Lynch Syndrome and Somatic Mismatch Repair Variants in Pancreas Cancer.

Importance: Microsatellite (MS) instability (MSI-H) occurs frequently in Lynch syndrome (LS)-associated tumors and is associated with response to immune checkpoint blockade (ICB) therapy. MSI-H is conferred by germline or somatic variants in mismatch repair genes. The contribution of somatic oncogenesis to MSI-H in pancreatic cancer (PC) is unknown. Objective: To evaluate an LS-related PC cohort to define clinicogenomic features, describe somatic MSI-H cases (germline negative), characterize response to ICB, and guide preferred MS testing methods. Design, Setting, and Participants: This single-institution, retrospective analysis was conducted from March 2012 to July 2023 at Memorial Sloan Kettering Cancer Center and included 55 patients with PC and either an LS germline pathogenic variant (gPV) or somatic mismatch repair (MMR) variant. Main Outcomes and Measures: Composite MMR and MS status determined using orthogonal methods. An artificial intelligence classifier was used to account for low-cellularity specimens. Demographic and clinical data were abstracted from medical record. Zygosity status and somatic comutation landscape analyzed. Results: Fifty-five patients (23 women [42%]) had PC and an MMR variant: 32 (58%) had LS (LS cohort) and 23 (42%) had a somatic MMR variant (no germline pathogenic variant, somatic MMR cohort). In the LS cohort, 10 (31%) had gMSH2, 9 (28%) gMSH6, 8 (25%) gPMS2, 4 (13%) gMLH1, 1 (3%) gEPCAM. The median age at diagnosis was 68 years (range, 45-88 years). For composite MS status, 17 (59%) were MSI-H, 12 (41%) MS stable, and 3 MS unknown. Five cases were reclassified as MSI-H by the artificial intelligence classifier. In the somatic MMR cohort, 11 (48%) had MSH6, 7 (30%) MLH1, 3 (13%) MSH2, and 2 (9%) PMS2. The median age at diagnosis was 72 years (range, 66-85 years). For composite MS status, 10 (43%) were MSI-H, 11 (48%) MS stable, and 2 (9%) MS indeterminate. Six cases were reclassified as MSI-H by the artificial intelligence classifier. For the LS and somatic MMR cohorts, 20 received ICB (n = 17 MSI-H). The median ICB duration was 27.7 months (95% CI, 11.5 to not reached); the disease control rate was 80%. Conclusion: The results of this cross-sectional study suggest that MSI-H occurs due to LS or somatic oncogenesis in PC. Orthogonal MS testing is key in PC; the artificial intelligence classifier reclassified approximately 20% of cases, most of which were low cellularity. ICB for patients with LS or somatic MSI-H PC provided significant benefit.