Researchers published a pan-cancer patient-derived organoid (PDO) platform in Science Advances that claims high fidelity to the original tumors and yields actionable therapeutic vulnerabilities. The work builds a dataset of 220 PDOs from 191 patients across 15 cancer types. The team reports strong concordance between organoids and parent tumors, citing 93% histopathology agreement, 80% median genomic concordance for driver mutations, and 0.85 median gene expression correlation across extended culture periods. They also report stable expression profiles over 10 passages and preservation of 85% of dominant tumor clones. Functionally, organoids retained clinically relevant behavior: in a subset deemed ineligible for PARP inhibitors under standard criteria, 58% showed substantial sensitivity when tested with talazoparib. The findings suggest organoid-based screening could identify patients who are excluded by current PARP inhibitor eligibility rules. Senior author Andrea Sboner, at Weill Cornell’s Sandra and Edward Meyer Cancer Center, described the system as practical for long-term preclinical modeling and reproducible therapeutic evaluation.