Federal Project Grant Summary

The National Institute of Biomedical Imaging and Bioengineering (NIBIB) awarded Massachusetts General Hospital a $133,077 Project Grant effective May 1, 2026 through April 30, 2028 under the Discovery and Applied Research for Technological Innovations to Improve Human Health program (CFDA 93.286). This award supports the development of a novel deep learning methodology for multi-modal biomedical data fusion that integrates three distinct clinical data sources—clinical notes, laboratory results, and medical images—to improve disease prediction and diagnosis accuracy. The research leverages three innovative computational techniques: an attention-mechanism-based deep learning model that learns the relative importance of each data source; a recurrent neural network (RNN) approach designed to analyze longitudinal data while capturing temporal relationships in disease progression; and a graph-based deep learning technique for population-level analysis to identify disease subtypes and patient subgroups based on similarities in multi-modal data profiles.

This research initiative advances the NIBIB program's mission to transform understanding of disease prevention, detection, diagnosis, and treatment through technology development. The project addresses a critical gap in artificial intelligence applications by developing methods that integrate multiple data types simultaneously, moving beyond current two-source fusion approaches to enable more comprehensive clinical decision support. By achieving improved predictive accuracy across individual patient snapshots, longitudinal disease progression tracking, and population-level patient stratification, the project aims to establish a foundation for personalized medicine with more targeted clinical interventions.