Comparison of the performance of manual and artificial intelligence in coronary CT angiography image post-processing and diagnosis of coronary artery stenosis

LIU Jian; JIANG Zengyu; MA Yanyun; HUANG Ruiliang

Abstract：Objective:To compare the application effects of artificial intelligence(AI)and artificial intelligence in the post-processing of coronary CT angiography(CCTA)and the diagnosis of coronary artery stenosis.Methods:The CCTA data of 120 patients with suspected coronary heart disease were retrospectively analyzed,and coronary angiogra-phy was used as the gold standard for diagnosis,and the CT coronary images were post-processed by manual method and AI,and the efficacy of AI in work efficiency,image quality,and coincidence rate of coronary stenosis diagnosis was summarized and analyzed.Results:The image post-processing time of the AI processing group was significantly shor-ter than that of the manual processing group(P＜0.05),there was no significant difference in the qualification rate of volume rendering(VR)image quality the diagnostic efficiency of coronary artery stenosis between the two groups(P＞0.05),and the sensitivity,specificity and accuracy of the AI group in diagnosing coronary artery stenosis were 92.80%,89.02%and 91.94%,respectively.However,the diagnostic efficacy of mild stenosis(grade 2)of right coronary artery(RCA),left circumflex coronary artery(LCX),and left anterior descending coronary artery(LAD)in the AI group was lower than that of the manual group(P＜0.05).Conclusion:The application of AI technology is helpful to improve the efficiency of CCTA image post-processing and shorten the image post-processing time,and the diagnosis results of AI in coronary artery moderate and severe stenosis are more accurate,which can be used as an important auxiliary tool for radiologists to diagnose.