ID #676 Precision-Guided Oncology for Pediatric Brain Tumors Using Functional Ex Vivo Drug Sensitivity Profiling
Gabor Tax, Jie Mao, Luke Fuller, Loretta M S Lau, Chelsea Mayoh, Roxanne Cadiz, Anthony Zhou, Mehrin Zaman, Nicholas Thomson, Jinhan Xie, Paul G Ekert, David S Ziegler, Emmy M DolmanAbstract
Functional precision medicine programs, including ZERO, combine tumor molecular profiling with ex vivo drug sensitivity profiling (DSP) to guide personalized therapy recommendations for children with cancer, including brain tumors. However, the clinical impact of integrating ex vivo tumor DSP into pediatric precision medicine programs remains unclear, and responses to personalized therapies are difficult to predict. This study aims to define the clinical utility of ex vivo tumor DSP and to characterize the pharmacogenomic landscapes of childhood cancers to uncover novel tumor-specific vulnerabilities.
Ex vivo DSP was performed on 236 ZERO patient samples, including 68 brain tumors, using pediatric-tailored drug libraries. A novel automated data analysis pipeline was established to generate drug efficacy parameters, including area under the dose-response curve (AUC) values, half-maximal inhibitory concentrations (IC50s), and half-maximal lethal concentrations (LC50s). Ex vivo efficacies were correlated with matched molecular profiling, in vivo response, and clinical outcome data.
DSP validated known pharmacologic vulnerabilities, including sensitivity of neurotrophic tyrosine receptor kinase (NTRK) fusion–positive brain tumors to NTRK inhibition. Importantly, patients treated with therapies identified as ex vivo DSP hits demonstrated significantly improved objective clinical benefit (p = 0.036) and progression-free survival (p = 0.035) compared with those receiving non-hit therapies, with a similar trend observed in brain tumor patients specifically (p = 0.073). Interestingly, mitogen-activated protein kinase kinase (MEK) and Bruton’s tyrosine kinase (BTK) inhibitors demonstrated efficacy in high-grade glioma samples lacking canonical driver pathway alterations, and MEK inhibitor sensitivity strongly correlated with JAK inhibitor response. Furthermore, LC50s provided approximately twofold greater predictive power for in vivo responses compared with IC50s.
Together, this study represents the largest effort to date demonstrating that integration of functional ex vivo tumor DSP into pediatric precision medicine programs improves clinical outcomes and identifies novel therapeutic vulnerabilities, expanding treatment options for children with cancer.