Mismatch repair deficient (MSI-high) pancreatic ductal adenocarcinoma

Mismatch repair deficiency (loss of MLH1, MSH2, MSH6 or PMS2) produces microsatellite instability and a very high mutation burden with many frameshift neoantigens, which is why these tumours respond to PD-1 blockade despite the immunosuppressive stroma that defeats immunotherapy in the rest of pancreatic cancer. In the pancreas the defect is found in about 1 percent of ductal adenocarcinomas, more often in Lynch syndrome carriers, in KRAS wild-type tumours and in tumours with medullary or mucinous colloid histology, including some arising in intraductal papillary mucinous neoplasms. Testing is by immunohistochemistry for the four proteins or by sequencing-based microsatellite analysis, and a positive result should prompt germline testing for Lynch syndrome.

The tumour-agnostic approval of pembrolizumab in May 2017, based on KEYNOTE-016 and related studies, covered pancreatic cancer; the pancreatic cohort of KEYNOTE-158 showed responses in a minority of patients, lower than in colorectal or endometrial cancer, but some responses were durable. Dostarlimab received a tumour-agnostic approval in 2021 for mismatch repair deficient solid tumours after chemotherapy. Reasons for the lower response rate include misclassification by immunohistochemistry, the pancreatic stroma and lower neoantigen burden in some tumours, and chemotherapy remains the first-line standard with a checkpoint inhibitor used after progression or first line in patients unfit for chemotherapy.

Open questions are whether checkpoint inhibitors should be used first line, whether dual checkpoint blockade (as in colorectal cancer) or combination with chemotherapy improves responses, and whether frameshift neoantigen vaccines being tested in Lynch syndrome could prevent pancreatic cancer in carriers. Because the group is so small, evidence comes from baskets and case series rather than pancreatic-specific trials.