Study On The Application Of Gd-EOB-DTPA Enhanced MRI Features For Predicting Vessels Encapsulating Tumor Clusters(VETC) In Hepatocellular Carcinoma

Objective:To investigate the relationship between vessels encapsulating tumor clusters(VETC)and other pathological characteristics of hepatocellular carcinoma(HCC).Materials and methods:One hundred and twenty-eight patients with pathologically confirmed HCC which were confirmed by postoperative pathology in our hospital were retrospectively analyzed.The pathological characteristics included VETC,capsule invasion,inflammatory infiltration of liver tissue,microvascular invasion(MVI),satellite focus,Ki-67 expression,CK-19 expression and Glypican-3 expression of HCC.Single factor chi square test was used to analyze the difference between VETC and the above pathological characteristics,between the macro/micro trabecular VETC and the above pathological characteristics;correlation coefficient heatmaps were plot to show the correlations between the VETC and other pathological characteristics,between the macro/micro trabecular VETC and other pathological characteristics.Results:There was significant differences between the VETC and MVI(p=0.015),Glypican-3 expression of HCC(P=0.060).There were significant differences between the macro/micro trabecular VETC and the degree of differentiation(high,high and medium differentiation defined as 0,medium,medium and low,low differentiation defined as 1;P=0.001),microvascular invasion(P=0.06),CK-19 expression(P=0.038),Ki-67 expression(P=0.045)and Glypican-3 expression(P=0.045).There was a moderate correlation between macro/micro trabecular VETC and differentiation of HCC with correlation coefficient r=0.5.Conclusion:The VETC is an a pathological characteristic independent of other pathological characteristics.The study of VETC was important for the precise diagnosis and treatment of HCC.Objectives:To investigate the noninvasive prediction of hepatocellular carcinoma(HCC)with vessels encapsulating tumor clusters(VETC)based on qualitative and quantitative imaging features of Gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid(Gd-EOB-DTPA)-enhanced MRI.Methods:One hundred and twenty-six patients with pathologically confirmed HCC who underwent Gd-EOB-DTPA enhanced MRI and immunochemical staining for CD34 were retrospectively evaluated in our institution.Preoperative imaging features of Gd-EOB-DTPA-enhanced MRI were qualitatively and quantitatively reviewed by radiologists.Significant variables for differentiating the VETC-positive and VETC-negative HCCs were identified in univariate and multivariate analyses.Receiver operating characteristic(ROC)analysis was performed to determine the optimal cut-off values for quantitative variables.The calibration curve was used to assess the goodness of fit and predictive performance of the multivariable logistic model.The nomogram based on the coefficient of multivariate analysis was constructed to evaluate the probability of VETC-positive HCCs.Results:The multivariate analysis showed that the serum AST level>40U 1-1(P=0.004),non-rim diffuse and heterogeneous arterial phase hyperenhancement(P=0.001),maximum diameter ≥ 5cm(P<0.001),tumor-to-liver signal intensity(SI)ratio of 1.135 or more on AP images(P<0.001),and tumor-to-liver SI ratio of 0.585 or less on HBP images(P=0.018)were significant predictors for predicting VETC-positive HCCs.Combing all five significant variables provided a diagnostic accuracy of 81.7%%,sensitivity of 83.9%,specificity of 80.9%for identifying VETC status.The area under the receiver operating characteristics curve(AUC)value of the logistical regression coefficient-based nomogram was 0.890(95%confidence intervals,0.824-0.946).Conclusions:Qualitative and quantitative imaging features of Gd-EOB-DTPA-enhanced MRI integrating laboratory examination can provide good diagnostic performance.Purpose:To develop and validate a radiomics model of gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid(Gd-EOB-DTPA)-enhanced MRI for prediction of VETC status of hepatocellular carcinoma(HCC).Patients and methods:This retrospective study included 135 pathologically confirmed HCC patients(training cohort:n=92;validation cohort:n=43).The 3D slicer was used to segment the region of interest(ROI)of the whole tumor on arterial phase(AP)and hepatobiliary phase(HBP)of Gd-EOB-DTPA)-enhanced MRI.Radiomics features were extracted from the AP and HBP.The least absolute shrinkage and selection operator(LASSO)logistic regression was used to select and build a radiomics score(Rad-score).AP,HBP,and combined AP and HBP radiomics model were constructed with their respectively Rad-score.ROC curve analysis,Delong test and net reclassification index(NRI)were used to compare the predictive performances of AP,HBP and combined radiomics model in training and validation cohort.Results:In the whole cohort,52 patients were VETC-positive,and 83 patients were VETC-negative.After LASSO analysis,in AP,HBP and combined radiomics models,there were 3,5 and 8 features were selected,respectively.In the training and validation cohorts,there were no significant differences between the AUC values of AP and HBP radiomics models(training cohort in AP and HBP:AUC=0.850 vs.0.810;validation cohort in AP and HBP:AUC=0.842 vs.0.822;P=0.282 vs.0.801).In the training cohort,the combined radiomics model showed the best predictive performance(AUC=0.870)compared with AP(AUC=0.850)and HBP(AUC=0.810)radiomics model with Delong test(P=0.050 vs.0.029);in the validation cohort,the AUC value of combined radiomics model(AUC=0.856)was higher than that of AP(AUC=0.837)and HBP(AUC=0.822)radiomics models with Delong test(P=0.525 vs 0.634),the combined radiomics model for predicting VETC-positive HCC could improve the diagnostic performance compared with AP and HBP radiomics model with NRI of 0.014 and 0.031,respectively.Conclusions:The radiomics model derived from AP and HBP images based on Gd-EOB-DTPA-enhanced MRI can predict VETC status of HCC.The combined radiomics model may be a better choice.