The Perelman School of Medicine at the University of Pennsylvania https://www.pennmedicine.org, Case Western Reserve University https://case.edu, Cleveland Clinic https://my.clevelandclinic.org, along with the Cedars-Sinai Medical Center https://www.cedars.sinai.org, were recently awarded a $3.2 million grant from NIH to enhance research for improving heart transplant outcomes.
One risk when undergoing a heart transplant is that the patient’s body’s immune system may see the donor heart as a foreign object and try to reject it which can then damage the organ. Rejections occur in 30 to 40 percent of patients during the first year after transplant. However, the current rejection grading standard indicates poor diagnostic accuracy and has limited ability to discern the mechanism of rejections. These limitations expose patients to risks of both over and under treatment
The universities involved in the research will provide the data to include digitized images of biopsies from patients who have already had transplants. Kenneth B. Margulies, MD, Professor of Cardiovascular Medicine at Penn in partnership with Anant Madabhushi, PhD, Professor of Biomedical Engineering at Case Western Reserve and Director, for Center for Computational Imaging and Personalized Diagnostics, will apply AI techniques to the data to see whether the initial biopsy images could have more accurately predicted which patients would accept or reject the new heart.
Improved diagnostic accuracy may enable recognition of serious rejection possibilities earlier and may also promote reduced rates of infection and other complications from immune-suppressing drugs taken by transplant patients. The team also hopes to identify patterns to help predict how patients will do over the long term which would enable fewer biopsies of the heart to take place.
In addition, the team will compare the relative performance of the AI analysis against human pathologists to compare their accuracy in identifying serious rejections. Previous research has shown that computers were more accurate than their human counterparts in diagnostic ability.
However, the team believes pathologists will not be replaced by computers. Instead, Dr. Margulies asserts. “Computer-aided tissue diagnostics will serve as a decision support tool for pathologists, consistently and efficiently and be able to identify subtle features that will increase the value of the diagnostic procedure and improve patient outcomes.”