Dual-phase T1-mapping (DHR-MAP) for non-invasive assessment of advanced liver fibrosis and moderate hepatitis: a single-center validation study

BACKGROUND & AIMS: Conventional non-invasive tests (NITs) for chronic liver disease (CLD) conflate fibrosis and inflammation, limiting diagnostic accuracy. We aimed to redefine the pathophysiological basis of key biomarkers and translate these insights into an integrated diagnostic strategy. METHODS: In a critically underpowered retrospective cohort of 99 patients with histologically confirmed CLD, we performed a comprehensive evaluation of FIB-4, MRE, and DHR-MAP. Analytical rigor was ensured through post-hoc power assessment and stringent multiple comparisons correction. Statistical approaches included partial correlation, hierarchical regression, and mediation analysis to establish biomarker specificity. Diagnostic performance and the proposed two-step strategy were validated via 5-fold cross-validation. RESULTS: Our analysis resolved a fundamental biomarker paradox. FIB-4 was identified as a marker predominantly reflective of inflammatory activity, showing no independent association with fibrosis after controlling for hepatitis (p=0.414). Conversely, MRE and DHR-MAP were established as fibrosis-specific indicators. This mechanistic clarity informed a novel two-step diagnostic framework: For advanced fibrosis (≥F3), initial screening with DHR-MAP+FIB-4 achieved high sensitivity (0.778 in cross-validation), followed by confirmatory MRE providing high specificity (1.000). For moderate hepatitis (≥G2), FIB-4 screening delivered perfect sensitivity (1.000), with subsequent DHR-MAP+FIB-4 confirmation offering balanced accuracy (specificity 0.857). This integrated strategy substantially reduced diagnostic errors compared to single-modality approaches (88.5% net error reduction for fibrosis; 30.4% for hepatitis) in this cohort, although performance metrics require validation in larger, prospectively designed studies. CONCLUSIONS: This study provides a pathophysiological redefinition of key CLD biomarkers, proposing FIB-4 as an inflammation-predominant marker and introducing a novel two-step diagnostic framework that strategically sequences biomarkers based on their specific strengths. The findings-particularly the dissociation of FIB-4 from fibrosis-offer a mechanistic hypothesis and a practical diagnostic template that must be validated in larger, appropriately powered prospective cohorts to assess generalizability and clinical impact.