急性主动脉综合征漏诊率被AI砍下九成

Core Viewpoint - The collaboration between Zhejiang University First Affiliated Hospital and Alibaba Damo Academy has led to the development of an AI model, iAorta, which can quickly identify potential acute aortic syndrome risks in routine CT scans, significantly improving diagnostic accuracy and speed [1][3]. Group 1: AI Model Development - The iAorta model was developed to address the challenge of diagnosing acute aortic syndrome, which often presents with non-specific symptoms leading to high misdiagnosis rates [2][3]. - The model integrates with hospital PACS systems, allowing simultaneous analysis of CT scans as doctors review them, providing alerts for potential risks [3][4]. Group 2: Performance Metrics - iAorta demonstrated a sensitivity of 97% and specificity of 94% in identifying acute aortic syndrome during validation [4]. - In a retrospective study involving 130,000 patients, iAorta achieved a sensitivity of 92.6% and specificity of 99.2%, significantly reducing the misdiagnosis rate from 48.8% to 4.8% [4][5]. Group 3: Clinical Impact - The average diagnosis time for acute aortic syndrome was reduced from 11.4 hours to 1.1 hours with the use of iAorta [5]. - In a prospective study with 13,000 patients, iAorta identified 9 cases of acute aortic syndrome, compared to only 2 identified by doctors alone [6]. Group 4: Deployment and Future Plans - iAorta has been deployed in the first 10 hospitals in Zhejiang, with plans for nationwide expansion [6]. - The model is expected to streamline the diagnosis, transfer, and treatment processes for patients with acute aortic syndrome [6].