Whole-genome sequencing of biliary tract cancer: Uncovering the genomic origins of evolutionary trajectories.

Abstract

4116 Background: Biliary tract cancers (BTC) are rare, highly aggressive malignancies with limited treatment options, leading to consistently poor outcomes. An improved understanding of BTC tumor evolution could inform enhanced screening strategies, identify useful prognostic markers, and discover novel therapeutic targets. Methods: We performed whole genome and transcriptome sequencing (WGTS) at high depth ( > 80X) on a prospective cohort of BTC tumors. After detecting mutations and mutational signatures, we developed two novel methods for driver identification and evolutionary reconstruction. First, we created new oncogene and tumor suppressor-specific models that integrate copy number profiles, structural variant breakends, and expression changes to distinguish structural drivers from neutral chromosomal rearrangements. Second, we applied population genetics techniques to fit demographic models to tumor allele frequencies across 20 paired primary and metastatic samples. Results: We analyzed 130 tumor samples from 110 patients, representing the largest BTC whole-genome cohort to date. We identified hypermutated tumors (50-150 mutations/mb) with distinct etiologies, including mismatch repair deficiency, platinum exposure, tobacco use, and aristolochic acid-related damage, the latter of which was associated with response to immunotherapy. Ourintegrated-driver approach identified an association between selection on RAD23A and an increased structural variant load, resulting in a tandem-duplicator-like mutational phenotype. This also revealed an underappreciated impact of SMAD4 in BTC, which is inactivated through multiple mutation types in 12%. Notably, BAP1 mutations occurred in 24% of cases, including 4% that were inactivated through deletion of the BAP1 promoter that lowered transcript expression, a previously undescribed mechanism. By pairing clinical data to the genomics, we observed a co-occurrence of BAP1 mutations and FGFR2 fusions in small-duct, mass-forming intrahepatic cholangiocarcinomas. These mutations were mutually exclusive with TP53 mutations, which were enriched in patients with primary sclerosing cholangitis. Finally, our novel method for subclonal population reconstruction on paired samples illuminated recent tumor evolutionary dynamics and identified an ARID1B fusion with a potential role in metastasis. Conclusions: This study uncovered novel genomic mechanisms underlying the evolutionary origins of BTC beyond those previously identified with exome and panel sequencing, highlighting the value of WGTS. These results highlight the complex genomic heterogeneity of BTC, with potential implications for precision therapy.