This retrospective multicenter study evaluated the contribution of artificial intelligence (AI) to radiologists’ diagnostic interpretation of arterial stenosis (AS) in CT angiography of the head and neck. The examinations of 268 patients were divided into two groups: without AI-washout-with AI and with AI-washout-with AI. Six independent readers reviewed the images with and without AI assistance. AI improved the sensitivity of all readers in detecting AS ≥30% by 5.2% (P <0.001) and increased patient-level diagnostic accuracy by 4.1% (P <0.001), while also reducing the number of false-positive findings. Additionally, AI assistance decreased median reading and reporting time for all readers by 241 s (P<0.001). These findings highlight the potential of AI solutions in improving both the diagnostic performance and efficiency of radiologists in this setting.
Home › News › Artificial intelligence enhances radiologists’ diagnostic performance in detecting arterial stenosis on contrast-enhanced CT angiography of the head and neck