Assessment of Hepatic Steatosis in Patients Undergoing Liver Resection: Comparison of US, CT, T1-weighted Dual-Echo MR Imaging, and Point-resolved1H MR Spectroscopy

Abstract

PURPOSE: To compare the diagnostic performance of ultrasonography (US), computed tomography (CT), T1-weighted dual-echo magnetic resonance (MR) imaging, and point-resolved proton (hydrogen 1[(1)H]) MR spectroscopy in the assessment of hepatic steatosis in patients undergoing liver resection. MATERIALS AND METHODS: This prospective study was approved by the institutional review board, and patients gave written informed consent. US, CT, T1-weighted MR imaging, and (1)H MR spectroscopy were performed preoperatively in 46 patients. Imaging results were correlated (Spearman correlation coefficient) with histopathologic analysis of results of intraoperative liver biopsies. To assess differences between groups, one-way analysis of variance was used. Sensitivity and specificity were calculated for each imaging modality by using receiver operating characteristic curve analysis, with a histopathologic cut-off value of 5% macrovesicular steatosis. Differences in sensitivity and specificity were assessed by means of McNemar analysis. RESULTS: At histopathologic examination, 23 patients had no (0%-5%) macrovesicular steatosis, 11 had mild (5%-33%), nine had moderate (33%-66%), and three had severe (>66%). MR imaging and (1)H MR spectroscopic measurements of hepatic fat had stronger correlation with histopathologic steatosis assessment (r = 0.85, P < .001 and r = 0.86, P < .001, respectively) than did US (r = 0.66, P < .001) and CT (r = -0.55, P < .001). Only T1-weighted MR imaging and (1)H MR spectroscopy showed differences across steatosis grades: none versus mild (P = .001 for both), mild versus moderate (P < .001 for both), and moderate versus severe (P = .04 and .01, respectively). Sensitivity of US, CT, T1-weighted MR imaging, and (1)H MR spectroscopy was 65% (13 of 20), 74% (17 of 23), 90% (19 of 21), and 91% (21 of 23), respectively, and specificity was 77% (17 of 23), 70% (14 of 20), 91% (20 of 22), and 87% (20 of 23), respectively. CONCLUSION: In contrast to US and CT, T1-weighted MR imaging and (1)H MR spectroscopy strongly correlate with histopathologic steatosis assessment and are able to demonstrate differences across steatosis grades. T1-weighted dual-echo MR imaging and (1)H MR spectroscopy had the best diagnostic accuracy in depicting hepatic steatosis.