The Application Research Of Coronary Artery CT Based On Artificial Intelligence In Diagnosing Coronary Occlusion

Objective To evaluate the accuracy of coronary computed tomography angiography(CCTA)based on artificial intelligence(AI)in diagnosing coronary artery plaque and coronary artery stenosis.Exploring the application value of artificial intelligence-based CCTA in the diagnosis of coronary artery occlusive disease,aiming to provide a basis for clinical application of AI to assist in the diagnosis of coronary artery disease.Methods A retrospective collection of 123 patients with suspected coronary heart disease who underwent CCTA and DSA examinations in Baotou Central Hospital from June 2018 to October 2019,and imported the patient’s CCTA data into the artificial intelligence-assisted diagnosis system.Using the CCTA as the gold standard to evaluate the performance of AI in diagnosing coronary artery plaque,calculate its specificity,sensitivity,accuracy,etc.;Using the CAG as the gold standard to evaluate the accuracy of AI and CT in diagnosing coronary artery stenosis,and analyze the accuracy of its diagnosis of stenotic segments caused by different plaque components.Diagnostic accuracy was determined by calculation of sensitivity,specificity,positive predictive values(PPV),negative predictive values(NPV),and by drawing receiver operating characteristic(ROC)curves.Results A total of 123 patients,496 vessels and 1107 vessel segments were enrolled in this study.All patients underwent CCTA and CAG examinations.(1)Among the 1107 vascular segments,126 calcified plaques,130 mixed plaques,87 non-calcified plaques,and 764 no plaques.The accuracy of AI diagnosis of calcified plaque,mixed plaque,and non-calcified plaque was 80.488%,91.14% and 93.496%,respectively.(2)The AUC of AI in the diagnosis of coronary stenosis at the patient level,vessel level,and segment level were 0.764,0.821,and0.853,respectively.(3)The AUC of AI in the diagnosis of significant coronary stenosis at the patient level,vessel level,and segment level were 0.76,0.879,and 0.899,respectively.(4)The accuracy of AI in diagnosing significant coronary artery stenosis in the corresponding segments of calcified plaque,mixed plaque and non-calcified plaque were 84.127%,88.462%,and85.057%,respectively.Conclusion AI shows good performance in diagnosing coronary artery occlusion.When diagnosing the characteristics of coronary artery plaque and coronary artery stenosis,the results of AI are trustworthy.Using AI to assist radiologists in diagnosis can further improve diagnosis efficiency.