PURPOSE: To evaluate the accuracy of the sonographic hepatorenal ratio (HRR) in the diagnosis and grading of nonalcoholic steatosis, using biopsy as the reference. METHODS: Ultrasound (US) and liver biopsy were performed in 42 patients with nonalcoholic fatty liver disease. Forty healthy volunteers without steatosis at US and without risk factors for nonalcoholic fatty liver disease were also studied. The HRR was obtained by dividing the mean brightness level of region-of-interest pixels in hepatic parenchyma by that in renal parenchyma. Needle biopsy samples (hematoxylin-eosin stained) were classified as mild (5-33% fatty infiltration), moderate (>33-66%), or severe (>66%) steatosis. Spearman coefficient was used to evaluate the correlation between HRR and steatosis grade, analysis of variance for differences between subgroups, and receiver operating characteristic curve analysis for sensitivity and specificity. RESULTS: Significant correlation was found between HRR and histologic steatosis (r = 0.80, p < 0.01). The HRR cutoff for predicting steatosis was ≥1.24 (sensitivity, 92.7%; specificity, 92.5%). The mean ± SD HRRs in controls and steatosis subgroups were control 1.09 ± 0.13, mild 1.46 ± 0.24, moderate 1.52 ± 0.27, severe 2.04 ± 0.3 and were significantly different from each other except between mild and moderate steatosis subgroups. CONCLUSIONS: The HRR is a noninvasive, objective, and simple method that could be used to diagnose and grade hepatic steatosis.
PURPOSE: To evaluate the accuracy of the sonographic hepatorenal ratio (HRR) in the diagnosis and grading of nonalcoholic steatosis, using biopsy as the reference. METHODS: Ultrasound (US) and liver biopsy were performed in 42 patients with nonalcoholic fatty liver disease. Forty healthy volunteers without steatosis at US and without risk factors for nonalcoholic fatty liver disease were also studied. The HRR was obtained by dividing the mean brightness level of region-of-interest pixels in hepatic parenchyma by that in renal parenchyma. Needle biopsy samples (hematoxylin-eosin stained) were classified as mild (5-33% fatty infiltration), moderate (>33-66%), or severe (>66%) steatosis. Spearman coefficient was used to evaluate the correlation between HRR and steatosis grade, analysis of variance for differences between subgroups, and receiver operating characteristic curve analysis for sensitivity and specificity. RESULTS: Significant correlation was found between HRR and histologic steatosis (r = 0.80, p < 0.01). The HRR cutoff for predicting steatosis was ≥1.24 (sensitivity, 92.7%; specificity, 92.5%). The mean ± SD HRRs in controls and steatosis subgroups were control 1.09 ± 0.13, mild 1.46 ± 0.24, moderate 1.52 ± 0.27, severe 2.04 ± 0.3 and were significantly different from each other except between mild and moderate steatosis subgroups. CONCLUSIONS: The HRR is a noninvasive, objective, and simple method that could be used to diagnose and grade hepatic steatosis.