Racial variation in the frequency and genotype–phenotype correlations of MLH1, MSH2, MSH6, and PMS2 gene mutations; and performance of PREMM1,2,6 in ethnically diverse individuals: A cross-sectional study

Purpose: To explore ethnic and racial variations in the frequency, genotype–phenotype correlations, and performance of the PREMM1, 2, 6 model in a large, ethnically diverse cohort of individuals undergoing DNA mutational analyses in the United States for germline mismatch repair genes associated with Lynch syndrome. Methods: Data from 52,758 individuals with personal and/or family history suggestive of Lynch syndrome who submitted blood samples for genetic testing of MLH1, MSH2, MSH6, PMS2, and EPCAM to Myriad Genetic Laboratories between July 2006 and June 2013, were analyzed. Patients were identified as Caucasian, Latin American, African American, Asian, or Other and only those with report of one race or ancestry were eligible. Statistical tests included Chi-square, Fisher’s exact test, analysis of variance (ANOVA), and receiver operator characteristic (ROC) with area under the curve (AUC). Results: A total of 37,665 subjects were eligible for analyses with the majority of subjects being Caucasian (29,771/37,665; 79%). Prevalence of MMR genes: Among subjects of all ethnicities, MLH1 (1036/2957, 36%) and MSH2 (1243/2957, 43%) were the most commonly mutated genes. Non-Caucasian mutation carriers had a higher proportion of MLH1 gene mutations, while the proportion of MSH2, MSH6 and PMS2 mutations was higher in Caucasian carriers (P\u3c0.0001). The detection of VUS was highest among non-Caucasians overall and for each of the MMR genes compared to Caucasians. Genotype-Phenotype correlation: Non-Caucasians more often reported a personal history of CRC than Caucasians and were younger at diagnosis. Overall, MLH1 mutation carriers reported a higher frequency of CRC (75%) compared with MSH2 (64%), MSH6 (52%) mutation carriers, followed by PMS2 (61%), across all racial/ethnic group. Conversely, extracolonic Lynch syndrome associated malignancies, including endometrial cancer, were more frequently reported in Caucasians and were associated with higher rates of MSH2 and MSH6 gene mutations. There were very few PMS2 and EPCAM mutation carriers overall and no statistical significance was observed between all carriers of diverse racial/ethnic groups. A similar burden of first degree relative family history of CRC was reported across racial/ethnic groups but non-Caucasians reported a lower frequency of first degree relative family history of endometrial cancer and other Lynch syndrome cancers. Non-Caucasians reported a lower frequency of second degree relative family history of CRC (P value\u3e 0.05). PREMM1,2,6: The PREMM1,2,6 had an overall c-statistic of 0.8252, performing equally well among non-Caucasians with a range between (0.81-0.84). Conclusion: This study provides novel data on the frequency and genotype-phenotype correlations for MMR gene mutation carriers among a large, racially and ethnically diverse population. Selection bias and changes in testing technologies over time may partially explain the differences in prevalence rates seen between the different ethnic groups, as well as the phenotypic expression of specific gene alterations. The PREMM1,2,6 model has excellent discriminatory capacity among all ethnic and racial groups and these results may reassure healthcare professionals that it is a useful strategy to identify individuals at risk for Lynch syndrome, regardless of race or ancestry