| Part 1 Preoperative histogram analysis model of intravoxel incoherent motion(IVIM)for predicting microvascular invasion in patients with single hepatocellular carcinomaPurpose:To evaluate the value of intravoxel incoherent motion(IVIM)histogram analysis in predicting microvascular invasion(MVI)of single hepatocellular carcinoma(HCC).Materials and Methods:The study enrolled 41 patients with pathologically proven HCCs who underwent IVIM diffusion-weighted imaging with nine b values and GD-EOB-DTPA contrast-enhanced magnetic resonance imaging(MRI).i)Patients’clinical characteristics such as age,sex,tumor size,origin of liver disease,Child-Pugh score,Edmondson-Steiner grade,background liver and a-fetoprotein(AFP)were recorded.ii)Two radiologists with experience in abdominal radiology,who were blinded to the clinical and histopathological information,independently and manually drew ROIs on axial DWI images with b1000.Whole-tumor ROIs were manually delineated to encompass as much of the entire HCC lesions in each slice as possible,which also included cystic necrotic regions.Axial T2WI SPIR images and hepatobiliary phase images were used as the reference to delineate the tumor,because the contour of the tumor was usually blured on DWI images.Histogram parameters including mean;skewness;kurtosis;and percentiles(5th,10th,25th,50th,75th,90th,95th)were derived from apparent diffusion coefficient(ADC),perfusion fraction(f),true diffusion coefficient(D),and pseudo diffusion coefficient(D*).Each tumor was measured twice by each observer,and the average values were considered.iii)In each HCC lesion,intratumoral fat deposition,gross classification of HCC,peritumoral enhancement,typical dynamic enhancement pattern,peritumoral hypointensity and tumor signal on hepatobiliary phase images were evaluated and recorded.Clinical data,imaging characteristics and quantitative histogram parameters were compared between HCCs with and without MVI.For significant parameters,receiver operating characteristic(ROC)curves were further plotted to compare the diagnosis performance for identifying MVI.Results:The tumor size and non-smooth tumor margin are significantly associated with MVI(all P<0.05).The mean,5th,10th,25th,50th,and 75th percentiles of D,and the 5th,10th,and 25th percentiles of ADC between HCCs with and without MVI were statistically significant(all P<0.05).The histogram parameters of D*and f showed no statistically significant differences between HCCs with and without MVI(all P>0.05).The area under the ROC curves(AUCs)were 0.707-0.874 for D and 0.668-0.720 for ADC.The area under the ROC curves(AUC)were 0.664 for tumor size and 0.683 for non-smooth tumor margin.The largest AUC of D(5th percentile)showed significantly higher accuracy than that of ADC,non-smooth tumor margin or tumor size(P = 0.009-0.046).With a cut-off of 0.403×10-3mm2/s,the 5th percentile of D value provided a sensitivity of 81%and a specificity of 85%in the prediction of MVI.Conclusion:Histogram analysis of IVIM based on whole tumor volume can be useful for predicting MVI.The 5th percentile of D was most useful value to predict MVI of HCC.Part 2 Preoperative CT texture analysis for predicting microvascular invasion in patients with single hepatocellular carcinoma(≤5cm)Purpose:This study was aimed to evaluate the diagnostic of texture analysis of computed tomography(CT)to discriminate between HCCs with MVI and without MVI in single hepatocellular carcinoma(HCC).Methods:The clinical date of 74 patients with pathologically proved HCC(≤5cm)were restrospectively collected.Patients were classified as microvascular invasion positive groups or microvascular invasion negative groups.Patients’ clinical characteristics such as age,sex,tumor size,origin of liver disease,Edmondson-Steiner grade,background liver and a-fetoprotein(AFP)were recorded.Two radiologists with experience in abdominal radiology,who were blinded to the clinical and histopathological information,independently and manually drew ROIs on axial arterial and portal venous phase CT images.Whole-tumor ROIs were manually delineated to encompass as much of the entire HCC lesions in each slice as possible,which also included cystic necrotic regions.The DICOM formal CT images were imported into Omni-Kinetics(GE Healthcare,China)software and the parameters of CT texture were calculated automatically.Texture analysis parameters including skewness,kurtosis,uniformity,energy,entropy,contrast,inverse difference moment(IDM)derived from CT imagings using unfilter fine texture.Each tumor was measured twice by each observer,and the average values were considered.CT texture parameters and clinical date were compared between HCCs with and without MVI.For significant parameters,receiver operating characteristic(ROC)curves were further plotted to compare the diagnosis performance for identifying MVI.Variables with P<0.05 in univariate logistic regression analysis were applied to a multivariate logistic regression analysis.Results:1.The diameter was statistically significant between HCCs with MVI and without MVI.2.The uniformity and energy texture parameters on arterial phase between HCCs with MVI and without MVI were statistically significant(all P<0.05).The uniformity,energy,entropy texture parameters on portal venous phase between HCCs with MVI and without MVI were statistically significant(all P<0.05).Other CT texture parameters between HCCs with MVI and without MVI were not statistically significant(all P>0.05).3.But,multiple logistic regression analysis showed that entropy on portal phase was independent risk factors for MVI(P=0.003).With a cut-off 9.931,the entropy parameter on portal phase provided a sensitivity of 73.8%and a specificity of 81.2%in the prediction of MVI.Conclusion:CT texture analysis is helpful in differentiated the HCCs with MVI and without MVI,especially the entropy on the portal venous phase has the most promising index in preoperative identifying MVI of single HCC(<5cm). |
PDF Full Text Request