Adopting a functional precision medicine approach to reduce cancer disparities in Hispanic and Black children of Miami

Cancer remains the leading cause of disease-related mortality in children in the U.S. Relapsed, refractory, and rare pediatric cancers remain major clinical challenges with dismal outcomes and limited treatment options. Most available therapies are repurposed adult regimens lacking pediatric-specific efficacy and safety data. To address this urgent unmet need, we propose to expand a clinically validated Functional Precision Medicine (FPM) program integrated with explainable artificial intelligence (xAI) to optimize individualized, low-toxicity therapies for pediatric patients with relapsed, refractory, or rare cancers who have exhausted standard-of-care options.

FPM combines comprehensive genomic and transcriptomic profiling with high-throughput ex vivo drug sensitivity testing (DST) on live patient-derived tumor cultures to generate actionable treatment recommendations. Our DST scoring system yields quantitative, actionable data to guide less toxic and effective individualized treatment regimens. Here we propose Specific Aim 1: To evaluate the efficacy of FPM as a tool for finding less toxic and effective, personalized treatment regimens to address poor outcomes in pediatric cancer care. We hypothesize that using novel tools such as FPM integrated with AI will improve cancer outcomes (measured as overall best response and progression-free survival) while concurrently reducing treatment-related toxicities and expanding access to advanced cancer therapies.

Aim 1A: Broaden access to personalized treatment options and clinical management recommendations using FPM profiling through Nicklaus Children’s Hospital, located in Miami, as the hub institution. Aim 1B: Compare individual outcomes in children with advanced cancers treated with FPM-guided therapy as compared to non-FPM-guided (conventional) therapy. Specific Aim 2: To elucidate novel correlations between tumor molecular alterations and ex vivo drug responses in advanced pediatric cancer patients. We hypothesize that distinct genetic, transcriptomic, and phenotypic properties exist within these groups, influencing drug response patterns.

We will perform multi-omics molecular profiling and then use an advanced artificial intelligence platform8,9 to discover novel associations between molecular features and DST response data. Our preliminary studies have demonstrated the feasibility of our approach, with >80% DST success rate and >65% actionable findings, with strong correlation to clinical benefit. This integrative FPM-xAI platform represents a scalable, logistically feasible strategy to personalize therapy, reduce treatment-related toxicities, and improve outcomes in pediatric oncology.

Grant Number: 5U54MD012393-09
NIH Institute/Center: NIH
Principal Investigator: Diana Azzam