The Value Of Quantitative Parameters Of Dual Energy CT In Displaying Early Hemorrhage And Predicting Delayed Intracerebral Hemorrhage After Thrombectomy In Patients With Acute Ischemic Stroke

Objective: To explore the value of quantitative parameters of dual energy CT Rho/Z in displaying early hemorrhage and predicting delayed intracerebral hemorrhage after thrombectomy in acute ischemic stroke.Methods: This is a retrospective study.During February 2018 to October 2019,80 patients at neurology department of our hospital underwent dual energy computed tomography scans immediately after arterial thrombectomy.According to the results of CT,the patients were divided into several groups: VNC ICH / non-VNC ICH,delayed ICH /non-ICH.The patients were divided into PH / non-PH by using the European Cooperative Acute Stroke Study as the classification method.A doctor manually sketched the ROI on the post-processing software DE Rho/Z twice every four weeks,and recorded the DECT parameters of each group: CT numbers of 80 k V and Sn150 k Vp,CT numbers of simulated120 k V,dual energy index,electron cloud density,effective atomic number and sketched area.The relationship between the parameters and the occurrence of VNC ICH,delayed ICH and PH was tested to determine the quantitative parameters which were significantly related to the occurrence of hemorrhage.To compare the diagnostic efficacy of different quantitative parameters and calculate the best cutoff for predicting delayed ICH after thrombectomy.Two radiologists determined whether there was ICH on the virtual unenhanhanced reconstructions and iodine map of the post-processing image,and compared with the follow-up CT to determinethe diagnostic efficacy of virtual unenhanhanced reconstructions and iodine map in the diagnosis of immediate ICH after thrombectomy.By comparing the diagnostic efficacy of quantitative parameters related to immediate postoperative ICH with that of virtual unenhanhanced reconstructions,a better post-processing method was determined.Results: According to the exclusion criteria,62 patients were finally enrolled in the group,14 patients had no hyper attenuation all the time,48 patients had hyper attenuation in the first DECT after thrombectomy,the number of each group was VNC ICH(12)/ non-VNC ICH(50),delayed ICH(25)/ non-ICH(37),PH(20)/ non-PH(42).(1)It was found that the CT numbers of 80 k V and Sn150 k Vp,the CT numbers of simulated 120 k V,the electron cloud density and the sketched area were not statistically significant.(2)The sensitivity 、 specificity 、 positive predictive value 、negative predictive value and accuracy of the diagnosis of intracranial hemorrhage were 65.79%、100%、100%、64.86% and 79.03%,respectively.Zeff is the better parameter to diagnose intracranial hemorrhage with Rho /Z.when the best cutoff value is 9.47,sensitivity 、 specificity 、 positive predictive value 、 negative predictive value and accuracy are 52.38% 、96.3%、91.67%、72.22% and 85.42%,respectively.Compared with the two diagnosis methods,the specificity and positive predictive value of virtual unenhanced reconstructions are higher.(3)DEI was 0.072(0.050-0.115),which was significantly higher than 0.042(0.029-0.052),u = 107,P <0.001,Zeff was 10.18(9.35-11.48),which was significantly higher than9.07(8.55-9.46),u = 110.5,P < 0.001,and the best cutoff values of DEI and Zeff were 0.045 and 9.355,respectively.(4)The parameters of DEI were 0.079(0.051-0.122),u = 73.5,P < 0.001,Zeff 10.40(9.39-11.69)and9.07(8.55-9.46),u = 77.5,P < 0.001,respectively.The best cut-off of DEI and Zeff were 0.058 and 9.09,respectively.Conclusion:1 Both virtual unenhanced reconstructions and Rho/Z post-processingof dual energy CT are valuable for the early recognition of intracranial hemorrhage after arterial thrombectomy in acute ischemic stroke.However,compared with them,the specificity and positive predictive of virtual unenhanced reconstructions are higher(100%).Therefore,the first choice for the recognition of early intracranial hemorrhage is virtual unenhanced reconstructions.2 The DEI and Zeff of Rho/Z are both valuable for distinguishing between delayed intracerebral hemorrahage and non-delayed intracerebral hemorrahage.In the recognition of delayed intracerebral hemorrahage,Zeff is more valuable,and the best cut-off is 9.355.3 The DEI and Zeff of Rho/Z are both valuable to distinguish parenchymal hematoma from non-parenchymal hematoma.If we identify parenchymal hematoma,Zeff is more valuable,and the best cutoff is 9.645.If we exclude non-parenchymal hematoma,DEI is more valuable,and the best cutoff is 0.058.