Research on the efficiency of CT vascular imaging diagnosis of coronary stenosis based on artificial intelligence technology

CAI Chuanqing; ZHAO Chunlin; MA Yan

doi:10.20258/j.cnki.1006-9011.2025.07.009

Abstract：Objective To explore the effectiveness of CT angiography(CTA)based on artificial intelligence(AI)technology in diagnosing the degree of coronary artery stenosis.Methods 82 patients suspected of having coronary atherosclerosis were in-cluded in the study.All patients underwent coronary CTA.The efficiency and image quality of the two methods were compared by the artificial and AI coronary image post-processing.The results of coronary angiography(CAG)were used as the gold standard to compare the efficacy of the artificial and AI in the diagnosis of left anterior descending branch,left circumflex branch and right coronary artery stenosis.Results The post-processing and diagnostic time of the AI group were lower than those of the artificial group(P<0.05).The image quality scores of the left anterior descending branch,left circumflex branch,and right coronary ar-tery in the AI group were higher than those in the artificial group(P<0.05).According to the results of CAG examination,among the 82 patients,there were 26 cases of single vessel lesions,26 cases of double vessel lesions,and 30 cases of triple vessel le-sions.Among them,there were 11 cases of no stenosis in the left anterior descending branch,6 cases of mild stenosis,22 cases of mild stenosis,22 cases of moderate stenosis,19 cases of severe stenosis,and 2 cases of vascular occlusion.30 cases had no stenosis in the left circumflex artery,8 cases had mild stenosis,13 cases had mild stenosis,16 cases had moderate stenosis,12 cases had severe stenosis,and 3 cases had vascular occlusion.37 cases had no stenosis in the right coronary artery,7 cases had mild stenosis,8 cases had mild stenosis,15 cases had moderate stenosis,10 cases had severe stenosis,and 5 cases had vascu-lar occlusion.The Kappa consistency test showed that the AI group had a strong consistency in diagnosing left anterior descend-ing branch stenosis with the artificial group(Kappa=0.891,P<0.001),the AI group had a strong consistency in diagnosing left circumflex branch stenosis with the artificial group(Kappa=0.826,P<0.001),and the AI group had a strong consistency in diag-nosing right coronary artery stenosis with the artificial group(Kappa=0.849,P<0.001).There was no statistically significant dif-ference in diagnostic accuracy between the artificial group and the AI group(P>0.05).Conclusion Compared with artificial methods,AI technology has higher work efficiency and better image quality in post-processing coronary CTA images,and their effectiveness in diagnosing the degree of coronary stenosis is comparable.AI technology can be an effective tool to assist clinical physicians in diagnosing coronary stenosis.