AI platform identifies pancreatic cancer patient subgroup with threefold Increase in survivability | 2026 ASCO Gastrointestinal Cancers Symposium

BullFrog AI Holdings, Inc. has utilized its proprietary bfLEAP platform to analyze clinical trial data from pancreatic adenocarcinoma patients, uncovering critical biomarkers linked to a nearly threefold enhancement in overall survival rates for particular subgroups receiving glufosfamide according to a recent news release. This post-hoc analysis of a phase 3 study highlights the potential of AI-driven precision oncology to enhance patient classification and improve therapy results, tackling enduring issues in one of the most lethal tumors.

Pancreatic adenocarcinoma continues to be a significant challenge in oncology, with a five-year survival rate of less than 10% and few therapeutic alternatives beyond conventional chemotherapy. In partnership with Eleison Pharmaceuticals, Nikolas Naleid, M.D., Pharm.D., from the H. Lee Moffitt Cancer Center and Research Institute in Tampa, Florida, and his team analyzed data from the TH-CR-302 study, which evaluated glufosfamide, an experimental alkylating drug, against best supportive care. Glufosfamide, initially formulated as a glucose-conjugated derivative of ifosfamide to target glucose-dependent tumor cells, demonstrated variable outcomes in the initial study but exhibited potential for biomarker-driven applications.

BullFrog AI utilized bfLEAP and its associated tool bfPREP to employ machine learning methodologies for the analysis of multimodal biological data, revealing heterogeneity in treatment effects. According to the abstract summarizing the study, a patient-to-patient similarity network was constructed from baseline demographics, screening labs and prerandomization clinical features. Clusters were then derived from this network by identifying recurring patient groupings. Patients without consistent groupings were not assigned a cluster.

Four distinct clusters were identified from 134 of the 281 (47%) patients: Cluster A (n=12), Cluster B (n=26), Cluster C (n=24), and Cluster D (n=72). Cluster A, characterized by lower baseline glucose values and higher baseline neutrophil and monocyte counts, was associated with improved overall survival compared with best supportive care.

The company news release noted that the platform revealed biologically significant patient clusters, highlighting subgroups where mean survival increased from the control arm to nearly three times greater in the glufosfamide-treated cohort. according to the news release.

The abstract put it this way: “Utilization of bfLEAP successfully identified patient subgroups within existing glufosfamide clinical trial data.”

Vin Singh, founder and CEO of BullFrog AI, said in the news release that "this glufosfamide case study in pancreatic cancer effectively demonstrated the capability of our platform to furnish drug developers with a comprehensive analytical tool designed to address multimodal biological complexity at scale." Singh highlighted the inefficiencies inherent in conventional drug research, stating that "excessive time and resources are squandered... resulting in patients lacking effective treatments." The AI technologies seek to alleviate this by concentrating efforts on viable pathways, lowering expenses for payers, and expediting access to focused medicines.

The abstract, titled "Data-driven subtyping and differential glufosfamide benefit in pancreatic adenocarcinoma," was scheduled to be presented at the 2026 ASCO Gastrointestinal Cancers Symposium in San Francisco tomorrow during Poster Session B. It will additionally be disseminated in a supplement to the Journal of Clinical Oncology. This study contributes to ongoing initiatives to include AI into oncology treatment. AI integration facilitates enhanced understanding of tumor heterogeneity, enabling cancer care to transcend a uniform treatment model. For managed care companies, such developments may result in more economical formularies, with biomarker testing informing payment decisions and enhancing population health measures.

This study is part of a larger movement in precision medicine, wherein data analytics connect trial results with real-world implementation. Early identification of responders may lead to decreased hospitalization rates and improved resource allocation for payers, especially in high-cost domains such as pancreatic cancer treatment.