The Value Of Preoperative Gd-EOB-DTPA Enhanced MRI And IVIM-DWI In The Prediction Of Hepatocellular Carcinoma Microvascular Invasion

Objective:To investigate the diagnostic value of Gd-EOB-DTPA enhanced MRI and Intravoxel incoherent motion diffusion weighted imaging(IVIM-DWI)in predicting HCC with MVI.Materials and methods:1.Case data:40 patients with HCC were diagnosed according to clinical symptoms,laboratory and imaging examinations from December 2015 to December2017,including 33 males and 7 females,aged 33-78 years,with a mean age of 57.4years.The inclusion criteria for our study were as follows:(a)single HCC without grossly vascular tumor thrombosis or extrahepatic metastasis on preoperative MR imaging evaluation;(b)patients who underwent preoperative gadoxetic acid–enhanced and IVIM-DWI liver MR imaging within 1 week before surgery;(c)patients without a history of previous HCC treatment;(d)a hepatic mass with full histologic description available in the pathologic reports;and(e)MR imaging quality adequate for analysis.We excluded patients with HCC larger than 5 cm in maximum diameter because such tumors are known to have a greater risk of spreading through MVI.2.MR scanning technology and image post-processing:(1)MR scanning parameters:Siemens 3.0T verio MRI superconducting scanner,using 4-channel phased array body coil,Scan sequences include:(a)MR plain scan:coronal T2WI fat suppression(FS),axial T1WI in/out of phase,axial TIWI,axial T2WI FS;(b)IVIM-DWI:single shot spin echo diffusion weighted planar echo formation(SE-DW-EPI)sequence;(c)Gd-EOB-DTPA enhanced.(2)image post-processing1).Gd-EOB-DTPA enhanced MRI image analysisGd-EOB-DTPA enhanced MRI images were analyzed by Siemens verio 3.0T MRI post-processing workstation.Qualitative data include:arterial peritumoral enhancement,tumor margin,radiological capsule,peritumoral hypointensity on HBP.Quantitative data include:measurement of tumor signal corrected by muscle signal is recorded as LMSI,and according to the formula RE=(Post-LMSI-Pre-LMSI)/Pre-LMSI calculated RE(relative enhancement),RE in arterial phase recorded as AP-REE,RE in hepatobiliary phase recorded as HBP-RE.The qualitative data were re-observed when the diagnosis was inconsistent,and a consistent conclusion was reached after discussion,and the quantitative data were taken as the average value of the two physicians'measurements.2).IVIM-DWI image analysisThe image processing and measurement are carried out by using MITK Diffusion software(version:2014.10.02-win64).First,after importing the original image with multi-b value,adjusting the channel value to make the image clear and convenient to observe the focus,and then choosing the double index model to fit the image.Two radiologists combined with Gd-EOB-DTPA enhanced MRI images to draw ROIs on the principle of avoiding cysts,bleeding and necrotic areas,and describing the largest ROI.The D value,D*value and f value were measured three times by using the above method,and the average values were calculated and recorded.3.Analysis of pathological resultsA retrospective light microscopic observation was conducted by pathologists without knowing the results of MRI.The focus was on observing the MVI around the lesion.At present,the definition of MVI is not uniform around the world.The purpose of this study was to use the MVI pathological diagnostic criteria described by the Chinese Pathology Association in the Standard pathological Diagnostic guidelines for Primary Hepatocellular carcinoma(2015):The nests of cancer cells(?50 cancer cells)can be seen in the vascular lumen lined with endothelial cells under microscope.4.Statistical(analysis software SPSS 24.0,?