Objective To evaluate the predictive value of pericoronary adipose tissue(PCAT)imaging features based on coronary CT angiography(CCTA)to identify major adverse cardiovascular events(MACE)in patients with suspected coronary heart disease in the next 5 years.Methods Retrospective analysis was performed on patients suspected of coro-nary heart disease who underwent CCTA examination.Patients with MACE events were selected as case group(205 cases),and patients without MACE events in the database during the same period were selected as the control group(205 cases).PC AT imaging features selected by LASSO and Fat Attenuation Index(FAI)were respectively modeled,and a combined model of the two was constructed,and the predictive efficiency of the three models was compared by ROC curve,decision curve and calibration curve.Results The PCAT radiomics model(AUC=0.94,0.89)was superior to the FAI model(AUC=0.59,0.53)in evaluating the predictive value of MACE events in the next 5 years,and the AUC values of the two models were significantly different(P＞0.05).The predictive power of the combined model(AUC=0.96,0.94)in asses-sing this event has improved,and the calibration curves of the above three prediction models all had good fit(P＜0.05).Conclusion The PCAT radiomics model based on CCTA can provide more predictive information than the FAI model in i-dentifying suspected CAD patients who may have MACE events in the future,and the combined model of the two can further improve the predictive ability of identifying possible MACE events.
NETL
NSTL
万方数据
