S1390 The Accuracy and Reliability of AST to Platelet Ratio Index, FIB-4, FIB-5, and NAFLD Fibrosis Scores in Detecting Advanced Fibrosis in Patients With Metabolic-Associated Fatty Liver Disease

Introduction: Metabolic-Associated Fatty Liver Disease (MAFLD) is a rapidly growing health issue, defined by hepatic steatosis with a disturbed metabolic state. The severity of liver damage in MAFLD can range from simple steatosis to advanced fibrosis and liver failure. MAFLD is also associated with a variety of extra-hepatic complications including cardiovascular diseases, insulin resistance, chronic kidney disease, and colorectal cancer. Although liver biopsy is the gold standard for assessing liver fibrosis, several non-invasive tools have been developed in this regard. We aimed to evaluate the accuracy and reliability of AST to Platelet Ratio Index (APRI), fibrosis index based on 4/5 factors (FIB-4/FIB-5), and non-alcoholic fatty liver disease fibrosis score (NFS) in predicting advanced liver fibrosis in MAFLD patients. Methods: We recruited adult patients (≥18 y/o) diagnosed with MAFLD in this cross-sectional study. Based on FibroScan® fibrosis (F) scores, patients were categorized into non-advanced (F0-F2) and advanced (F3-F4) fibrosis groups. APRI, FIB-4, FIB-5, and NFS were calculated using laboratory biomarkers and anthropometric data. We measured the area under receiver operating characteristic (AUROC) curve for the aforementioned tests in detecting advanced fibrosis. Sensitivity, specificity, positive/negative predictive values, positive/negative likelihood ratios, and accuracy were also calculated for each test using the optimal cut-off. Reliability analysis was done using intra-class correlation coefficient (ICC). Results: A total of 100 eligible MAFLD patients were recruited. Six (6%) patients had advanced fibrosis (F3-F4) and 95 (95%) were non-advanced (F0-F2). The AUROC curves for APRI, FIB-4, FIB-5, and NFS were 0.913 (95% Confidence Interval (CI): 0.843-0.983), 0.894 (95%CI: 0.767-1.00), 0.020 (95%CI: 0-0.054), and 0.945 (95%CI: 0.864-1), respectively. We calculated the ICC of 0.780 (95%CI: 0.701-0.843) for F score, APRI, FIB-4, and NFS. Sensitivity, specificity, positive/negative predictive values, positive/negative likelihood ratios, and accuracy of each test are summarized in Table 1. Conclusion: APRI, FIB-4, and NFS are accurate and reliable measurement tools in detecting advanced fibrosis in MAFLD patients, while FIB-5 score failed to demonstrate enough accuracy. Using available data to develop a modified non-invasive test comprising of strongest predictors of advanced fibrosis, may attain the optimal accuracy (Figure 1).Figure 1.: ROC curves for non-invasive tools for detecting advanced liver fibrosis (F3-F4) in patients with MAFLD. A: AST to Platelet Ratio Index (APRI) B: NAFLD fibrosis score (NFS) C: FIB-5 D: FIB-4. Table 1. - Descriptives, optimal cut-offs, and diagnostic values of non-invasive tools for assessing advanced liver fibrosis (F) APRI FIB-4 FIB-5 NFS Mean score (Min-Max) 0.36 (0.14-0.95) 0.94 (0.14-3.27) 28.89 (20.17-38.47) -2.34 (-6.25-1.56) Optimal cut-off 0.54 1.13 21.24 0.22 Sensitivity 83.33% 83.33% 33.33% 83.33% Specificity 87.23% 81.91% 1.06% 96.81% Positive likelihood ratio 6.53 4.61 0.34 26.11 Negative likelihood ratio 0.19 0.20 62.67 0.17 Positive predictive value 29.41% 22.73% 2.11% 62.50% Negative predictive value 98.80% 98.72% 20.00% 98.91% Accuracy 87.00% 82.00% 3.00% 96.00% APRI, AST to Platelet Ratio Index; FIB-4, fibrosis index based on 4 factors; FIB-5, fibrosis index based on 5 factors; NFS, non-alcoholic fatty liver disease fibrosis score.