=0.05)The counting data were described by the number of cases(percentage)and the measurement data by the mean±standard deviation(??s).1).Gd-EOB-DTPA enhanced MRI count data were checked by chi-square test to analyze whether it had statistical significance in predicting HCC MVI and to calculate its diagnostic efficacy.2).Gd-EOB-DTPA enhanced MRI and IVIM-DWI series were measured by independent sample t-test to analyze whether they had statistical significance in predicting MVI.3).The correlation between the quantitative data of MRI and IVIM-DWI sequence and the occurrence of MVI was analyzed by Spearman correlation analysis.4).The diagnostic efficacy of the statistical significance of the predicted MVI was compared using the Receiver Operator Characteristic Curve(ROC).Results:1.40 cases of HCC pathology and imaging general resultsIn the 40 lesions,18 lesions confirmed by pathology with MVI,and 22 lesions without MVI,the incidence of MVI was 45%.37.5%with arterial peritumoral enhancemen and 62.5%without arterial peritumoral enhancemen;77.5%had radiological capsule and 22.5%without radiological capsule;had smooth and non-smooth margin is 52.5%and 47.5%,respectively;55%of the patients with peritumoral hypointensity on HBP and 45%of the patients without peritumoral hypointensity on HBP;the tumor was high,middle and low differentiation accounted for 15%?62.5%?22.5%respectively.2.Comparison of 40 cases of HCC lesion count data between MVI group and non-MVI group(1)40 cases of HCC lesions with arterial peritumoral enhancement,non-smooth tumor margin,peritumoral hypointensity on HBP was significant difference in the prediction of tumor(P=0.011?0.028?0.012).Moreover,worse differentiated tumors were more likely to develop MVI than well differentiated ones.But tumor capsule was not significant difference in the prediction of tumor(P=0.546).It cannot be used as a risk factor for predicting MVI.(2)Gd-EOB-DTPA enhanced MRI imaging features peritumoral hypointensity on HBP has the best diagnostic efficacy,sensitivity?specificity?PPV?NPV in predicting MVI were 88.9%?72.7%?72.7%?88.9%,respectively.The specificity of arterial peritumoral enhancement was 85.7%,but the sensitivity was lower(61.1%),and the sensitivity and specificity of tumor margin both low in predicting MVI.3.Comparison of 40 cases of HCC focus measurement data between MVI group and non-MVI group(1)ADC,D of IVIM-DWI and AP-RE and HBP-RE of Gd-EOB-DTPA enhanced MRI had statistical significance in predicting MVI,but IVIM DWI parameter f and D~*had no statistical significance in predicting MVI.Moreover,the Dand HBP-RE of MVI group were lower than that of MVI group and AP-RE was on the contrary.(2)ADC,D of IVIM-DWI and HBP-RE of Gd-EOB-DTPA enhanced MRI were negatively correlated with MVI,and D was significantly correlated with occurrence of MVI,while AP-RE of Gd-EOB-DTPA enhanced MRI was positively correlated with MVI.(3)The area under the curve of HCC with MVI predicted by D value was 0.864,larger than AUC of ADC?AP-RE and HBP-RE(0.773?0.260?0.734).When the cut-off value of D?ADC?AP-RE and HBP-RE was 0.898?0.983?48.688?16.060 respectively,the sensitivity of predicting MVI was 100%?100%?63.6%?92.9%,and the specificity was 63.6%?54.5%?78.6%?45.5%,respectively.Conclusion:1.Gd-EOB-DTPA enhanced MRI imaging features arterial peritumoral enhancement,non-smooth tumor margin,peritumoral hypointensity on HBP can be used as a risk factor for preoperative predicting HCC with MVI.2.ADC?D of IVIM-DWI and AP-RE and HBP-RE of Gd-EOB-DTPA enhanced MRI can be used to predict MVI before operation.3.ADC?D of IVIM-DWI and AP-RE and HBP-RE of Gd-EOB-DTPA enhancement were negatively correlated with the occurrence of MVI.AP-RE was positively correlated with the occurrence of MVI.D of IVIM-DWI was significantly correlated with the occurrence of MVI.4.The peritumoral hypointensity on HBP was the best diagnostic efficacy of qualitative parameters in predicting preoperative HCC with MVI,and D of IVIM-DWI was the best diagnostic efficacy of quantitative parameters in predicting preoperative HCC with MVI.