The comparison between patients with or without steatosis are shown in Table 2. By univariate analysis, ALT (P = 0.0004), AST (P = 0.0005), HOMA-IR (P = 0.0043), GGT (P = 0.0044), BMI (P = 0.0049), fasting insulin (P = 0.0050) and age (P = 0.0379) were significantly associated with liver steatosis. By multivariate analysis, L/S ratio (P = 0.012; OR, 0.501; CI, 0.29–0.86), AST (P = 0.022; OR, 1.253; CI, 1.03–1.52), and HOMA-IR (P = 0.041; OR, 1.335; CI, 1.01–1.76) were significantly associated with liver steatosis (Table 3). The percentage of steatosis was calculated by using Dynamic cell

count BZ-H1C Selleck Y27632 software after measuring the real steatotic area of liver specimens by BIOREVO BZ-9000 microscope. Figure 2 (a) shows the relationship between steatotic grades and percentage of steatosis. Percentage of steatosis was significantly correlated with the steatotic grade (r = 0.86, P = 1.85 × 10−18). L/S ratios were: S0, 1.16 ± 0.20 (mean ± SD); S1, 0.88 ± 0.28; S2, 0.76 ± 0.20; and S3, 0.40 ± 0.18, respectively (Fig. 2a). L/S ratio was negatively

correlated with steatotic grade (r = −0.77, P = 7.94 × 10−13) (Fig. 2b). Because the percentage of steatosis was evaluated from the liver specimens, this was expected to show the real fat deposition in the liver. Therefore, the relationship between the percentage of steatosis and L/S ratio was compared Urease (Fig. 3), and showed the significant negative correlation. Based on this curve, L/S ratio could Apitolisib manufacturer be expected by the formula: (−0.6356 × [log percentage of steatosis] + 1.2964). The AUROC for the diagnosis of steatosis was 0.886 for L/S ratio (Fig. 4). The optimum cut-off value for L/S ratio to exclude steatosis was 1.1 which produced sensitivity and specificity values of 83.3% and 93.3%, respectively, as well as a positive predictive value of 97.6% and a negative predictive value of 63.3%. THIS STUDY ATTEMPTED to elucidate

the accuracy of histological diagnosis of fatty deposition by pathologists and also to demonstrate the optimal cut-off value for the diagnosis of fatty liver on CT. As a result, evaluation by histological findings and steatotic grades assessed by a pathologist were comparatively accurate compared with imaging modalities such as CT in this case. That is, steatotic grades diagnosed by a pathologist were significantly correlated with the percentage of steatosis and L/S ratio. From the point of laboratory findings, patients with steatosis were younger, had higher BMI, and increased AST, ALT, GGT and HOMA-IR. By multivariate logistic regression analysis, AST and HOMA-IR were independently associated with steatosis. For the non-invasive assessment of liver steatosis, imaging analyses and histological findings were compared to elucidate the utility of L/S ratio on CT.