Thursday, December 3 | 11:00 a.m.-12:00 p.m. | R4-SSHN03-5 | Room N228
An AI algorithm can help spot intracranial aneurysms on routine contrast-enhanced neck CT exams, according to this scientific paper.
Routine neck CT scans include part of the brain’s blood vessels, which can be a blind spot when the exams are ordered for unrelated reasons, noted presenter Owais Ahmad Bhat, MD, of the University of Louisville, and colleagues. The group evaluated how an AI aneurysm detection algorithm (Aidoc) performed as a secondary reviewer on these studies, with an emphasis on its performance for identifying clinically occult intracranial aneurysms.
The researchers prospectively applied the algorithm to 631 contrast-enhanced neck CT exams performed for routine clinical indications, excluding dedicated CT angiography studies. They then compared AI-flagged cases with the original radiology reports.
The algorithm flagged three exams as positive for aneurysm. In two, it found subtle aneurysms under 2 mm that the interpreting radiologist hadn’t initially identified, prompting addenda to the reports. The third, a 4.5- to 5-mm aneurysm, had already been spotted by the radiologist and was later confirmed on CT angiography.
“AI reduces perceptual oversight in routine CT neck, enhancing detection of incidental aneurysms and overall interpretive quality,” the researchers wrote.
Check out this Thursday morning session to learn more.
