Application value of coronary computed tomography angiography based on artificial intelligence post-processing in diagnosis of coronary artery lesions

GENG Hequn; SHI Jinzheng; MENG Choushuan; ZHANG Xiaochen; ZHANG Xiaoping

doi:10.3969/j.issn.1009-0126.2026.02.007

Abstract：Objective To explore the application value of coronary computed tomography angiography(CCTA)based on artificial intelligence(AI)post-processing in the diagnosis of coronary artery lesions.Methods 200 elderly patients with suspected coronary heart disease admitted in our hospital from February 2023 to February 2025 were selected.All patients received CCTA at admission,and coronary angiography(CAG)within 3 months after CCTA.The results of CAG were regarded as the gold standard.All CCTA images were imported into the workstation for manual post-processing and AI software post-processing to analyze the main blood vessels,such as the left main coronary artery(LMCA),left anterior descending artery(LAD),left circumflex artery(LCX),and right coronary artery(RCA)and their lesion conditions.Based on different post-processing methods,they were divided into a manual group and an AI group.The time for post-processing and subjective and objective scores of image quality were compared between the two groups.Kappa test was used to compare the consistency,sensitivity,specificity,and accuracy of the diagnosis of the above coronary lesions in the two groups.Receiver operating characteristic(ROC)curves were drawn to analyze the diagnostic efficacy of the two post-processing methods for coronary artery lesions.Results There were no statistical differences in subjective and objective scores of image quality between the post-processing methods(P>0.05).The AI group exhibited a significant shorter time for image post-processing than the manual group(2.67±0.26 min vs 13.31±0.60 min,P<0.01).There were no obvious differences in the diagnostic results of coronary artery lesions by CAG,and AI and manual post-processing(P>0.05).The Kappa value for the consistency in diagnosing the lesions in LMCA,LAD,LCX,and RCA was 0.827,0.900,0.657,and 0.661,respectively,between the AI group and CAG examination,and the Kappa value for the consistency in diagnosing above lesions was 0.641,0.669,0.904 and 0.863,respectively,between the manual group and CAG examination.The AUC value of AI post-processing in diagnosing coronary artery lesions was 0.848(95%CI:0.807-0.889),and the AUC value of manual post-processing was 0.842(95%CI:0.797-0.890).There was no statistical difference in the AUC values between the two post-processing methods(P>0.05).Conclusion CCTA based on AI post-processing has the advantage of shorter processing time in the diagnosis of coronary artery lesions.It also shows high diagnostic efficacy for coronary artery lesions,and can be used as an auxiliary tool for clinical diagnosis of coronary artery lesions.