Site-Dependent Decoupling of Drug-Biomarker Associations in Clear Cell Renal Cell Carcinoma Revealed by Functional Profiling of Patient-Derived Cell Models

Clear cell renal cell carcinoma (ccRCC) frequently exhibits primary and acquired resistance to standard-of-care therapies, highlighting the need for improved understanding of treatment failure and more effective, patient-specific, therapeutic strategies. Although recent multi-omic and single-cell atlases have provided detailed insight into the molecular landscape of ccRCC, translating these discoveries into individualized treatment remains challenging, in part because molecular alterations alone often incompletely predict therapeutic response. To bridge this gap, we prospectively profiled a cohort of 28 patients with localized and metastatic ccRCC by integrating comprehensive molecular characterization with functional drug screening in patient-derived cell (PDC) models, and longitudinal clinical data. Functional drug profiling identified recurrent sensitivities to selected kinase inhibitors, apoptotic modulators, and metabolic regulators across subsets of PDCs, alongside patient-specific vulnerabilities. We identified putative actionable therapies in 27/28 patients (96%) based on genomic biomarkers previously described in ccRCC, other renal and/or non-renal cancers. However, integration with functional data revealed substantial discordance between genomic actionability and ex vivo drug sensitivity. Linear mixed-effects modelling identified 16 novel copy-number-based features associated with sensitivities to 11 drugs. Importantly, genotype-drug response associations were largely preserved between primary tumors and vena cava (VC) thrombi but frequently disrupted in distant metastatic samples, suggesting site-specific evolutionary decoupling of genomic alterations and therapeutic phenotype. Together, these findings demonstrate that integrating functional drug testing of PDCs with multi-omic profiling refines therapeutic actionability in ccRCC, revealing vulnerabilities not apparent from genomic data alone. This functional precision oncology framework provides a scalable strategy to complement molecular profiling, account for interpatient and intersite heterogeneity, and support hypothesis-driven, patient-specific treatment prioritization. SignificanceFunctional drug screening of PDCs reveals context-dependent vulnerabilities in ccRCC. Copy number-driven genotype-drug associations are preserved locally but frequently lost in metastases. This scalable framework refines actionability for guiding treatment prioritization.