Abstract Drug development is notoriously inefficient, slow, and expensive. In oncology, only 5% of new compounds identified during the drug discovery phase successfully progress to clinical trials. This high attrition rate is attributed to patient intra- and inter-tumor heterogeneity, as well as an over-reliance on 2D cell cultures and animal models, which often fail to accurately replicate patient tumor biology and physiology. The adoption of advanced, clinically relevant preclinical models at earlier stages of drug development is essential for better identifying and prioritizing agents with a higher likelihood of success in clinical trials. The Human Cancer Models Initiative (HCMI) is a global initiative founded by the National Cancer Institute (NCI), in collaboration with the Cancer Research UK, Wellcome Sanger Institute, and the foundation Hubrecht Organoid Technology. The mission is to generate patient-derived next-generation cancer models from diverse tumor types as a community resource. Unlike traditional cancer models, the new models are cultured under optimized, predominantly 3D conditions that better preserve the characteristics of the parental tumors than historical culture conditions. This preservation is validated through phenotypic and molecular analyses of tumor tissue and models, which are shared with the community alongside standard operating procedures, informed consent templates, and associated clinical and molecular data. To contribute towards the goal of the HCMI, several Cancer Model Development Centers (CMDC) have been generating cancer models from a variety of tumor types. This includes the Stanford CDMC which is currently dedicated to models of pediatric solid tumors, emphasizing central nervous system (CNS) tumors, the leading cause of cancer-related death in children. We developed a standardized bioprocessing pipeline which yielded a functional tumor bank, achieving a 60-70% success rate in establishing next-generation cancer models for long-term passaging. We have successfully generated and submitted 85 pediatric cancer models along with case-associated clinical and biospecimen data, as well as internal QC data validating the derived cancer models, for further characterization and distribution via the HCMI pipeline. Our next-generation cancer models partially capture the heterogeneity of pediatric CNS tumors, neuroblastoma, hepatoblastoma, Wilms tumor, and brain metastases from neuroblastoma and rare sarcoma-related cancers. Longitudinal biobanking has identified and characterized novel onco-fusion proteins, rare tumor entities, and recurrences, and therapeutic vulnerabilities through multi-omics. In summary, next-generation cancer modeling overcomes known challenges by benchmarking models and creating standardized protocols and procedures, thereby enhancing their predictive capabilities, therapeutic efficacy and promoting personalized treatment strategies. Citation Format: Emon Nasajpour, Dena Panovska, Ruolun Wei, Conrado Soria, Calvin J. Kuo, Rachana Agarwal. Next-generation models to advance pediatric solid cancer treatments abstract. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 6166.
Nasajpour et al. (Fri,) studied this question.