Texture Analysis Of Gd-EOB-DTPA Magnetic Resonance Imaging Of On Prediction Of Therapeutic Response Of Hepatocellular Carcinoma To Transcatheter Arterial Chemoembolization:A Pilot Study

Objective:The objective of this study was to assess the value of texture analysis for prediction of therapeutic response of hepatocellular carcinoma to transcatheter arterial chemoembolization with pretherapeutic Gd-EOB-DTPA magnetic resonance imaging.Material and methods:In total 41 patients(male 35 cases,female 6 cases,range 20-80 years,average age 56.17±11.84 years)with confirmed to hepatocellular carcinoma by histopathology or American association for the study of liver diseases criteria in our hospital,between January 2014 to November 2016.The clinical and imaging data of these patients were collected,and there were 66 liver cancer lesions.Among them,16 patients had multiple lesions(9 had 2,5 had 3,and 2 had 4).All the patients were scanned with a 3T clinical scanner(Ahiveva TX,Philips,the Netherlands).Gd-EOB-DTPA was used as contrast agent during the exam.Conventional MRI sequences included mDixon,axial T2WI fat suppression,IVIM DWI sequence,Dynamic enhanced scan and hepatobiliary phase in 20 minites after injection of Gd-EOB-DTPA,using e-Thrive interpolation(three-dimensional volume rapid scanning T1 high-resolution isotropic displacement excitation)fat suppression of T1 weighted gradient echo imaging technology,TR = 3,TE = 1.51 ms,FOV = 375 mm × 304 mm,matrix=384 = 1 mm,with a thick layer of layer spacing = 0 mm.The maximal diameter of the lesion was measured according to T1WI sequence,T2WI/FS,enhanced scan sequence,IVIM DWI sequence and hepatobiliary phase sequence in the PACS workstation.Late arterial phase,portal vein phase,hepatobiliary phase image were derived from PACS workstation in DICOM format,then imported into the Omni-Kinetics(OK)software,two abdominal physicians with 3 years experience in MR imaging diagnosis sketched region of interest manually.Two kinds of texture parameters,including Histogram and Gray Level co-occurrence Matrix(GLCM),were extracted from the tumor region from the hepatobiliary phase image.TACE was conducted using a mixture of oxaliplatin,epirubicin,mitomycin,lipiodol,and contrast agent.The dose of lipiodol was dependent on tumor size and vascularity,with 50-100 mg of oxaliplatin,10-20mg of epirubicin,10-20mg mitomycin and lipiodol used per session.Subsequently,embolization was performed using polyvinyl alcohol particles or gelatin sponge,and occlusion of target vessels and absence of additional tumor blood supply confirmed.All patients was reviewed 4-6 weeks after TACE,the digital subtraction angiography was performed to clarify the deposition of iodine oil and tumor necrosis in patients with liver lesions.In the current study,each tumor was classified into two groups:without residual or recurrence and with residual or recurrence.SPSS 20.0 software was used for statistical analysis of each parameter.The measurement data is expressed as the mean number plus or minus standard deviation.Calculate the intraclass correlation coefficient evaluation consistency between two physicians measuring characteristic parameters of observer,greater than 0.8 as a good consistency.Independent-Sample T test was used to calculate in all the parameters which were in normal distribution and homogeneity test of variance.On the contrary,mann-whitney U test is adopted.The 3 different phases image parameters applied Logistic regression analysis for multiparameter joint analysis.The receiver operating characteristic curve was calculated,obtained the area under the curve of the MR texture analysis quantitative parameters to predict without residual or recurrence for hepatocellular carcinoma after TACE.In this study,the confidence interval of the mean was set to 95%,and P<0.05 was considered statistically significant.Results:1.Therapy ResultsAccording to the standard of DS A assessment of lesion treatment response,41 patients with a total of 66 lesions,of which 36 assessed to without residual or recurrence group,30 assessed to with residual or recurrence group.2.General data analysisThere was statistical difference in tumor size between the without residual or recurrence group and the with residual or recurrence group(P=0.041),and the without residual or recurrence group had a smaller tumor size.There was no statistical difference between the two groups’ tumor location and the presence of the capsule appearance(P>0.05).3.Comparison of texture analysis parameters between the two groupsAll the intraclass correlation coefficient of parameter between the two observers were more than 0.8.There were 3 meaningful parameters in arterial phase:Entropy,Inverse Difference Moment and Prominent clustering.The area under curve(AUC)of Entropy and Prominent clustering were 0.713 and 0.739 respectively,indicating good diagnostic efficiency.There were 2 meaningful parameters in portal vein phase:Variance and Prominent clustering.There were 3 meaningful parameters in hepatobiliary phase:Skewness,Energy and Entropy.The AUC of them were 0.769,0.734 and 0.739 respectively,indicating good diagnostic efficiency.4.Multiparameter joint analysis of texture analysisThe AUC of the combination of texture analysis parameters in arterial phase,portal vein phase,hepatobiliary phase were 0.858,0.715 and 0.854,respectively.The efficiency of all the eight parameters joint diagnosis is highest,which is higher than a single parameter,and it also has improved sensitivity and specificity.Conclusions:1.There were 8 texture parameters able to predict postoperative curative effect after TACE:Entropy,Inverse Difference Moment and Prominent clustering in arterial phase,Variance and Prominent clustering in portal vein phase,Skewness,Energy and Entropy in hepatobiliary phase.2.The AUC value is greater in multi-parameter joint ROC curve analysis than singel parameter.The efficiency of all the eight parameters joint diagnosis is highest,which has improved sensitivity and specificity.3.The Texture analysis of Gd-EOB-DTPA magnetic resonance imaging can predict therapeutic response of hepatocellular carcinoma to transcatheter arterial chemoembolization,providing an important basis for preoperative evaluation,and is of great guiding significance for the selection of clinical treatment options in the next stage.