| BACKGROUND AND AIMS: The prevalence of non-alcoholic fatty liver disease(NAFLD)is increasing gradually,and NAFLD-related complications and disease mortality have also increased.Therefore,the development of early screening strategies is essential to reduce the burden of NAFLD disease.In order to achieve the early diagnosis of NAFLD patients,this study aims to explore the clinical value of MRI proton density fat fraction in the quantitative diagnosis of liver and pancreatic steatosis in patients with nonalcoholic fatty liver disease and to compare the relationship between pancreatic fat content and liver fat content.METHODS: We screened 2083 patients with suspected NAFLD in Shanghai General Hospital from March 2018 to January 2020 who underwent MRI-PDFF test,excluding patients with other chronic liver diseases and incomplete clinical data.Of the patients,611 patients were selected and included in the study which aims to quantitatively detect liver and pancreas fat content with MRI.611 patients were randomly divided into a model group(n = 456)and a verification group(n = 155).Taking MRI-PDFF as the reference standard,the influencing factors that were significantly related to the degree of fatty changes in the model group were selected to establish a binary logistic In the regression model,evaluate the diagnostic value of the model,and explore the correlation between pancreatic fat content and liver fat content.RESULTS: We include gender,AST/ALT,TG,LDL,BMI,and paraumbilical fat thickness as independent variables to establish a binary logistic regression model A and B.The results showed that models A and B had 0.73(95% CI: 0.68-0.78)and 0.74(95% CI: 0.69-0.79)AUROC when diagnosing NAFLD,respectively,and the ability to predict liver steatosis was better than FLI index and HIS index.All patients were divided into normal group and NAFLD group with liver MRI-PDFF = 5%,3.71%,5.2%as the boundary.There was no statistically significant difference in pancreatic MRIPDFF between the two groups;The fat content in the head of the pancreas is less than the fat content of the body and tail of the pancreas;the fat content of the pancreas is not statistically different between patients with different grades of liver steatosis;there is no significant correlation between the fat content of the pancreas and the liver fat content of the patients(rs = 0.047,p = 0.222).CONCLUSIONS: The proton density fat fraction of MRI can be used to quantify the fatty change of liver and pancreas in patients with nonalcoholic fatty liver disease.Models A and B based on serological and imaging information have certain clinical value in diagnosing the degree of liver fat in patients with NAFLD. |
