Objective:To investigate the predictive value of mean whole-lung iodine perfusion(MeanIP)and perfusion de-fect volume ratio(PDVR)as derived from spectral CT in the risk stratification of acute pulmonary embolism(APE).Methods:A retrospective analysis was conducted on 80 patients diagnosed with APE via spectral CT scans.Patients were classified based on clinical and imaging data into high-risk(n=14),intermediate-high risk(n=30),and intermediate-low risk(n=36)groups.Clinical and spectral CT data were compared across these groups,and a Logistic model was developed using perfusion parame-ters to predict risk stratification of intermediate-high or high-risk groups.The model's efficacy was evaluated using the area un-der the ROC curve(AUC),and differences between models were assessed with the DeLong test.Results:Statistically significant differences were observed in MeanIP and PDVR of spectral parameters among the three groups(P＜0.05).The AUCs for the MeanIP,PDVR,and MeanIP+PDVR prediction models were 0.822,0.810,and 0.887,respectively.The DeLong test revealed significant differences between the combined MeanIP+PDVR model and the models using only MeanIP or PDVR,indicating the superior predictive performance of the combined model(P＜0.05).Conclusion:MeanIP combined with PDVR can effectively predict the risk stratification of APE,providing critical imaging evidence for clinical decision-making.