Abstract A140: Patient-derived organoids and precision medicine: Insights from the PASS-01 clinical trial in PDAC

Abstract

Pancreatic ductal adenocarcinoma (PDAC) is frequently diagnosed at an advanced stage with limited treatment options. Patient-derived organoids (PDOs) can model tumors, identify determinants of therapeutic response, and support precision medicine. In the multi-institutional PASS-01 trial of metastatic PDAC, biopsies from six US and Canadian institutions were collected for RNA sequencing, whole-genome sequencing (WGS), and PDO establishment. KRAS mutation status by droplet digital PCR was used to confirm neoplastic cultures and identify normal outgrowth. Among 186 biopsies from 183 consented and screened patients, the malignant PDO establishment rate was 50%. Establishment was more frequent in patients with rapid disease progression (p<0.001) and varied by tumor features. Success was highest in hybrid tumors (69%), compared with classical (48%) and basal-like tumors (17%), and was enriched in tumors with both KRAS and TP53 mutations. Initial PDO morphology, classified as cystic, dense, or mixed, was associated with patient progression-free survival (p=0.019). WGS showed strong concordance between matched patient tumors and PDOs, including preservation of genomic alterations and ploidy. Selected cases showed divergence, including whole-genome duplication during PDO establishment or propagation. Single-cell RNA sequencing characterized malignant epithelial sub populations and inferred copy-number clones. Across matched patient tumors, PDOs, and PDO-derived xenografts, dominant features were generally retained, although the relative abundance of transcriptional states and clones varied across models. These findings support PDO fidelity while revealing clonal and cell-state selection during model generation. A subset of PDOs was screened with RAS-directed agents, revealing heterogeneous responses, including among KRAS wild-type tumors. Transcriptional KRAS-dependency signatures varied across subtypes, inferred clones, and models, with greater predicted dependency in selected basal-like populations. These findings demonstrate substantial interpatient and intratumoral heterogeneity in RAS pathway dependence. All PASS-01 PDOs underwent high-throughput pharmacotyping with standard-of-care chemotherapies. Integration of drug-response data with molecular features enabled development of transcriptomic chemotherapy-response signatures associated with patient overall survival (p=0.012). The average time from tissue receipt to initial drug-screen results was 65 days, with six PDOs screened in fewer than 30 days. Although these results were able to be presented at monthly molecular tumor boards potentially informing second-line therapy, ongoing work is evaluating image-based and accelerated screening approaches that may inform first-line treatment. By integrating clinical outcomes, molecular profiling, PDO pharmacotyping, and cross-model analyses, PASS-01 demonstrates how PDOs can be incorporated into prospective clinical trials as a functional precision medicine platform to complement molecular profiling and support individualized treatment selection.