Clinical Synthesis: dMMR/MSI-H Across Six Papers

The Six Papers at a Glance

I. Universal Testing: The Most Consistent Cross-Paper Recommendation

• Kwak 2026: Test all solid tumors; IHC is the first-line method when tissue is limited (needs only 4 slides); ctDNA-based MSI is research-only and not yet a clinical-grade diagnostic.
• Lee/Diaz 2016: Predicted (correctly) that MSI testing would become a pan-cancer biomarker; argued it should expand beyond CRC and endometrial cancer to all tumor types with documented dMMR rates.
• Latham 2021 (SBA): Demonstrated that SBA has a 26% dMMR rate - higher than CRC - and that 50% of Lynch syndrome cases in this series would have gone undetected without MMR testing of the SBA. Mandates universal testing for all small bowel adenocarcinomas.
• Gutierrez 2023 (NCDB): Shows that 14.2% of all US CRC is MSI-H, with specific populations at higher risk (elderly patients: ~20% MSI-H; medullary histology: 41%; mucinous: 24.6%). Testing cannot be selectively skipped in elderly patients or on the basis of histological assumption.
• Ercan/Tabori 2022 (CMMRD): Demonstrates the catastrophic consequences of missed testing - patients with Constitutional MMR Deficiency develop cancers starting in infancy; 90% cumulative cancer incidence by age 18.

II. Immunotherapy Selection: Who Benefits and How Much

1. First-line metastatic MSI-H/dMMR CRC: Pembrolizumab is FDA-approved (KEYNOTE-177); nivolumab ± ipilimumab received CRC approval in April 2025 (Kwak 2026). Offer ICI first-line without delay.
2. All advanced solid tumors with confirmed dMMR: The FDA's 2017 tumor-agnostic pembrolizumab approval applies. Tumor histology should not determine ICI eligibility - biomarker status alone should.
3. CMMRD patients with advanced cancers: ICI should be introduced early and universally, given that every CMMRD tumor is ultra-hypermutated regardless of histology.
4. Gene-specific benefit in CMMRD (Ercan/Tabori): ICI benefit was most pronounced in PMS2/MSH6 patients (5-yr OS 58% vs. 33% without ICI). MLH1/MSH2 CMMRD patients benefit more from surveillance than from ICI alone, likely because CNS tumors in this group have distinct dynamics.

III. Chemotherapy Avoidance: Consensus Across All Applicable Papers

• Stage II CRC (Lee/Diaz 2016; Kwak 2026; Gutierrez 2023): Multiple prospective trials and meta-analyses confirm 5-FU monotherapy does not benefit MSI-H stage II colon cancer. NCCN guidelines advise against adjuvant 5-FU monotherapy. The mechanism is clear: an intact MMR system is required to recognize 5-FU-induced DNA damage; dMMR cells cannot register this damage to trigger cell death.
• Stage III CRC (Kwak 2026): FOLFOX/CAPOX (oxaliplatin-containing regimens) retains benefit in MSI-H stage III disease; the detrimental effect of 5-FU monotherapy appears to be overcome by adding oxaliplatin.
• Gastric cancer (Kwak 2026; Lee/Diaz 2016): MSI-H gastric cancer appears not to benefit from standard capecitabine+oxaliplatin adjuvant regimens post-D2 resection. Retrospective Korean data showed that adding adjuvant chemotherapy in MSI-H gastric cancer actually flipped a survival advantage (HR 0.49 favoring MSI-H) into a disadvantage (HR 1.16). ESMO recommends careful consideration before using adjuvant chemo in this subset.
• Endometrial cancer (Kwak 2026): dMMR EC should NOT receive chemoradiation purely on the basis of high histologic grade. ESGO/ESTRO/ESP guidelines mandate molecular classification first - it is the p53-mutant (POLE-wt) subgroup, not dMMR, that gains from chemoradiation.
• Small bowel adenocarcinoma Stage II (Latham 2021): MMR-D stage II SBA had a recurrence rate of 18.2% vs. 88.2% in MMR-P patients - an 8-fold difference. By analogy with dMMR CRC, consideration should be given to withholding adjuvant chemotherapy in MMR-D stage II SBA.

IV. Lynch Syndrome Detection: A Shared Obligation

• Latham 2021 (SBA): 38.5% of dMMR SBAs harbor Lynch syndrome - twice the LS rate seen in dMMR CRC (19%). Critically, 50% of these LS patients had SBA as their very first cancer, and 80% had no family history of SBA. Without MMR testing, Lynch syndrome would be entirely missed in these patients and their families.
• Gutierrez 2023 (NCDB CRC): Bimodal age distribution of MSI-H (peaks in patients <30 and ≥80) reflects the dual LS (early-onset, germline) and sporadic (late-onset, CIMP-driven) etiologies, underscoring the need for different downstream workups at different ages.
• Kwak 2026: Complete algorithm for distinguishing sporadic dMMR from Lynch: MLH1 promoter methylation testing + BRAF V600E. If MLH1 is lost on IHC and methylation is absent and BRAF is wild-type, Lynch syndrome should be presumed until ruled out by germline testing.

V. Diagnostic Pitfalls: Where Testing Goes Wrong

1. Missense MMR mutations producing immunoreactive but non-functional proteins: In one Latham 2021 case, a patient with germline MLH1 c.55A>T missense variant had MSI-H SBA confirmed by MSIsensor (score 25.06) but showed intact MLH1 staining on IHC - because the mutant protein was present but non-functional. She had a prior dMMR endometrial cancer that was also IHC-negative. This patient's Lynch syndrome would have been missed by IHC alone.
2. Atypical punctate nucleolar MLH1 staining (Kwak 2026): A pattern that can be misread as retained expression (false negative) by inexperienced pathologists.
3. Subclonal MMR loss: Heterogeneous staining where some tumor cells retain expression.
4. Spatial/temporal heterogeneity (Kwak 2026): ~12% of metastatic CRC patients have MMR-discordant primary vs. metastasis. Testing the primary alone is insufficient in metastatic disease; rebiopsy and retesting of a metastatic site should be considered.

VI. Population-Specific and Equity Considerations

• Non-Hispanic White: 16.9% MSI-H
• Hispanic: 15.1%
• Asian/Pacific Islander: 12.4%
• Non-Hispanic Black: 11.3%

VII. Emerging and Neoadjuvant Frontiers

• Neoadjuvant ICI for locally advanced dMMR CRC: Phase II trials show substantial pathological responses; non-operative ("watch and wait") organ-preservation strategies are being investigated. This is not yet standard of care but represents an active and rapidly evolving area.
• ctDNA-based MSI testing: Sensitivity as low as 6% at tumor fractions below 1%. No FDA CDx approval as of 2025; ESMO recommends only when tissue is unavailable. Not reliable enough for routine clinical decision-making.

Master Summary: Seven Actionable Clinical Recommendations