Diagnostic accuracy and significance of clinical analysis of liver stiffness measurement by Fibro touch in metabolic-related fatty liver disease

Abstract Background and Aim: We aimed to explore risk factors and construct a model of MAFLD with fibrosis fibrosis and compare the performance of the panels of APRI, FIB-4 and BARD.Methods: We retrieved clinical information and enrolled 3671 patients. Subjects were assigned to four groups: MAFLD, MAFLD with fibrosis, and MAFLD with advanced fibrosis. Multivariate regression analysis and randomforest model were to construct the model of MAFLD with fibrosis, the receiver operating characteristics (ROC) were used to compare the diagnostic efficacy of predictive model, APRI, BARD, FIB-4. Results: ①The proportion of fibrosis in MAFLD was higher than that of non-MAFLD group; Compared with the control group, advanced fibrosis (≥F3) has no statistical significance. ②Four variables were selected to build the model of MAFLD. The diagnostic accuracy (AUROC=0.730) was superior to that of APRI and random forest model. ③Multivariate logistic analysis showed that increased BMI and AST was risk factors for advanced fibrosis and elevated PLT was a protective factor for advanced fibrosis; The diagnostic power of the model (AUROC=0.714) was superior to that of the FIB-4 and BARD.Conclusion: The model of fibrosis stage with MAFLD is superior to other non-invasive markers of fibrosis of MAFLD, which can assist clinicians in screening and warning for potential risks of MAFLD with fibrosis.