Accuracy and clinical value of artificial intelligence-assisted diagnosis of coronary CT angiography images in patients with acute coronary syndrome
10.3969/j.issn.1002-1671.2024.07.010
- VernacularTitle:冠状动脉CT血管成像图像人工智能辅助诊断用于急性冠状动脉综合征患者评估的准确性和临床价值
- Author:
Genyi FENG
1
;
Gang WANG
;
Jinsong LI
;
Jiangang WANG
;
Honghong GUO
;
Xueyan LI
;
Qing HU
;
Zhiming ZHAO
;
Chao HE
Author Information
1. 西安宝石花长庆医院影像科,陕西 西安 710200
- Keywords:
artificial intelligence;
coronary computed tomography angiography;
acute coronary syndrome;
coronary artery stenosis
- From:
Journal of Practical Radiology
2024;40(7):1079-1082
- CountryChina
- Language:Chinese
-
Abstract:
Objective To explore the accuracy and clinical application value of artificial intelligence(AI)-based coronary computed tomography angiography(CCTA)in the evaluation of coronary artery stenosis in patients with acute coronary syndrome(ACS).Methods Fifty-four patients with suspected ACS who underwent CCTA examination and invasive coronary angiography(ICA)within 72 h were retrospectively selected.The CCTA images of all patients were processed by AI(AI group)and manual post-pro-cessing(manual group),respectively.The image quality,work efficiency and detection rate of coronary artery stenosis were compared between AI group and manual group.With ICA results as the gold standard,the sensitivity,specificity,positive predictive value,negative predictive value and accuracy of AI in the diagnosis of ACS patients with coronary artery stenosis(≥50％)in CCTA were analyzed,and the consistency of AI and ICA examination results was tested.Results The image quality of CCTA in AI group(grade Ⅰ 27.8％)was better than that in manual group(grade Ⅰ 14.8％),but there was no statistical difference between the two groups(X2=2.707,P＞0.05).The average diagnosis time of AI group(89.67 s±33.21 s)was significantlyshorter than that of manual group(813.33 s±301.84 s)and the difference was statistically significant(t=-17.512,P＜0.001),and the average time gain rate was 88.97％.There was no statistical difference in the detection rate of coronary artery stenosis(≥50％)between AI group and manual group(x2=0.003,P＞0.05).The sensitivity,specificity,positive predictive value,negative predictive value,and accuracy of AI in diagnosis of ACS were 87.60％,96.44％,80.30％,97.92％,and 95.19％,respectively,which were significantly consistent with the results of ICA examina-tion(Kappa=0.810,P＜0.05).Conclusion AI-assisted diagnosis can correctly identify the coronary artery tree with better image,significantly shorten the diagnosis time of CCTA in ACS patients with high accuracy,and can provide a strong basis for the early treat-ment of patients with acute chest pain.
