Comparative Study Of The MRI With Hispathology In Quantifying Hepatic Steatosis

Purpose: Using histopathologic of hepatic steatosis as the gold standard,evaluation the value of MRI(in-phase and out-of-phase magnetic resonanceimaging, proton magnetic resonance spectroscopy) in quantitativediagnosis of nonalcoholic fatty liver disease(NAFLD) in rabbits.Materials and methods:40Male New Zealand rabbits were randomly dividedinto two groups: control group(n=15), fed chow diet; and model group(n=25),these rabbits were fed high fat diet to establish the model of nonalcoholicfatty liver disease. After6weeks, all rabbits underwent DGE-MRI and1H-MRS, subsequently were sacrificed, and their livers were taken forhistopathology observation. Histopathology was a gold standard,two-sample independent t test was employed to analysis theinter-comparison of hepatic fat fraction (HFF) and lipid water ratio (LWR).The sensitivity and specificity of HFF、LWR for diagnosing hepaticsteatosis (HS)≥5%were compared by receiver operating characteristic(ROC)analysis. Employing Kappa values to comparison the accuracy betweenHFF and LWR in quantitative estimation of HS.Result:1、Modeling success rate. High fat diets could successfully induceNAFLD model. there were none rabbits with hepatic steatosis in the controlgroup. Among25rabbits served as experimental animals,20rabbits weresuccessful to steatosis, the remain5rabbits were failed to formsteatosis. In the model group, the success rate was80%(20/25).2、Basedon histopathology diagnosis, the result was divided into negative groupand positive group, compare mean differences between these two groupsof HFF and LWR. Inter-comparison among different histopathology diagnosis showed that the values of HFF had distinct statistical significance,inhomogeneity of variance, t=-3.777，P＜0.05. The means of negative andpositive group was （8.710±1.886）%、（11.987±3.590）%, respectively.Inter-comparison among different histopathology diagnosis showed thatthe values of LWR also had distinct statistical significance,heterogeneity of variance, t=-5.639，P＜0.05. The means of negative andpositive group was （0.419±0.401）%、（1.684±0.995）%, respectively.3、The threshold values of DGE-MRI,1H-MRS for diagnosing HS≥5%. Thethreshold value for DGE-MRI to evaluate HS was9.027%, the sensitivityand specificity was0.90and0.80respectively. Youden index was0.7. Thethreshold value for1H-MRS to evaluate HS was0.852%, the sensitivity andspecificity was1.0and0.80respectively. Youden index was0.80.4、To analysis the consistency between DGE-MRI,1H-MRS and histopathology.The Kappa values were0.651,0.865respectively. Both were exceeded to0.40, implying that two indicator had a strong consistency withhistopathology.Conclusion: High fat diets could successfully induce NAFLD model. DGE-MRI,1H-MRS were two excellent indicators for quantitative identify NAFLD byanimal experient.