Iron Quantification Using Spectral CT-Based Material Decomposition Technique for Noninvasive Hepatic Fibrosis Staging: An Experimental Study.

Abstract

BACKGROUND: Accurate staging is crucial for effective treatment of hepatic fibrosis (HF). This study evaluates whether spectral CT-derived liver iron concentration (LIC) quantification correlates with HF stages and validates its efficacy in diagnostic staging. METHODS: One hundred seventy-five New Zealand white rabbits were included in the prospective study (35 normal controls; 140 HF models). Rabbits underwent spectral CT scans before modeling and at 4, 6, 8, 10, and 12&#x2009;weeks. HF was staged by METAVIR system. Using iodine and water as reference materials paired with iron, the original LIC and the normalized LIC (NIC&#x2009;=&#x2009;LIC/aortic-iron) were measured. The material decomposition (MD) technique of spectral CT was used to quantify LIC and evaluate the correlation between HF stage and LIC. ANOVA, chi-square, and ROC were employed using SPSS, Python, and MedCalc. RESULTS: One hundred fifty rabbits (30-per-stage, F0-F4) were selected. LICand NICin the noncontrast (NC) phase, and LICand NICin arterial phases (AP), increased with HF progression (p&#x2009;<&#x2009;0.05) consistent with pathological results. Multiple LIC parameters exhibited significant differences in 126 comparisons (p&#x2009;<&#x2009;0.05). ROC showed excellent diagnostic performance of NIC-AP across HF stages (AUC: 0.884-1.000), especially in distinguishing F1 from F0, F2, F3, and F4 (AUC: 0.979, 0.979, 0.918, 0.934). NIC-NC exhibited moderate efficacy in F0 versus F1 (AUC&#x2009;=&#x2009;0.767) but high in other stages (AUC: 0.826-0.994). A multiparameter model achieved a macro-AUC of 0.9917. CONCLUSION: Spectral CT-based liver iron concentration quantification demonstrates a strong correlation with HF stages, providing precise HF staging and excellent diagnostic performance.