A retrospective study analyzed emergency CT scans of the abdomen and pelvis from Yale New Haven Hospital (July 31 to Dec. 31, 2020 and Jan. 1 to May 10, 2021). An FDA-cleared AI algorithm (Aidoc) processed the images, while a natural language processing (NLP) program evaluated the corresponding radiology reports. Cases where the AI detected an abdominal aortic aneurysm (AAA), but NLP did not, were flagged as potential discrepancies and independently reviewed by an emergency radiologist.
A total of 4,023 abdominal and pelvic CT exams were analyzed, with 98.3% (3,955) reported as negative for abdominal aortic aneurysm (AAA) by natural language processing (NLP). AI flagged 16 potential discrepancies, and secondary review confirmed previously undocumented AAA in 31% (5/16) of these cases, increasing detection by 7.4% (5/68) compared to radiologists alone. In 69% (11/16) of false-positive cases, AI overestimated aortic diameter due to lumen obliquity, which was corrected using multiplanar reconstruction.
AI algorithms can potentially enhance incidental AAA detection on CT imaging but require further advancements before being viable for fully automated clinical screening.
Disclaimer: This links to a third-party website, which may require credentials to review.
Aidoc experts, customers and industry leaders share the latest in AI benefits and adoption.
Explore how clinical AI can transform your health system with insights rooted in real-world experiences.
Learn how to go beyond the algorithm to develop a scalable AI strategy and implementation plan.
Explore how Aidoc can help increase hospital efficiency, improve outcomes and demonstrate ROI.
