MSH2

It is the second half of the argument for unselected germline testing, and it widens the target beyond BRCA: two thirds of the actionable inherited findings in prostate cancer are in other genes, including the mismatch repair genes that open a checkpoint inhibitor route.

It is the argument for sequencing every man with advanced prostate cancer rather than only the ones who look high risk: the phenotype is uncommon, it is invisible clinically, it opens the only durable immunotherapy route in this disease, and in one man in five it also identifies Lynch syndrome in the family.

The germline finding is the practical lesson: a meaningful minority of biliary cancer patients carry inherited repair-gene mutations that matter for PARP inhibitor eligibility and for their relatives, which argues for germline testing alongside tumour sequencing.

Guidelines recommending universal mismatch repair testing in pancreatic cancer, by immunohistochemistry or sequencing rather than PCR alone, rest on this and similar series.

It gives a cheap way to find the rare men who could benefit from checkpoint blockade: an immunohistochemical stain on the highest-grade primary tumours, where the yield is twenty times higher than average. It also shows the loss is clonal and early, so the diagnostic block is an adequate place to look.

It matters for how the test is done. A mismatch repair assay designed around the colorectal mechanism, looking for MLH1 promoter methylation and simple mutations, can miss the prostate mechanism entirely, which is one reason microsatellite instability was under-recognised in this disease for so long.

It is the evidence behind universal mismatch repair testing of every colorectal cancer, which is now standard in most guidelines and which, as a by-product, identifies everyone eligible for immunotherapy.

EPCAM deletion testing is now part of Lynch syndrome panels, and it explains the MSH2-deficient tumours in families where sequencing of the four repair genes comes back clean.