Reviews/Perioperative Immunotherapy Solid Tumours/Microsatellite instability-high colon cancer immune checkpoint inhibitor surrogate biomarker
Msi Colon · PERIO-MSI-02 · v2
Borderline: within 4 points of the Established/Likely boundary, so a re-grade could land either side.
In early-stage colon cancer with a faulty DNA mismatch-repair system (dMMR/MSI-H), giving immunotherapy before surgery shrinks or clears the tumour in a remarkably high share of patients. Separate studies run by different teams found that roughly two-thirds of these patients had no viable tumour left at surgery [13][17][3], and real-world use produced broadly similar results [10][12]. By contrast, tumours with a working repair system did not respond on the same terms and needed added radiotherapy and chemotherapy to reach a comparable result [5]. The panel judged that this consistent, repeated benefit in the biomarker-positive group, together with biomarker-confined benefit in advanced disease [7], supports the idea that the marker predicts who responds. It stopped short of the top tier because no single trial directly compared marker-positive against marker-negative patients, and because the trials measured tissue response at surgery rather than long-term survival.
Across four independent groups, neoadjuvant immune checkpoint inhibition in dMMR/MSI-H early-stage colon cancer produced exceptional pathological responses: 68% pathological complete response in NICHE-2 [13] and NICHE-3 [17], 65% in resected patients on pembrolizumab [3], and 64% major response with 42% complete response in a real-world cohort [10]. A retrospective series reported 73.4% complete response in colorectal patients [12]. In mismatch-repair-proficient disease, TORCH-C [5] reached a 45.3% complete response only by adding short-course radiotherapy and chemotherapy to immunotherapy, versus 10% with chemotherapy alone. Metastatic phase 3 data confirm checkpoint benefit confined to the MSI-H/dMMR subtype [7]. The contested point: The division is whether these data demonstrate a predictive biomarker or merely an enrichment biomarker: every pivotal neoadjuvant trial enrolled only dMMR/MSI-H patients [13][17][3][10][12], so differential benefit is inferred from cross-trial contrast with [5] rather than from any within-trial marker-positive versus marker-negative comparison. There is also tension in [5], where proficient-repair tumours did respond once the regimen was intensified.
All 23 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Perioperative Immunotherapy Solid Tumours, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
Every pivotal neoadjuvant trial — NICHE-2 [13], NICHE-3 [17], Ludford [3], AGEO [10], Wang [12] — enrolled ONLY dMMR/MSI-H patients. They establish high response within the biomarker-positive group but contain no within-study pMMR comparator, so they do not formally demonstrate differential benefit, which is the defining property of a predictive (versus enrichment/selection) biomarker in this neoadjuvant early-stage setting.
Primary endpoints are surrogate (pathological complete/major response), not long-term disease-free or overall survival. Follow-up remains immature in the strongest trials (NICHE-3 [17] median 8 months; NICHE-2 [13] median 26 months), so the biomarker's predictive value for durable clinical benefit is not yet established.
Where it remains uncertain. No trial compared marker-positive against marker-negative patients directly, and proficient-repair disease still reached 45.3% response with an intensified regimen [5].
Aleatory
Even in biomarker-positive disease, response is not universal; NICHE-2 [13] and NICHE-3 [17] each had 2 of 59-111 patients without pathological response, and progression events occurred in [3].
Epistemic
Follow-up is immature (NICHE-3 median 8 months [17]; NICHE-2 median 26 months [13]), so whether pathological response translates to long-term disease-free or overall survival is not yet established in this corpus.
The load-bearing papers from a corpus of 23, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Prospective phase II neoadjuvant pembrolizumab in localized MSI-H/dMMR tumors; concordant high response in the biomarker-positive group.
“Among 33 evaluable patients, best overall response rate was 82%. Among 17 (49%) patients who underwent surgery, the pathologic complete response rate was 65%.”
Panel auditInternal validity: medium · scope of inference: Neoadjuvant pembrolizumab response in localized MSI-H/dMMR solid tumours, predominantly colorectal.Key limitations: single-arm, no pMMR contrast · mixed solid tumours not colon-only · small sample
Supports the predictive (not merely prognostic) reading: biomarker-negative disease is the setting where ICI progress has lagged and required intensified regimens.
“Immunotherapy has become the standard treatment for deficient mismatch repair or microsatellite instability-high (dMMR/MSI-H) colon cancer. However, there has been little progress for the majority of cases, which was mismatch-repair proficient or microsatellite stable (pMMR/MSS).”
Panel auditInternal validity: medium · scope of inference: Neoadjuvant chemoradiation plus PD-1 inhibitor versus chemotherapy in high-risk pMMR/MSS locally advanced colon cancer.Key limitations: pMMR/MSS only, no MSI arm · regimen adds radiotherapy, not ICI alone · conference abstract
Randomized phase-3 confirmation that ICI benefit tracks MSI-H/dMMR status (metastatic setting), the biological basis extended to the neoadjuvant claim.
“Pembrolizumab led to significantly longer progression-free survival than chemotherapy when received as first-line therapy for MSI-H-dMMR metastatic colorectal cancer, with fewer treatment-related adverse events.”
Panel auditInternal validity: high · scope of inference: First-line pembrolizumab versus chemotherapy in metastatic MSI-H/dMMR colorectal cancer.Key limitations: metastatic not neoadjuvant · within-MSI-H comparison · no pMMR contrast
3 of 23 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Exceptional, independently replicated response in the biomarker-positive population plus a documented pMMR contrast strongly support a predictive role, but the absence of a within-trial pMMR comparator and reliance on surrogate pathological endpoints hold the verdict below Established.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Within-trial marker-positive versus marker-negative comparison. Every pivotal neoadjuvant trial [13][17][3][10][12] is biomarker-enriched, so differential benefit, the defining property of a predictive biomarker, is never shown within a study. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: A neoadjuvant trial enrolling both dMMR/MSI-H and pMMR/MSS early-stage colon cancer on the same checkpoint regimen, with marker status as a prespecified subgroup.
Open question: Does dMMR/MSI-H status predict differential neoadjuvant checkpoint benefit when marker-positive and marker-negative patients are treated and compared within the same trial? Why it is still open: All pivotal neoadjuvant trials in the corpus [13][17][3][10][12] enrolled only biomarker-positive patients, leaving no within-study comparator.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19