Beyond TMB: Characterizing tumor-intrinsic genomic architectures of immune resistance.

2586 Background: Immune checkpoint inhibitors (ICI) are increasingly selected based on tumor mutational burden (TMB); however, clinical benefit remains inconsistent. We hypothesized that tumor-intrinsic genomic alterations associated with immune suppression, immune exclusion, and antigen presentation failure frequently coexist with ICI-enabling biomarkers and define distinct resistance architectures. Methods: We analyzed tissue-based genomic profiling data from 7,773 solid tumors. TMB-high (TMB-H) tumors were evaluated for genomic alterations implicated in ICI resistance, including PTEN loss-of-function (LOF), STK11 and KEAP1 LOF, WNT /β-catenin pathway alterations ( CTNNB1 activating mutations, APC truncation/biallelic loss), B2M LOF (antigen presentation), JAK1/2 LOF (interferon signaling), and MDM2/MDM4 amplification. Results: Among 908 TMB-H tumors, 30.6% harbored at least one immune resistance–associated alteration. PTEN LOF was the dominant functional resistance event (7.6%) and represented the central hub in tumors with multiple resistance alterations (65.6%; 21/32). WNT pathway alterations were frequent, with APC truncation (15.9%; 144/908) exceeding CTNNB1 mutations (2.6%; 24/908), indicating heterogeneous modes of WNT activation. Importantly, WNT -driven tumors demonstrated divergent immune escape architectures: CTNNB1 -mutant tumors were enriched for B2M LOF (16.7%; 4/24), whereas APC -mutant tumors rarely harbored B2M loss (0.7%; 1/144), suggesting antigen presentation-dependent versus immune-exclusion-dominant resistance, respectively. Across the TMB-H cohort, increasing mutational burden was significantly associated with higher B2M LOF frequency (Wilcoxon rank-sum p = 3.6 × 10⁻⁶), consistent with immune editing under high neoantigen pressure. Interferon signaling alterations were rare ( JAK1 : 0.2%; 2/908; JAK2 : 0.0%; 0/908). Conclusions: TMB identifies immune pressure but not immune competence. Intrinsic ICI resistance in TMB high tumors is structured around a PTEN-WNT-B2M genomic axis, with distinct immune escape architectures determined by the mode of WNT activation. Integrating negative genomic predictors of immune response with established biomarkers may improve immunotherapy stratification and inform development of rational combination strategies. Prevalence of genomic alterations associated with intrinsic ICI resistance in TMB-high solid tumors. ICI resistance marker Prevalence in TMB-High Tumors STK11 2.4% KEAP1 0.3% PTEN 7.6% CTNNB1 2.6% APC 15.9% B2M 2.4% JAK1 0.2% JAK2 0.0% MDM2 2.5% MDM4 0.4%