in biopsy specimens presaged their presence in subsequent resection specimens. Two hundred fifty-eight specimens from 52 patients were examined for K-ras mutations by direct sequencing. Seventy-one of the specimens were neoplastic. MSI was evaluated after polymerase chain reaction (PCR) amplification using primers directed at 8 microsatellite loci. Of the patients, 18.2% had K-ras mutations, and 30.8% had MSI in at least 1 locus. Of K-ras mutations, 81.8% were G to A substitutions involving the second nucleotide of codons 12 or 13. Only 0.7% of the samples showed a high level of MSI. No relationship existed between MSI and K-ras mutations, even in the 2 samples with high-level MSI. The numbers are small, but it appeared that MSI in biopsies failed to predict its presence in resection specimens. In contrast, K-ras mutations present in biopsy specimens tended to predict their presence in resections. K-ras mutations were found predominantly in neoplastic mucosae, whereas MSI was found predominantly in regenerative mucosae. The lack of any relationship between MSI and K-ras mutations suggests that MSI in the UC replicative mucosa does not predispose to colonic neoplasia via a K-ras-mediated pathway. This is probably related to the fact that the MSI is generally low-level MSI.