Genomic landscape and clinical outcomes of MDM2 amplifications in biliary tract cancers.

589 Background: Biliary tract cancers (BTCs), including intrahepatic cholangiocarcinoma (iCCA), extrahepatic cholangiocarcinoma (eCCA), and gallbladder cancer (GBC), are molecularly heterogeneous malignancies with limited effective treatments. MDM2 may represent a therapeutic target in BTCs, but the frequency and prognostic impact of MDM2 amplifications are not well defined. Methods: Under an IRB-approved, prospective genotyping protocol (NCT01775072), patients with histologically confirmed BTC underwent next-generation sequencing using the FDA-authorized, MSK-IMPACT platform at Memorial Sloan Kettering Cancer Center. A cohort of 94 MDM2 amplified (MDM2 amp ) samples were extracted. A matched cohort of an additional 92 patients was identified for comparison by using exact matching on stage and nearest matching on sex, subtype, and age at diagnosis from a MDM2 wild-type cohort. Clinicopathologic parameters were retrospectively extracted to (1) assess fold change of MDM2 (2) and compare co-occurring genomic and pathway alterations (3) and outcome. Results: The MDM2 amplified cohort included 50% female patients and 27% early (I/II) stage disease, with 30% eCCA, 36% GBC and 33% iCCA cases and a median age of 64 (range 25-89). Stratification by MDM2 log2 fold change quartiles demonstrated no significant differences in overall survival from date of BTC diagnosis or start date of first line or second line treatment. The median overall survival from date of diagnosis in the amplified cohort was 30 months (95% CI 21-37). Multivariate cox regression confirmed stage IV patients had significantly worse outcomes (p-value < 0.001) compared to patients with early-stage disease. Comparative analysis with a matched BTC cohort revealed that the MDM2 amp group was enriched for ERBB3 (27% vs 5% q-value 3e-3) and CDK4 (24% vs 0% q-value 5e-6) alterations. The MDM2 wild-type group showed significantly higher frequencies of TP53 (53% vs 12% q-value 2e-7) and ERBB2 (36% vs 3% q-value 7e-7) alterations. Despite genomic differences, overall survival showed no significant difference between the wild-type and amplified groups with a median survival of 31 (95% CI 25-45) and 30 (95% CI 21-37) months respectively. OncoKB level 1 and 2 actionable alterations were significantly more prevalent in the wild-type cohort (62% vs 10%). Conclusions: MDM2 amp tumors demonstrate distinct genomic characteristics, including enrichment for ERBB3 and CDK4, likely caused by their proximity to MDM2, and significantly fewer alterations in ERBB2 and TP53. MDM2 amp samples lack established level 1/2 actionable targets, suggesting a favorable context for precision therapeutic strategies despite no significant outcome differences from wild-type. Ongoing work includes a comprehensive comparison of the TP53 alterations and fold change of CDK4 and ERBB3 in both groups.