Abstract A053: CURE AI predicts immunotherapy treatment benefit for SHH-activated/TP53 wild-type but not SHH-activated/TP53 mutant medulloblastomas

Abstract Clinical trials are not able to study individual treatment effects as methods to do so have not yet matured into practical use. For decades, we have studied groups of patients enrolled on a trial, comparing groups of patients as large cohorts and losing signal for important, complex features that define individuals. Foundation models take a different approach. Using the CURE AI foundation model generated from clinical and multi-omics data from hundreds of thousands of patients, complex, non-linear biological patterns can be found in new datasets that can provide insights into disease biology. We previously identified a complex signature predictive of immunotherapy response relative to chemotherapy response by analysis of a set of non-small cell lung cancer clinical trials (Weiss et al., AI in Precision Oncology, 2025). In the current study, we theorized that we could apply predictors of immunotherapy response/nonresponse from adult clinical trials to pediatric cancer patients. If feasible, this approach could dramatically accelerate progress in pediatric cancer treatment development by leading to the development of better treatment approaches that could immediately be assessed on pediatric clinical trials. We analyzed RNA sequencing from over 50 types of pediatric cancers to predict treatment benefit of immunotherapy relative to chemotherapy. Among all pediatric cancers, SHH-activated/TP53 wild-type medulloblastomas demonstrated among the highest predicted immunotherapy benefit relative to chemotherapy (strong benefit to immunotherapy). Interestingly, SHH-activated/TP53 mutant medulloblastomas demonstrated among the lowest predicted responses to immunotherapy (strong benefit to chemotherapy instead of immunotherapy). SKP2, UBE4A, and CUL4A genes were found to be strongly associated with differences among TP53 wild-type and mutant SHH-activated medulloblastomas, but interestingly these genes were not independently associated with differences in predicted treatment response. Using CURE AI, IFNGR1, DOCK2, and PLEK were found strongly associated with predicted treatment response in both types of medulloblastoma (i.e., these genes are enriched in patients with high predicted immunotherapy response), suggesting that high expression of these genes is predictive of immunotherapy response in SHH-activated medulloblastoma. Notably, PTCH2 was significantly upregulated in immunotherapy resistant (nonresponsive) patients in both types of SHH-activated medulloblastomas, suggesting that targeting Notch signaling could overcome immunotherapy resistance in this subset of medulloblastoma patients with CURE AI-predicted immunotherapy resistance. These results demonstrate that foundation models can be used to analyze adult clinical trials to gain novel insights into different cancer types and form the basis for patient-level data generation that could accelerate clinical trial concept development in rare diseases. Citation Format: Vitalay Fomin, Ofir Landau, Daniel Khankin, Tal Shor, Tzvi Lederer, Gregory Koushnir, Amit Weiss, Neil Pfister. CURE AI predicts immunotherapy treatment benefit for SHH-activated/TP53 wild-type but not SHH-activated/TP53 mutant medulloblastomas [abstract]. In: Proceedings of the AACR Immuno-Oncology Conference (AACR IO): Discovery and Innovation in Cancer Immunology: Revolutionizing Treatment through Immunotherapy; 2026 Feb 18-21; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Immunol Res 2026;14(2 Suppl):Abstract nr A053.