Chanchan Shan, Jian Zhang, Liyuan Zhang
Immune checkpoint inhibitors (ICIs) have transformed treatment for deficient mismatch repair/high microsatellite instability (dMMR/MSI-H) colorectal cancer (CRC), but durable benefit remains uncommon in microsatellite-stable/proficient mismatch repair (MSS/pMMR) disease. Negative combination trials may reflect heterogeneous resistance programs tested without adequate biological stratification or prospective pharmacodynamic verification. This narrative and critical Review examines MSS/pMMR CRC through tissue-level immune-desert, immune-excluded, and immune-suppressed contexts, with active liver metastasis treated as a putative organ-context modifier rather than a validated predictive phenotype. We develop a hypothesis-generating, barrier-matched, pharmacodynamic (PD)-gated approach for future trials, not routine care. The workflow separates Tier 1 core stratification-liver-metastasis status, disease burden, core genomics, and baseline plasma circulating tumor DNA (ctDNA)-from Tier 2 tissue-based immune-spatial assessment required for mechanism-matched allocation to most non-hepatic barriers. Therapeutic modules are aligned with the dominant resistance context and reassessed after short induction using prespecified GO/ADAPT/STOP rules. For spatial or tissue-based mechanisms, GO requires a barrier-specific readout when tissue assessment is feasible; ctDNA is a complementary systemic-response signal and cannot by itself establish barrier engagement. Mixed or lesion-discordant phenotypes and inadequate tissue remain exploratory unless a prespecified dominance rule is met. A prototype master protocol would combine mechanism-defined cohorts, contemporaneous controls where feasible, early barrier-specific PD conversion as the primary biological endpoint, 6-month progression-free survival as the primary clinical endpoint, and prespecified expansion, futility, safety, and adaptation rules. Prospective validation is needed before this approach can inform practice.