- Artificial Intelligence May Improve Heart Transplant Outcomes
- Racial Disparities Seen in Heart Transplant Patients with COVID-19
“Our retrospective pilot study demonstrated that combining artificial intelligence and human intelligence can improve expert agreement and reduce the time needed to evaluate biopsies,” said senior author Faisal Mahmood, PhD, from the Mahmood Lab at Brigham's Department of Pathology, in the press release. “Our results set the stage for large-scale clinical trials to establish the utility of AI models for improving heart transplant outcomes.”
CRANE was trained using pathology images from over 1,300 procedures at Brigham and Women's, learning to perform tasks such as detecting, subtyping, and grading transplant rejections.
Researchers then validated the system, using test biopsies from Brigham and test sets from hospitals in Switzerland and Turkey.
Following this process, researchers concluded that CRANE effectively detected and monitored the rejection process, and it reduced the potential for disagreement between providers.
Though the researchers recognized room for improvement, they concluded that the results produced by CRANE show that using AI in heart transplant procedures has significant potential.
“Throughout the history of medicine, diagnostic assessments have been largely subjective,” said Mahmood. “But because of the power and assistance of computational tools, that’s beginning to change. The time is right to make a shift by bringing together people with clinical expertise and those with expertise in computational science to develop assistive diagnostic tools.”
Interest in using AI to improve heart procedures is growing.
In 2020, the National Institute of Health granted $3.2 million to studies examining the use of AI in heart transplants. With this grant, researchers from Perelman School of Medicine at the University of Pennsylvania, Case Western Reserve University, Cleveland Clinic, and Cedars-Sinai Medical Center aimed to study what causes the rejection of a new heart and how to manage it. They used AI to assist in analyzing cardiac biopsy tissue images, finding the types of immune cells that reject the new heart.
A January 2019 study also described research regarding transplant rejections. However, it used genomic sequencing rather than AI. Researchers used a blood test to determine if the immune system attacked a new lung. The study also notes that this presents an opportunity to integrate personalized medicine into monitoring transplant reactions because providers can customize treatment based on who is at high risk.