부산대학교 도서관

Introduction. The increasing prevalence of nonalcoholic fatty liver disease (NAFLD) worldwide has become a significant burden. NAFLD can progress to non-alcoholic steatohepatitis (NASH), NASH-related cirrhosis, and hepatocellular carcinoma. Scientific efforts are focused on developing non-invasive tests (NITs) to predict clinically significant portal hypertension (CSPH) and avoid invasive, costly investigations. Vibration-controlled transient elastography (VCTE) has become part of many algorithms, including the recent Baveno VII criteria. A new model (ANTICIPATE-NASH) has been proposed to identify CSPH in obese patients. Recently, other NITs, such as Agile 3+ and Agile 4, have been validated for predicting advanced fibrosis and cirrhosis, respectively, in NAFLD/NASH patients, but their performance in assessing CSPH has not been tested yet. Objective. This study aimed to evaluate the performances of Agile 3+ and Agile 4 in identifying CSPH in patients with NAFLD/NASH. Materials and Methods. The study included seventy-six consecutive patients with biopsy-proven NAFLD/NASH. Liver stiffness was measured by VCTE and fibrosis was assessed histologically using the Metavir scoring system. The performance of NITs was assessed using AUROC analysis and the DeLong protocol for comparison. Differences in classification between NITs were tested using McNemar's test. Data analysis was performed in MedCalc v20, considering p-value < 0.05 as statistically significant. Results. The liver histology fibrosis scoring identified 1 (1.3%), 10 (13.2%), 18 (23.7%), 15 (19.7%), and 32 (42.1%) patients as F0, F1, F2, F3, and F4, respectively. The performance of VCTE in identifying CSPH was excellent, with an AUC of 0.95 (95% CI: 0.86 - 0.99, p < 0.001). The ANTICIPATENASH score had slightly lower but still excellent performance, with an AUC of 0.935 (95% CI: 0.84 - 0.98, p<0.001). Agile 3+ had the best performance in identifying CSPH, with an AUC of 0.96 (95% CI: 0.89-9.99, p < 0.001), and was significantly better than FIB-4 (p = 0.04) and FIB-4+ (p = 0.02). The Baveno VII criteria for CSPH had excellent rule-out (Se = 96%, NPV = 96.3%) and rule-in (Sp = 100%, PPV = 100%) performance, with 21 (33.9%) patients left unclassified. Agile 3+ was superior to the Baveno VII criteria in identifying patients with CSPH, with 15 (24.2%) patients still in the "grey zone," and no significant difference in classification (3.17%, CI: -5.59-11.94, p = 0.22). Conclusions. Newly developed NITs, such as Agile 3+ and Agile 4, show good performance in identifying patients with CSPH at risk for decompensation in NAFLD/NASH. [ABSTRACT FROM AUTHOR]