UCSF Pediatric Focus

Using Artificial Intelligence to Speed Imaging of Brain Emergencies

Computed tomography (CT) scanning is currently the only type of imaging used worldwide to diagnose neurologic emergencies. Immediate diagnosis, aided by rapid, automated evaluation of head CT could greatly improve care in situations where minutes count.

When a patient has a neurological emergency, such as a traumatic brain injury, physicians are in a race against time to minimize the destruction or deterioration of fragile brain tissue. To speed the diagnostic testing needed to triage patients and guide treatment, Pratik Mukherjee, MD, PhD, and Esther Yuh, MD, PhD, of the UCSF Department of Radiology and Biomedical Imaging are testing whether they can use artificial intelligence technology to automatically recognize life-threatening findings on emergency CT scans of the head.

“The idea is to accelerate the detection of emergency features on CT scans of the head, so that critical decisions about patient care can be made more rapidly,” Mukherjee says. “These are life-and-death decisions where minutes count. Anything that accelerates [the detection of emergency features] is crucial and can save lives and reduce long-term disability from these disorders.”

Another potentially important advance stems from the technology’s ability to catalog clinically significant “digital markers” that are recognizable across scans. This will facilitate future precision medicine research by combining data from quantitative image analyses with other types of data.

This project is one of six that the California Initiative to Advance Precision Medicine recently selected to receive up to $1.2 million in grant funding. Initially, the technology will be used in research, allowing imaging to be more easily incorporated into large-scale studies that determine how to provide the best treatment and care to those who experience a neurological emergency.