Assessing colorectal cancer mismatch repair status in the modern era: a survey of current practices and re-evaluation of the role of microsatellite instability testing.

Hissong E1, Crowe EP1, Yantiss RK1, Chen YT2.

Author information

Abstract

Results of DNA mismatch repair testing are used to detect Lynch syndrome and have prognostic and therapeutic implications among patients with sporadic colorectal carcinomas. Immunohistochemistry for mismatch repair proteins (MLH1, PMS2, MSH2, MSH6) and PCR for microsatellite instability are two established methods for assessing mismatch repair function. Older literature suggested a discordance rate of approximately 5% between these assays, leading some institutions to perform dual testing on all cases. Although universal mismatch repair testing is now recommended by multiple professional organizations, none provide guidelines regarding preferred assays. We surveyed 96 academic and nonacademic institutions to assess Lynch syndrome screening practices and evaluated discordance rates between immunohistochemistry and PCR among 809 colorectal cancers tested in our own institution. Our survey demonstrated no significant differences between academic and nonacademic practices with respect to testing strategies. Eighty six percent performed universal screening, and usually (76%) employed immunohistochemistry on initial biopsy samples. Only 20% employed PCR; these were mostly academic practices that used both immunohistochemistry and PCR (p < 0.01 compared with the nonacademic groups). Loss of MLH1/PMS2 staining was often (90%) followed by either BRAF mutational analysis or MLH1 methylation assays. Only 24% adhered to WHO recommendations to assign histologic grade based on mismatch repair status. We found only 3 cases (0.4%) with discordant immunohistochemistry and PCR results in our own practice: 1 reflected decreased MSH-6 staining in a neoadjuvantly treated microsatellite stable tumor, 1 MLH1-deficient tumor showed diminished MLH1/PMS2 in the tumor compared with internal control, and 1 case reflected an error in the molecular laboratory. Overall, our results showed extremely low discordance between methods assessing mismatch repair status and would suggest immunohistochemistry as the preferred single screening test. PCR can be reserved for cases that show equivocal immunostaining patterns.