The aim of this study was to evaluate the usefulness of gadobenate dimeglumine–enhanced magnetic resonance imaging in characterizing the grade of hepatocellular carcinoma (HCC) using the signal intensity (SI) of the erector spinae as internal reference.Materials and Methods
Clinical data of 40 patients (a total of 44 lesions) confirmed by pathology for HCC were retrospectively reviewed. Gadobenate dimeglumine–enhanced magnetic resonance imaging was performed in all patients, and SI of lesions (SIles), liver parenchyma around the lesions (SIhep), erector spinae (SImus) and standard deviation of SI of the surrounding noise (SDnoi) on nonenhanced T2WI, nonenhanced T1WI, and contrast-enhanced T1WI (in both arterial and hepatobiliary phase [AP and HBP]) were measured, respectively. Contrast-to-noise ratio (CNR) were separately defined as CNR1 ([SIles − SIhep]/SDnoi) and CNR2 ([SIles − SImus]/SDnoi). Statistical analyses were performed using one-way analysis of variance, least significant difference test, logistic regression analysis, Spearman rank correlation, and receiver operating characteristic curves analysis.Results
Forty-four HCCs, including 3 well-differentiated HCCs, 26 moderately differentiated HCCs, and 15 poorly differentiated (PD) HCCs, were confirmed. On logistic regression analysis, only CNR2 in the HBP was predictor of PD HCCs (P = 0.015, odds ratio = 1.040). The size of lesions, CNR1 in the AP, CNR2 in the AP, and CNR2 in the HBP, showed significant correlations with the degree of differentiation (correlation coefficients = −0.371, 0.435, 0.503, and 0.512, respectively; P = 0.013, 0.003, 0.001, and 0.000, respectively). Contrast-to-noise ratio 2 in the HBP with the cutoff of less than 4.56 could distinguish moderately differentiated HCCs from PD HCC with the sensitivity and specificity of 84.6% and 60.0%, respectively.Conclusions
Relatively low arterial enhancement and low CNR2 value in the HBP are predictive for poor histological grade of HCCs.