Project Grant R01HL188153

The National Heart, Lung, and Blood Institute awarded The Children's Hospital Corporation, doing business as Boston Children's Hospital, $857,878 for a Project Grant (R01HL188153) on September 1, 2026, under the Cardiovascular Diseases Research program (CFDA 93.837).

The award funds development and evaluation of pediatric cardiac foundation models—machine learning models pre-trained on large volumes of unlabeled electrocardiogram (ECG), echocardiogram (ECHO), chest X-ray, and clinical text data to enable accurate diagnosis and prognostication even for rare pediatric cardiovascular diseases. The research team will train modality-specific foundation models using millions of unlabeled records from Boston Children's Hospital, Texas Children's Hospital, and public repositories, fine-tuning on site-specific data to predict ventricular dysfunction, perform diagnostic phenotyping, and estimate survival. The team will develop multimodal integration strategies to merge modality-specific models through bridge networks into a shared representation space and will evaluate model interpretability and clinical deployment readiness.

Performance occurs in Boston, Massachusetts, from September 1, 2026, through May 31, 2030. Boston Children's Hospital, a nonprofit pediatric teaching hospital and Harvard Medical School affiliate, brings access to one of the world's largest pediatric cardiac datasets and prior experience developing and deploying AI-ECG and AI-ECHO models.