Accelerating and de-risking late pre-clinical drug development with the zebrafish tumor xenograft (ZTX) platform.

e15111 Background: Insufficient drug efficacy due to patient response heterogeneity is the main reason why 19 of 20 oncology drug candidates fail during clinical development. To improve success during clinical development, pre-clinical studies should investigate the need for, and approach to, personalized therapy. Importantly, the personalization approach should be clinically actionable and later included in the clinical trial design. While patient-derived organoid or xenograft models in mice are often used in pre-clinical investigations of patient response heterogeneity, these models are not clinically relevant due to low success rates for establishing such models within certain indications, long assay times, and prohibitively high costs. The zebrafish tumor xenograft (ZTX) model has emerged as an innovative and customizable alternative, enabling the individualized assessment of patient responses to novel drug candidates within just five days, with a sensitivity and specificity similar to mouse PDX models, and at a cost similar to or lower than NGS analyses or other molecular precision medicine approaches. Furthermore, ZTX models recapitulate the invasive and metastatic potential of patient tumors, offering unique opportunities to effectively investigate the anti-metastatic property of novel drug candidates in vivo. Methods: As a demonstration of the benefit of using ZTX models during late pre-clinical drug development, this study evaluated the anti-tumor efficacy of novel drug candidates, alone or in combination with other oncological therapies, on xenografts generated from established cell lines, patient-derived cell lines, and primary clinical tumor samples using the ZTX platform. Results: The results showed that the ZTX platform effectively identified a test compound that synergized with a standard of care to inhibit prostate cancer cell metastasis, demonstrating the superior efficacy of this clinical-stage compound compared to standard therapies and other emerging prostate cancer therapies. Additionally, glioblastoma (GBM) patient-derived cell lines showing poor engraftment rates in mice, all engrafted in the ZTX platform, and the ZTX models accurately identified the responders and non-responders to standard treatments used in the clinic. Furthermore, using clinical samples from colorectal cancer patients, the ZTX platform identified responders and non-responders to a novel test compound, which together with molecular characterization of the samples can provide critical insights into potential avenues for patient stratification. Conclusions: These findings highlight the ZTX platform as a robust tool for evaluating inter-patient response heterogeneity, anti-metastatic potential, and clinical patient stratification strategies, for personalized use of novel oncology therapies.