Abstract
<title>Abstract</title> <p> <bold>Background</bold> Detection of fibrosis in metabolic dysfunction–associated steatotic liver disease (MASLD) is crucial for referral to hepatology care. Current guidelines recommend a two-step strategy based on FIB-4 (cut-offs 1.3 or 2.0 if > 65 years) followed by liver stiffness measurement (LSM ≥ 8 kPa). We assessed the real-world FIB-4 performance in a large Italian MASLD cohort and identified factors determining misclassification. <bold>Methods</bold> We enrolled 2638 ultrasound-proven MASLD patients from 9 Italian hepatological centers at their first evaluation. All subjects underwent simultaneous Fibroscan® and FIB-4 calculation within 6 months from referral. <bold>Results</bold> LSM ≥ 8 kPa in 21.9% of patients. Among those with FIB-4 < 1.3/2.0, 13.2% were false negatives (LSM ≥ 8 kPa), 58.7% of patients with FIB-4 ≥ 1.3/2.0 were false positives (LSM<8kPa). Diabetes, hypertension, overweight and obesity, elevated ALT, and grade 3 steatosis independently predicted false-negative results, whereas older age, diabetes, and elevated ALT reduced the likelihood of false-positive classification. FIB-4 showed 59% sensitivity, 77% specificity, 87% negative predictive value, and an AUROC of 0.72 (95% CI 0.69–0.74). Diagnostic accuracy improved with age (AUROC 0.59 in patients < 45 years vs 0.77 in ≥ 65 years), declined with increasing BMI (0.82 in lean vs 0.71 in obesity), and in severe steatosis (0.64 vs 0.76). Lower rule-out thresholds were required in younger, obese, and severe steatosis patients. <bold>Conclusions</bold> FIB-4 showed only moderate performance for detecting elevated LSM in MASLD. Patients with multiple cardiometabolic factors were frequently misclassified, supporting the need for a direct fibrosis assessment in high-risk individuals and for individualized FIB-4 thresholds. </p>