LBPO.BCS02 · 生物信息与计算 · Late-Breaking

Multi-omic reconstruction of metastatic prostate cancer evolution from rapid autopsy reveals polyclonal seeding and implications for liquid biopsy

编号 LB433 展板 1 时间 4/22 09:00–12:00 区域 Section 52 主讲 Azhar Khandekar
分会场 Late-Breaking Research: Bioinformatics, Computational Biology, Systems Biology, and Convergent Science 2
该海报暂无可访问的完整资料 AACR 官方页面 ↗

作者与单位

Azhar Khandekar, Manasvita Vashisth, Pushpa Itagi, Robert Patton, Mohamed Adil, Patricia Galipeau, Michael Haffner, Colm Morrisey, Peter Nelson, Gavin Ha

Fred Hutchinson Cancer Center, Seattle, WA

摘要 Abstract

Introduction: Metastatic castration-resistant prostate cancer (mCRPC) remains invariably lethal due to acquired resistance to androgen receptor signaling inhibitors (ARSIs) and taxane chemotherapy. While genomic drivers of mCRPC are well-cataloged, the spatiotemporal evolution of resistance and the precise architecture of lineage plasticity-specifically the transition from AR-positive adenocarcinoma (ARPC) to neuroendocrine (NEPC) or double-negative (DNPC) phenotypes-remain poorly understood due to the scarcity of multi-site metastatic tissue. We present a comprehensive genomic and transcriptomic atlas of end-stage mCRPC constructed via a rapid autopsy program, defining the evolutionary history of lethal disease. Methods: We performed whole-genome sequencing (WGS, 60×) on 180 metastatic tumors from 70 mCRPC patients (average 3 tumors/patient) collected through the University of Washington Rapid Autopsy Program, with matched germline and post-mortem plasma ctDNA for a subset of 58 patients. Our analytical framework incorporated: consensus somatic variant calling (VarScan, Mutect2, Strelka2); structural variant detection (SVABA, Manta, GRIDSS); copy number analysis (TitanCNA); mutational signature analysis (SigProfilerExtractor); phylogenetic reconstruction (PyClone, LICHeE); and ctDNA phenotype classification (Keraon). Results: Phylogenetic reconstruction revealed that lethal mCRPC is driven by complex, polyclonal seeding rather than simple linear progression. We identified distinct evolutionary trajectories: while ARPC metastases frequently shared late-arising subclonal drivers, NEPC tumors demonstrated early, truncal divergence characterized by simultaneous RB1 and TP53 loss. Concordance analysis between autopsy ctDNA and synchronous tumor WGS demonstrated that liquid biopsy accurately recovers truncal driver events (>90% concordance). Additionally, Keraon phenotype analysis from ctDNA revealed mixed ARPC/NEPC phenotypes in 17% of patients with heterogeneity. Conclusions: We present the largest multi-site genomic atlas of mCRPC to date, revealing that lethal resistance is driven by early branching events and maintained through extensive polyclonal seeding. Critically, we show that liquid biopsy accurately delineates complex phenotypes (including mixed ARPC/NEPC states), and recapitulates truncal drivers while also detecting spatially segregated subclonal heterogeneity, although this is mediated by tumor fraction.
利益披露 Disclosure
A. Khandekar, None.. M. Vashisth, None.. P. Itagi, None.. R. Patton, None.. M. Adil, None.. P. Galipeau, None.. M. Haffner, None.. C. Morrisey, None.. P. Nelson, None.. G. Ha, None.

在会议检索中打开