Caspian J Intern Med. 2026 Mar 10;17(2):436-448. doi: 10.22088/cjim.17.2.436. eCollection 2026.
ABSTRACT
BACKGROUND: Accurate image extraction is essential for simplifying the diagnosis and treatment of coronary artery disease (CAD) and reducing the need for multiple imaging procedures. This study compared the diagnostic accuracy of AngioBoost knowledge-based software with routine angiography in assessing coronary artery lesions.
METHODS: This cross-sectional study included 174 patients undergoing coronary angiography at Shahid Rajaei Heart Hospital, Tehran. Patients with signs and symptoms of ischemic heart disease and an indication for coronary angiography were enrolled. All underwent angiography using the Seldinger method. Angiographic films were reviewed by two independent cardiologists. AngioBoost was then applied to the same images, and its findings were interpreted by two additional cardiologists.
RESULTS: Comparison of routine angiography and AngioBoost showed contingency coefficients of 74.7% (P = 0.01) and 90.1% (P = 0.01) for the first and second cardiologists, respectively. AngioBoost correctly identified 96.3% of patent vessels, 79.9% of mild lesions, 58.0% of moderate lesions, 91.2% of significant lesions, and 96.8% of occluded vessels. Overall diagnostic agreement was 88.2% (P = 0.01). Kappa coefficients for vascular occlusion were 46.8% for Left Main, 71.7% for LAD, 70.1% for LCX, and 77.8% for RCA (all P = 0.01).
CONCLUSION: AngioBoost demonstrated acceptable diagnostic accuracy in assessing coronary artery occlusion, showing strong agreement (88.2%) with routine angiography. These findings suggest that AngioBoost may serve as a useful complementary tool for CAD diagnosis, potentially reducing physician workload and improving diagnostic efficiency.
PMID:42668867 | PMC:PMC13525821 | DOI:10.22088/cjim.17.2.436