Engineers Aim to Use AI on Chest X-rays to Predict Disease Severity

Samueli School engineers are using artificial intelligence, based on COVID-19 patients’ chest X-rays, to help determine disease severity.

April 16, 2020 - COVID-19 has imposed a significant burden on hospitals and medical centers around the world. Experts predict the demand for ventilators and beds in U.S. hospitals’ intensive care units will far exceed capacity for coronavirus patients during the pandemic.

An engineering team is investigating the use of artificial intelligence on COVID-19 patients’ chest X-rays to foresee disease severity. This information would allow medical personnel to prioritize urgent cases by being able to predict which patients will require imminent ventilation and intensive care.

Principal investigator Dr. Arash Kheradvar, professor of biomedical engineering, is working with Hamid Jafarkhani, Chancellor’s Professor of electrical engineering and computer science, and Dr. Alpesh Amin, professor of medicine at UCI Medical Center, on the research.

“We aim to establish a cloud-based AI platform to quantify the progression of the disease during the 14 days after admission to the emergency room, based on daily chest X-rays and lab results,” said Kheradvar. “We previously have been working collaboratively on a fully automated platform for cardiac segmentation using a variety of methods involving artificial intelligence. We would like to use our expertise in designing AI-based medical imaging tools to help with mitigating the COVID-19 pandemic.”

To train the simulation models, the researchers will use COVID-19 patients’ chest X-rays taken on the first day of admission and daily for up to two weeks, in addition to pertinent clinical information and patients’ final outcomes. Accordingly, they will design a deep learning network that can predict, based on the first chest X-ray taken in the emergency room, whether a patient will develop a more severe case of the disease that may require a higher level of care.

– Lori Brandt