AI tool predicts HF mortality with 88% accuracy
A collaboration between physicists and cardiologists at the University of California San Diego has resulted in an AI tool that can predict heart failure patients’ life expectancy with 88% accuracy.
UCSD physics professor Avi Yagil, PhD, teamed up with cardiologists Eric Adler, MD, and Barry Greenberg, MD, after the pair first treated him for heart failure some seven years ago. Yagil was diagnosed with HF after returning home from a trip to Europe in 2012, and within four years he’d undergone a heart transplant.
“I consider June 17 my second birthday,” Yagil said in a release.
The professor said the idea for an AI-driven prediction tool for HF began to form while he was recovering from his transplant.
“In my day job, I use machine learning to understand a vast amount of information and measurements of particles and how they interact,” he said. “The human body is even more complex, but the medical profession isn’t utilizing the technologies that are needed to capture the multi-dimensional correlations between the measurements, such as lab tests and vital signs, and the outcomes. We hypothesized that such methodology and techniques could contribute to improving the prognosis and treatment of patients with HF.”
Yagil, Adler, Greenberg and colleagues at UCSD developed a machine learning algorithm based on the de-identified electronic heart records of 5,822 hospitalized or ambulatory patients with heart failure at UC San Diego Health. The model considered patients’ diastolic blood pressure, creatinine levels, blood urea nitrogen, hemoglobin concentration, white blood cell and platelet counts, albumin, and red blood cell distribution before coming up with a mortality risk score that indicated either a low or high risk of death.
According to the team, the model was able to accurately predict life expectancy 88% of the time, performing “substantially better” than other published models.
“This tool gives us insight, for example, on the probability that a given patient will die from heart failure in the next three months or a year,” Adler said in the release. ‘This is incredibly valuable. It allows us to make informed decisions based on a proven methodology and not have to look into a crystal ball.”
The researchers said they also tested their model—successfully—using de-identified patient data from the University of California, San Francisco, and a database of 11 European medical centers. They all said the collaboration between cardiologists and physicists was what made the innovation possible.
“It’s been a wonderful collaboration with two groups that don’t usually join forces,” Adler said. “Our findings need further validation, but we are thrilled to have these results to build upon. Avi has a first-hand perspective as a patient and a strong motivation to help improve existing medical strategies and approaches. Working with him has been a highlight of my career.”
The team’s findings were published in the European Journal of Heart Failure on Nov. 12.
After graduating from Indiana University-Bloomington with a bachelor’s in journalism, Anicka joined TriMed’s Chicago team in 2017 covering cardiology. Close to her heart is long-form journalism, Pilot G-2 pens, dark chocolate and her dog Harper Lee.