Performance of PREMM5, clinical criteria, and immunohistochemistry for MMR proteins in genetic risk assessment of Mexican patients with colorectal cancer.

Germline variants in the mismatch repair genes MLH1, MSH2, MSH6 and PMS2 are identified using clinical criteria, the PREMM5 prediction model and tumour immunohistochemistry.

In resource-limited settings, universal screening by immunohistochemistry for MMR proteins combined with clinical criteria remains the main route to germline testing in patients with colorectal cancer. Two hundred and eight patients were enrolled at two Mexican centres between September 2018 and October 2024, with a median age at diagnosis of 45.0 years and 48.6% mismatch repair-deficient tumours, germline multigene panel sequencing serving as the reference standard. A germline pathogenic variant was identified in 32.2% of patients (n = 67), of whom 77.6% (n = 52) had Lynch syndrome (30 MLH1, 14 MSH2, 6 MSH6, 2 PMS2) and 22.4% (n = 15) carried a variant in another gene (4 ATM, 3 BRCA1, 3 CHEK2, 2 TP53, 1 BRCA2, 1 BRIP1, 1 PALB2). The Amsterdam II criteria showed the highest specificity (56.0% sensitivity, 91.9% specificity), the revised Bethesda criteria (98.1% sensitivity, 19.9% specificity) and MMR immunohistochemistry (91.2% sensitivity, 68.7% specificity) the highest sensitivity, and the area under the curve for PREMM5 reached 0.822 (95% CI 0.746-0.898).

Synthesis written by Geno'X. For the full original abstract, please refer to the source publication.

The figure to remember is not the overall 32.2% detection rate, which mainly reflects tight upstream selection and does not transfer to an unselected colorectal cancer population, but the fact that one carrier in five falls outside the MMR genes. That fraction rules out a sequence of immunohistochemistry followed by a single targeted test: ATM, CHEK2, TP53 or BRCA1 will never be reached that way, although each changes surveillance and family investigation. The Amsterdam II criteria, missing 44% of carriers, incidentally confirm that referral based on family history alone is no longer sufficient, including in resource-constrained settings.

Analysis by Dr Thibaut Benquey