Standardized approach for microsatellite instability detection in colorectal carcinomas

Background: Ubiquitous mutations in microsatellite DNA sequences define a s pecific type of genetic instability, termed microsatellite instability (MSI ). Various approaches have been used to identify the presence and degree of AISI. To define standard diagnostic criteria for MSI, we developed and tes ted a mathematical model. Methods: We designed an algorithm for the efficie nt characterization of MSI and used it to analyze data on six microsatellit e markers in colorectal carcinoma and normal tissues from 415 patients. The oretical models considering one, two, or three populations were tested agai nst the data collected. Results: The observed frequencies of MSI in our ser ies of samples best fit a two-population model, stable and unstable, define d by instability in tno or more of four to sis markers analyzed. MSI was ob served in 7.5% of the tumors. The misclassification rate was less than 5% w hen any four loci were analyzed and less than 1% when the sis markers were used. A stepwise strategy, consisting first of a bulk screening of two loci and then a second screening of two to four additional markers, provided ex cellent sensitivity (greater than or equal to 97%) and specificity (100%). Tumors with MSI had distinctive genetic and clinicopathologic features, inc luding better patient survival. Conclusion: To assess the presence of MSI i n colorectal cancer, we have developed a simple, sensitive, and specific ap proach based on the apparent good fit of the data to a two-population model . Its application to a prospective series of patients with colorectal carci nomas demonstrates that the presence of MSI characterizes a subset of less aggressive tumors.