All stage IV colorectal cancer patients should be screened for Lynch syndrome, according to Dr. Douglas J. Hartman, and his colleagues at the University of Pittsburgh.
Published pathology models for predicting high microsatellite instability failed to identify up to 13% of Lynch syndrome–associated colorectal carcinomas in a study conducted by Dr. Hartman and his associates.
Current predictive pathology models for the detection of high microsatellite instability (MSI-H) colorectal cancers, including PREDICT and MSPath, rely primarily on right-sided tumor detection; however, Dr. Hartman and his associates determined that MSI-H tumors within the left colon and rectum were associated with Lynch syndrome in 57% of cases. In addition, revised Bethesda Guidelines for Lynch syndrome screening in colorectal cancer patients do not call for microsatellite instability testing in patients over age 60 years, yet 32% of left-sided tumors were identified in patients over age 60 years (Hum Pathol. 2013 Sep 10 [doi: 10.1016/j.humpath.2013.06.012]).
The researchers used MSI polymerase chain reaction (MSI-PCR) and DNA mismatch repair protein immunohistochemistry to prospectively analyze 1,292 colorectal cancers between January 2009 and July 2012. During 2009-2011, cases were evaluated using MSI-PCR only. In 2012, cases were first evaluated with immunohistochemistry. MSI-PCR was performed if there were any equivocal findings in the first round of testing.
MSI-H was found in 150 tumors; 112 were sporadic and 38 were probable Lynch syndrome as determined by BRAF V600E mutation, MLH1 promoter hypermethylation, cancer history, and germline mismatch repair gene mutation. All MSI-H tumors were analyzed for their grade, location, and histology; left-sided MSI-H tumors (n = 12; 57%) were more likely to be Lynch syndrome–related than were right-sided tumors (n = 26; 20%).
Neither PREDICT nor MSPath identified all Lynch syndrome–related tumors: PREDICT found 87% (33 of the 38), while MSPath identified 92% (35 of the 38). The researchers attributed these shortcomings to current predictive models’ "reliance on right-sided location."
"At our institution, we employ a universal screening approach for all resected colorectal carcinomas and biopsies from stage IV colorectal carcinomas regardless of patient age, tumor histology, and tumor location. Patients with tumors exhibiting loss of MSH2 and/or MSH6 and isolated loss of PMS2 are referred for further genetic counseling. Tumors with loss of MLH1 and PMS2 protein expression are further evaluated with BRAF mutation analysis and MLH1 promoter hypermethylation analysis before referral to genetic counseling," the authors wrote.
Contrary to some previously published data, the researchers also found that sporadic MSI-H tumors shared similar morphologies with tumors associated with probable Lynch syndrome, with the exception of a slightly higher proportion of sporadic MSI-H tumors (81%) demonstrating tumor infiltrating lymphocytes as compared with Lynch syndrome–associated tumors (61%).
The authors wrote that this is the "largest study" to date of consecutively identified Lynch syndrome–associated and sporadic tumors. As previous reports were based on retrospective data of primarily Lynch syndrome–associated tumors, those results were "vastly enriched for patients selected for analysis based on cancer-related personal or family history."
The failure of current predictive pathology models to predict all MSI-H and Lynch syndrome–related tumors, particularly left-sided ones, is a strong argument for universal screening, which "does not require widespread application of clinicopathologic criteria by clinicians and pathologists in order to select patients for testing." Further, universal screening is cost effective, and provides useful predictive and prognostic data in reducing cancer risks for patients and relatives who may or may not benefit from genetic counseling, particularly in patients over 60 years, for whom most of the MSI-H tumors were sporadic (91.5%), requiring no additional genetic counseling or testing.
Dr. Hartman and his associates reported no relevant disclosures.