AI Model Pairs EKGs With Blood Tests to Spot Heart Transplant Rejection

An alternative to the biopsy

Doctors currently diagnose heart transplant rejection with a biopsy, in which a small piece of heart muscle is surgically removed and inspected under a microscope for inflammation and other changes that may show the immune system is rejecting the organ.

The NYU Langone team explored a different route. They asked whether electrocardiograms (EKGs), which record the heart's electrical activity through sensors placed on the skin, could be combined with blood tests to identify rejection without an invasive biopsy in many cases.

How the models were built and tested

The researchers trained AI models to recognize patterns in 5,300 EKG readings taken from 2,357 adult heart transplant recipients. One model was trained on EKG data alone. Another used EKG readings plus the results of two blood tests commonly used to predict rejection risk. Biopsy records paired with EKG readings from the same patients served as the measure of accuracy.

The combined model was trained on data from recipients treated between 2018 and 2024. Each EKG was matched to a biopsy performed within the past month, a window the researchers say is supported by earlier research. Rejections were sorted into two groups: no or mild rejection, and moderate or severe rejection. Treatment changes, such as adjusting immune-suppressing medications, are typically made only for the more serious cases.

In a test group of an additional 38 male and female heart transplant recipients, the combined model performed better, correctly identifying 94 percent of patients who were not experiencing rejection. The model based on blood tests, by contrast, incorrectly flagged 19 patients as potentially needing a biopsy. The researchers said the combined model would have spared those patients the procedure.

Why blood biomarkers alone fall short

The blood biomarkers measure gene activity linked to cellular rejection and fragments of donor DNA in the recipient's bloodstream. Previous work has shown they are effective at spotting rejection, but they frequently produce false-positive results that lead to unnecessary biopsies, the researchers said.

Published online Sept. 25 in the Journal of Heart and Lung Transplantation, the study is the first to combine these blood biomarkers with EKG readings in a single AI-enabled model and compare the analysis directly against biopsy results, according to the researchers.

Senior author Lior Jankelson, MD, PhD, of NYU Grossman School of Medicine and NYU Tandon School of Engineering, said: "Our results highlight that electrocardiograms contain an abundance of physiological information that can be used to substantially improve the accuracy of detection and enable earlier diagnosis and treatment for patients with cardiac transplant rejection."

What comes next

The findings come from a relatively small test group of 38 recipients, and the researchers plan to test their model in more patients at several transplant centers as a next step. The study was funded by NYU Langone.

Dr. Jankelson said: "Heart transplantation is a major, resource-intensive procedure, and rejection is common but treatable if caught in time." He added that combining the availability of EKGs with the support of AI "may help identify rejection earlier and save more lives." If confirmed in larger groups, the approach could mean fewer biopsies for transplant patients who are not experiencing rejection.