Tuesday, June 22, 2021

Allegheny Health Network Specialists Explore Use of Artificial Intelligence to Improve Stroke Diagnosis

PITTSBURGH - Doctors at Allegheny Health Network (AHN) are studying the effectiveness of an innovative artificial intelligence (AI) capability called Rapid CTA to more accurately and quickly diagnose stroke. The AHN team, led by neuroradiologist Warren Chang, MD and radiology resident Julie Adhya, MD, has published its initial experience with the technology in The Neuroradiology Journal (“Positive predictive value and stroke workflow outcomes using automated vessel density (Rapid-CTA) in stroke patients: One year experience”). Other co-authors of the study included AHN neuroradiologists Charles Li, MD and Michael Goldberg, MD, AHN neurologist Russell Cerejo, MD and neuroradiologist Laura Eisenmenger, MD.

Through an industry partnership with RapidAI, stroke specialists at AHN have access to real-time stroke imaging analyses to help them determine the most appropriate course of treatment for a patient experiencing stroke symptoms. Specifically, Rapid CTA compares a patient’s blood vessels against each other, then highlights brain regions with reduced blood vessel density, helping clinicians to quickly identify suspected large vessel occlusions. The most common type of stroke is an ischemic stroke, which occurs when there is a blockage in a vessel supplying blood to the brain. When one of the major arteries of the brain is blocked, it is considered a large vessel occlusion (LVO) stroke.

As a person is having a stroke, the clock is ticking, explained Dr. Chang. “There is a finite window of opportunity for treatment to begin, while the brain is suffering irreparable damage,” he said. The standard treatment for LVO stroke is a minimally invasive procedure called mechanical thrombectomy performed by an interventional radiologist using specialized equipment threaded into the patient’s arteries to remove the clot. Rapid CTA can rapidly notify radiologists, neurologists and endovascular therapists of suspected LVOs amenable to mechanical thrombectomy, minimizing stroke-related permanent disabilities for the patient in the long run.

 The journal article documents the group’s findings obtained through a one-year study of 310 patients who arrived at an AHN emergency department with stroke symptoms or neurologic deficits. The analysis compared two cohorts of patients with similar vessel blockages; physicians evaluating the patients in one cohort had access to Rapid CTA imaging, while those evaluating the other group did not. Ultimately, patients who had Rapid CTA received more timely treatment