Preoperative MMR testing in colorectal cancer: Large-scale cohort confirms reliability but identifies tumor heterogeneity as a challenge.

BACKGROUND

Although previous studies have indicated that preoperative biopsy is reliable for MMR testing in colorectal cancer, they were limited by small sample sizes and interpretation based on traditional binary criteria.

METHODS

This retrospective cohort of 5307 CRC patients with paired biopsy-resection specimens underwent a two-tier re-review: first with NCCN/CAP criteria, then with an optimized integrated interpretation standard, plus retesting of indeterminate cases.

RESULTS

Concordance between biopsy and resection was 99.6% and the kappa for binary MMR status was 0.984 (95% CI 0.972-0.996). Non-definitive interpretations comprised 0.13% definite discordance and 0.28% equivocal (MMRe, defined as weak expression or intraglandular heterogeneity), mostly due to these atypical patterns. The incidence of non-definitive results was 0.60% (8/1336) after neoadjuvant therapy vs 0.35% (14/3971) after direct surgery (OR 1.70, 95% CI 0.71-4.06); the small number of events precluded definitive comparison.

CONCLUSIONS

Preoperative biopsy is highly reliable for MMR assessment in this large-scale CRC cohort. Tumor heterogeneity (i.e. weak staining and intraglandular heterogeneity) is the major contributing factor to discordant/equivocal interpretations. Adoption of refined criteria improves interpretation of atypical MMR patterns. Neoadjuvant therapy does not appear to substantially increase unreliable interpretations, and rare therapy-associated alterations warrant awareness.