Project Grant R01HL188930

The National Heart, Lung, and Blood Institute awarded New York University a Project Grant of $847,373 on August 1, 2026, under the Cardiovascular Diseases Research program (CFDA 93.837).

The project will use national organ transplant registry data, including structured clinical data and unstructured narratives describing donor chest imaging findings, to improve evidence for lung offer decisions in transplantation. Large language models will be trained to extract structured data elements from unstructured text, capturing the presence and severity of lung abnormalities such as edema, pneumonia, and atelectasis. The research will quantify risk-adjusted associations between these newly structured data elements and post-transplant outcomes, including graft survival time, using statistical and causal inference methods. Machine learning methods will identify patterns and interaction effects in the data that suggest specific subgroups of patients who might benefit from particular donor lung profiles, addressing the substantial variability in clinical donor characteristics that currently complicates acceptance decisions and contributes to hundreds of patient deaths annually on the national lung waiting list.

Work is performed at New York University School of Medicine in New York, New York. The period of performance extends from August 1, 2026, through April 30, 2031. The award is classified as a Project Grant, an assistance type reflecting the research funding structure.