Project Grant R01HL187015

The National Heart, Lung, and Blood Institute awarded the University of Pittsburgh $3,229,552 on August 15, 2026, under the Cardiovascular Diseases Research program (CFDA 93.837) to develop DART-HIV, an artificial intelligence-driven framework for modeling physiological vulnerability and predicting adverse health outcomes in people living with HIV over a 10-year horizon.

DART-HIV comprises two core components: a time-dependent risk trajectory forecasting four major adverse outcomes—all-cause death, hospitalizations, emergency department visits, and major health events such as cancer or cardiovascular disease—and a directed causal graph identifying population- and individual-specific drivers of vulnerability to support targeted intervention. The model integrates longitudinal, multi-institutional cohort data spanning ten years of clinical follow-up, including chest CT scans, clinical data, and molecular markers such as HIV viral load and CD4 cell count. A distinguishing element is the use of quantitative image biomarkers extracted from chest CT images, including three-dimensional assessments of body composition (fat, muscle, bone, and spatial distribution) and cardiopulmonary structures (heart, lungs, airways, vasculature, parenchyma).

The award is classified as a Project Grant under the assistance type. Work is performed at the University of Pittsburgh in Pittsburgh, Pennsylvania, with a period of performance from award date through June 30, 2030.