Effect Analysis Of Field Strength In MR Diffusion Weighted Imaging Measurement And The Diagnostic Performance Of DWI In Predicting Histological Differentiation Of Hepatocellular Carcinoma

Purpose:To quantitatively compare the apparent diffusion coefficient (ADC) values of 3.0-T and 1.5-T MR diffusion-weighted imaging (DWI) protocols for the detection of liver focal lesions.Materials and Methods:In total,26 consecutive patients with 28 liver focal lesions prospectively underwent both 1.5-T and 3.0-T DWI of the liver. The protocols included respiratory-triggered (RT), breath-holding (BH), and free-breathing (FB) acquisitions. The ADC values were measured at both field strengths using three methods:ROIs with the minimum ADC, the largest solid part, and the maximum diameter of lesions. Bland-Altman tests and paired t-tests were used to compare ADC values in the liver focal lesions obtained at 1.5-T and 3.0-T.Results:The 3.0-T and 1.5-T protocols differed significantly with regard to the ADC values of the RT, BH, and FB acquisitions, for ROIs of both the largest solid part (p= 0.005,p= 0.014, and p= 0.022, respectively) and maximum diameter of lesions (p< 0.001,p= 0.001, and p= 0.001, respectively).Conclusions: When using DWI for quantitative analysis of liver focal lesions, field strength may be a negative factor, and ADCs from different field strength could be not mixed together, no matter with RT, BH, or FB-DWI.Purpose:To define correlations between the pathological grades of hepatocellular carcinomas (HCCs) and apparent diffusion coefficients (ADCs) derived using breath-holding diffusion-weighted imaging (BH-DWI).Methods:We retrospectively evaluated 94 patients (105 lesions) with pathologically proved HCC who underwent hepatic DWI on a 3.0-T MR platform. HCCs were divided into five groups:well (w-) differentiated (n=10), well-to-moderately (wm-)differentiated(n=11), moderately(m-) differentiated(n=51), moderately-to-poorly (mp-) differentiated(n=20), and poorly(p-) differentiated(n=13). The ADCs of carcinomas among different histological grades were compared by one-way analysis of variance. Spearman’s rank correlation test was used to analyze correlations between the degree of histopathological differentiation and ADC value. The Bonferroni correction for the comparison between groups of differentiation was calculated.Results:The BH technique yielded ADC values that differed significantly by the extent of differentiation (F=8.392, p<0.001), and there was significant negative correlation between the extent of differentiation and ADCs (r=-0.462, p< 0.001).The mean ADC values of poorly differentiated HCCs was significantly lower than the well, well-to-moderately, moderately and moderately-to-poorly differentiated HCCs(p values were 0.000,0.000,0.003 and 0.031, respectively).Conclusion:ADC obtained with BH-DWI may be of value to noninvasively predict HCC tumor differentiation, and the extent of histological HCC differentiation was inversely correlated with ADCs.Objective: To investigate whether the histological differentiation of the hepatocellular carcinoma (HCC) could be predicted by the parameters of intravoxel incoherent motion diffusion weighted imaging (IVIM-DWI) and apparent diffusion coefficient (ADC).Materials and Methods:73 patients with 73 surgically confirmed HCCs were prospectively performed with liver diffusion-weighted imaging (DWI) with nine b values (0-1000 sec/mm2). The apparent diffusion coefficient (ADC) values, diffusion coefficient (D), pseudodiffusion coefficient (D*), and perfusion fraction (f) were calculated. The region of interest (ROI) was placed in the solid part of the maximum axial of the lesion. Two radiologists each measured once. Variance analysis was used to analyze the differences of ADC, D, D*, and f in different HCCs’pathological differentiations, and the correlations between the four parameters and the pathological differentiations of HCCs were respectively analyzed by Spearman correlation test. For D values and ADC values, the diagnostic efficiencies in the different pathological differentiation of HCCs were performed with receiver operating characteristic (ROC) curve analysis.Result: The D values (F=28.66,p<0.001) and ADC values (F=12.54, p<0.001) from the IVIM-DWI in the three different HCCs’ pathological differentiations were both statistically significant. And the D values (r=-0.748,p<0.001) and ADC values (r=-0.329,p=0.005<0.05) were both negatively correlated with the pathological differentiation of HCCs. For the well-differentiated HCC, the diagnostic sensitivity of D value was 90%, the specificity was 82.5%, the Youden index was 0.725, area under the curve (AUC) was 0.896, and the p value was less than 0.001; while the ADC values had no significant diagnostic efficiency (p=0.949>0.05). For the poorly differentiated HCC, the diagnostic sensitivities of the D values and the ADC values were respectively 78.7% and 85.2%, the specificities were respectively 100% and 83.3%, the Youden index were respectively 0.725 and 0.686, and AUCs were respectively 0.934 and 0.878.Conclusion:The D values and ADC values from IVIM-DWI were both negatively correlated with the pathological differentiations of HCCs, while the D values had better diagnostic performances in differentiating the well-differentiated and poorly differentiated HCCs.