Study On The Correlation Between Radiomics Based On Pericoronary Adipose Tissue And The Severity Of Coronary Artery Disease

Objective To compare the histogram parameters of pericoronary adipose tissue between patients with coronary artery disease and those without coronary artery disease,and between patients with acute coronary syndrome and patients with stable coronary artery disease,and to explore the clinical significance.Materials and Methods Clinical data and imaging data of 93 patients admitted to the department of cardiology from January 2013 to December 2018 who underwent coronary CT angiography were retrospectively analyzed.And 58 patients who underwent coronary CT angiography in the department of cardiology from January to December 2018 and had no coronary artery disease in all three coronary arteries were retrospectively analyzed.39 coronary arteries in 39 patients with coronary artery disease for acute coronary syndrome patients(acute coronary syndromes group)and 54 cases with 54 coronary arteries in patients with stable coronary artery disease(stable coronary artery disease group),combined them,93 cases of 93 coronary arteries as coronary artery disease group,58 patients with 174 coronary arteries without coronary artery disease(coronary artery disease-free group).A region of interest(ROI)was delineated around the stenosis of coronary arteries in patients with coronary artery disease.In the coronary artery disease-free group,starting from 30mm away from the coronary root,a segment of pericoronary adipose tissue with a length of 20mm was delineated,and the range of CT value was-190～-30HU to eliminate non-adipose tissue.CT value of ROI after excluding non-adipose was measured and histogram analysis was performed.The parameters obtained included mean,median,5th,10th,45th,55th,70th,and 95th percentiles.The histogram parameters between the 4 groups were compared,and receiver operating curve(ROC)analysis was made to evaluate of the value of the parameters.Stepwise regression multiple logistic regression analysis was used for parameters screening and the final prediction model was constructed.Draw the ROC curve of the final model and analyze its value.Results There were significant differences in the mean,median,5th,10th,45th,55th,70th and 95th percentiles between the coronary artery disease group and the coronary artery disease-free group,and between the acute coronary syndrome group and stable coronary artery disease group(all P<0.05).The sensitivity,specificity and area under the curve were 82.1%,89.1%and 0.90 respectively for the diagnosis model between acute coronary syndrome and stable coronary artery diseaseConclusion The parameters of pericoronary adipose tissue histogram have high value in differentiating patients with coronary artery disease from those without coronary artery disease,as well as patients with acute coronary syndrome and stable coronary artery disease.The multi-parameter model established by multiple logistic regression has better diagnostic value than the single parameter model.Objective To evaluate the value of radiomics model based on pericoronary adipose tissue in predicting coronary flow with CT fractional flow reserve(CT-FFR)as reference.Materials and Methods Retrospective analysis was performed on inpatients in the department of cardiology who underwent coronary CT angiography(CCTA)from January 2018 to December 2020.60 patients with 68 coronary arteries with CT-FFR less than or equal to 0.8 and stenosis rate greater than 50%were selected as the ischemic specific disease group.171 cases with 189 coronary arteries with CT-FFR greater than 0.8 and stenosis rate greater than 50%were selected as the control group.Imaging and clinical data of the two groups were collected.The pericoronary adipose tissue of the plaque was segmented on the CCTA images of each enrolled coronary artery.The radiomics features of pericoronary adipose tissue were extracted using perivascular fat analysis tool software.All 257 coronary arteries were divided into the training group and the testing group by stratified random sampling in a 7:3 ratio(179 arteries in the training group and 78 arteries in the testing group).Then features were selected by single factor logistic regression,the least absolute shrinkage and selection operator(LASSO)and multiple logistic regression.The support vector machine(SVM)method was used to build the radiomics model.Results The diagnostic efficacy of the radiomics model in the training group was evaluated using the receiver operation curve(ROC)and tested independently in the testing group.In the testing group,the accuracy,sensitivity,specificity,and area under the curve(AUC)for identifying whether coronary arteries are ischemic were 82.1%,52.4%,93%,and 0.79,respectively.Conclusion The radiomics model based on pericoronary adipose tissue has good accuracy in predicting coronary flow,which may be a new method for evaluating coronary flow.