Clinico-genomic landscape and prognostic impact of APC/CTNNB1 oncogenic alterations in non–small cell lung cancer.

Abstract

8552 Background: Genomic activation of the WNT/β-catenin pathway via APC loss-of-function or activating CTNNB1 mutations is well established across multiple cancer types, but its prevalence and genomic context in non–small cell lung cancer (NSCLC) remain poorly characterized. We systematically characterized the prevalence of APC and CTNNB1 alterations in NSCLC, and defined their mutational landscape and prognostic impact. Methods: We conducted a multicenter cohort study including patients with NSCLC whose tumors underwent targeted exon sequencing at Dana-Farber Cancer Institute (DFCI) and Memorial Sloan Kettering Cancer Center (MSKCC). Tumors were classified as APC-mutated, CTNNB1-mutated, or WNT-mutated (APC and/or CTNNB1 mutated). Pathogenic alterations were defined as oncogenic or likely oncogenic per OncoKB, or as missense variants with a REVEL score > 0.5. Gene enrichment analyses used gene-wise logistic regression adjusted for tumor mutational burden (TMB), testing genes meeting a predefined prevalence threshold, with Bonferroni correction within each analysis. Results: In the DFCI cohort (N = 4,717), APC or CTNNB1 alterations were identified in 4.5% of NSCLC (212/4,717), including APC in 1.7% (79/4,717) and CTNNB1 in 2.8% (133/4,717), and were mutually exclusive. APC alterations were predominantly loss-of-function events (85.3% of qualifying APC variants) and splice-region/site variants (14.8%). CTNNB1 alterations were all missense mutations (100%), clustering in exon 3, consistent with β-catenin stabilization and gain-of-function. Relative to WNT–wild-type tumors, WNT-mutated NSCLC occurred more frequently in never-smokers (32% vs 24%, p = 0.03), patients of Asian race (11.8% vs 4.2%, p < 0.001), and non-squamous histology (99.5% vs 88.6%, p < 0.001). These tumors exhibited lower median PD-L1 expression (p < 0.001) and TMB (p = 0.03). Similarly, in the MSK cohort (N = 4710), the prevalence of APC and CTNNB1 alterations was 0.7% and 2.7%, respectively. Relative to WNT–wild-type tumors, WNT-altered NSCLC was significantly enriched for EGFR mutations (19.3% vs 10.5%; OR 2.18; p < 0.01) and SMAD4 mutations (7.1% vs 2.1%; OR 3.47; p = 0.0026), with relative depletion of KRAS alterations (20.8% vs 32.1%; OR 0.55; p = 0.027). External validation reproduced these findings, with WNT-altered tumors enriched for EGFR (42% vs 21.7%; OR 3.21: p < 0.001) alterations and depleted for KRAS (16.7% vs 29.4%; OR 0.47; p = 0.011). In the DFCI cohort, APC/CTNNB1 alterations were not associated with overall survival [HR 1.05 (0.88-1.25 95% CI; p = 0.57)]. Conclusions: Genomic activation of the WNT/β-catenin pathway is characterized by EGFR and SMAD4 co-alterations, relative KRAS depletion, and lower PD-L1 and TMB. These results suggest that WNT pathway alterations contribute to molecular heterogeneity in NSCLC with potential relevance for future biomarker and therapeutic studies.