Routine testing for mismatch repair deficiency in sporadic colorectal cancer is justified

Abstract

This study prospectively examines the accuracy of immunohistochemical staining in the identification of mismatch repair defective (MMRD) colorectal cancer in routine clinical practice. The potential impact of this information on decisions regarding adjuvant treatment and germline testing were quantified. A consecutive series of fresh tissue (836 cancers) was obtained from 786 individuals undergoing curative surgery for colorectal cancer at one institution. As part of normal practice, each tumour was screened for the expression of MLH1 and MSH2 by immunohistochemical staining (IHC) and relevant clinicopathological details were documented. Microsatellite instability (MSI) was assessed using standard markers. Overall, 108 (13%) tumours showed loss of staining for either MLH1 (92 tumours) or MSH2 (16 tumours). The positive predictive value of mismatch repair IHC when used alone in the detection of MSI tumours was 88%, and the negative predictive value was 97%. Specificity and positive predictive value were improved by correlation with microsatellite status. Tumour stage (HR 3.5, 95% CI 2.0-6.0), vascular space invasion (HR 1.9, 95% CI 1.2-3.0) and mismatch repair deficiency (HR 0.2, 95% CI 0.05-0.87) were independent prognostic factors in stages II and III disease. Screening by mismatch repair IHC could reasonably have been expected to prevent ineffective treatment in 3.6% of stage II and 7.6% of stage III patients. The frequency of germline mismatch repair mutations was 0.8%, representing six unsuspected hereditary non-polyposis colorectal cancer (HNPCC) cases. Routine screening of colorectal cancers by mismatch repair IHC identifies individuals at low risk of relapse, and can prevent unnecessary adjuvant treatments in a significant number of individuals. Abnormal immunohistochemistry should be confirmed by microsatellite testing to ensure that false-positive results do not adversely impact on treatment decisions.