Multimodal CT Based On Artificial Intelligence In Evaluating The Prognosis Of Patients With Acute Ischemic Stroke After Endovascular Therapy

Objective:To investigate the value of multimodal CT based on artificial intelligence in evaluating influencing factors of AIS patients with anterior circulation large vessel occlusion(AIS-LVO)after endovascular therapy(EVT)and to establish a predictive model.The aim of which is to provide theoretical basis for improving the clinical prognosis of patients.Methods:Patients with AIS who underwent EVT from February 2021 to January 2023 in the Second Affiliated Hospital of Nanchang University were retrospectively collected.General clinical data of the patients were collected.All patients performed non-contrast CT(NCCT),CT angiography(CTA),and CT perfusion(CTP)before treatment.According to the modified Rankin Scale(m RS)score determined at 3months after treatment,the patients were divided into good prognosis group and poor prognosis group.Multivariate Logistic regression analysis was used to screen the independent predictors for prognosis of AIS patients with anterior circulation large vessel occlusion after EVT and to establish predictive models based on clinical factors,imaging factors and combined clinical and imaging factors.ROC curve was used to evaluate the predictive effect of 3 models.Results:A total of 114 patients were included,66 patients in good prognosis group and48 patients in poor prognosis group.Statistically significant differences in age(61.39±12.05 VS 66.71±11.83),NIHSS score at admission [7.00(4.00-9.00)VS14.25(11.25-17.00)],puncture to reperfusion time(PTR)(43.48±24.07 VS 56.85±34.85),ASPECTS score [8.00(7.00-8.00)VS 4.00(3.00-7.00)],Infarct core volume(12.32±10.24 VS 45.16±30.28),CBF in the infarct core(16.96±5.47 VS 14.63±4.87),clot burden score(CBS)[8.00(7.00-8.25)VS 6.50(4.00-8.00)] and Miteff score were statistically significant(P<0.05).For ASPECTS regions,there were significant difference in C,L,I,M1,M2,M4 and M6(P<0.05).The factors with statistical difference between the two groups were analyzed by multivariate Logistic regression.Clinical model showed age(OR=1.061,95%CI:1.013-1.112,P=0.013),NIHSS score at admission(OR=1.482,95%CI:1.289-1.705,P< 0.001)were independent predicto-rs for prognosis of AIS,and the AUC of model was 0.899(95%CI:0.842-0.956)with sensitivity of 72.9%,specificity of 90.9%,accuracy of 81.6%,positive predictive value of 80.0%,negative predictive value of 82.6%.Imaging model showed infarct core volume(OR=1.060,95%CI:1.009-1.115,P=0.021),Miteff score(OR=1.014,95%CI 1.001-1.027,P=0.035)were independent predictors for prognosis of AIS,and the AUC of model was 0.914(95%CI:0.859-0.968)with sensitivity of 81.3%,specificity of 89.4%,accuracy of 85.9%,positive predictive value of 84.8%,negative predictive value of 86.8%.Combining clinical and imaging model showed age(OR=1.082,95%CI:1.018-1.150,P=0.011),NIHSS score at admission(OR=1.285,95%CI:1.078-1.531,P=0.005),Infarct core volume(OR=1.068,95%CI:1.005-1.135,P=0.035)and Miteff score(OR=1.019,95%CI:1.003-1.035,P=0.022)were indepen-dent predictors for prognosis of AIS,and the AUC of model was 0.953(95%CI:0.916-0.990)with sensitivity of 87.5%,specificity of 93.9%,accuracy of 89.5%,positive predictive value of 92.9%,negative predictive value of 87.5%.Delong’s test found that clinical model(Z=-2.422,P=0.015)and imaging model(Z=2.196,P=0.028)had lower predictive effect compared with the clinical combined with imaging model,and the difference was statistically significant.Conclusions:(1)Age,NIHSS score at admission,infarct core volume and Miteff score are independent predictors for prognoses of AIS patients with anterior circulation large vessel occlusion after EVT,which have certain value in predicting the postoperative functional prognosis of AIS patients with EVT.(2)AUC value of the predictive model of combining clinical and imaging was0.953,with sensitivity of 87.5%,specificity of 93.9%,accuracy of 89.5%,positive predictive value of 92.9%,negative predictive value of 87.5%.The predictive model of clinical combined with imaging was superior to the clinical model and imaging model,and could effectively improve the predictive effect of AIS-LVO patients with EVT.