Conference Abstracts - Summit on Cancer Health Disparities (SCHD26)
Vol. 6, Issue Supplement 1, 2026 · S1-1
OncoSphere AI: A Scalable AI-Driven Tumor Board Platform to Bridge the Expertise Gap in Low-Resource Oncology Settings
Aakash Desai, MD, MPH, FASCO,Sanad Alhushki, MD,Ellen McNeeley, MA,Sandeep Bodduluri, PhD
Submission received: 2025-11-25 / Accepted: 2026-01-07 / Published: 2026-01-25
Abstract
Background
Disparities in cancer survival between academic centers and rural or Low- and Middle-Income Country (LMIC) settings are largely driven by the "Expertise Gap." While precision oncology requires sub-specialist interpretation of complex genomic and clinical data, frontline providers in resource-constrained settings function as generalists with limited access to Multidisciplinary Tumor Boards (MTB). We propose OncoSphere AI, a digital clinical decision support system designed to democratize academic-level expertise and ensure guideline-concordant care at the point of service.
Method
OncoSphere AI utilizes Natural Language Processing (NLP) and interoperable standards to ingest fragmented patient data (pathology, imaging, history) and auto-generate a comprehensive "Patient Journey" dashboard. For this implementation study, the platform is adapted to integrate multi-agentic specialty personas through NLP and Retrieval Augmented Generation (RAG) frameworks to enable guidelines based care. The system functions as an asynchronous "Virtual Fellow," synthesizing data and flagging guideline-aligned therapeutic options and clinical trial eligibility.
Results
In preliminary internal validation using retrospective datasets (N=25), OncoSphere AI demonstrated 95% concordance with NCCN Guidelines for standard-of-care systemic therapy recommendations and reduced administrative preparation time for case review by 85% (from 20 mins to 3 mins per case). By automating the clerical burden of data synthesis, the platform allows local providers to focus on patient counseling and care delivery. The pilot aims to demonstrate a 15% increase in guideline adherence and clinical trial screening efficiency within the first 12 months of deployment.
Conclusion
OncoSphere AI provides a scalable, low-latency mechanism to export sub-specialist expertise to the periphery of the health system. By reducing the cognitive burden on generalist oncologists, this tool holds the potential to standardize high-quality care delivery and reduce survival disparities across diverse geographies.