Hepatocellular Carcinoma Histological Grade Prediction

OBJECTIVES This study aimed to compare quantitative assessments for predicting hepatocellular carcinoma (HCC) histological grades using magnetic resonance imaging. METHODS We retrospectively reviewed magnetic resonance imaging data from 49 patients with 54 surgically resected HCCs (11 well differentiated, 29 moderately differentiated, and 14 poorly differentiated). We measured the lesion-to-liver relative contrast ratio (RCR) on diffusion-weighted (DW), T2-weighted (T2W), and ethoxybenzyl-hepatobiliary images. We also evaluated contrast-to-noise ratio (CNR) on DW images, and the apparent diffusion coefficient. We compared the feasibility of these parameters in predicting HCC histological grade. RESULTS Areas under the receiver operating characteristic curves of both the DW RCR (97%) and DW CNR (97%) were significantly greater compared with those of the apparent diffusion coefficient (79%), T2W RCR (81%), and ethoxybenzyl-hepatobiliary RCR (80%) (P < 0.05, 0.001, and 0.001, respectively). CONCLUSIONS The DW RCR and DW CNR might represent accurate predictors of HCC histological grade.