Background: To explore whole-liver histogram analysis (HA) with blood oxygen level-dependent functional magnetic resonance imaging (BOLD-fMRI) in evaluating and diagnosing hepatic fibrosis (HF) in a CCl4-induced rabbit model.
Methods: One hundred rabbits were classified as CCl4-induced HF groups (n=80) and control group (n=20), and were scheduled for BOLD-fMRI examination on a 3.0T scanner. Whole-liver HA metrics, including the mean, median, skewness, kurtosis, inhomogeneity, entropy, and nth percentiles, were extracted from the R2* map. Parameters were compared among the different HF stages using receivers operating characteristic analysis and Mann-Whitney U-test.
Results: In all, 17, 18, 19, 17, and 15 rabbits were pathologically diagnosed as F0, F1, F2, F3, and F4, respectively. HA parameters, including the median, inhomogeneity, entropy, and the 75th and 90th percentiles of the BOLD R2* map, demonstrated significant positive correlations with the fibrosis stage (r=0.226-0.718, P≤0.039). The 75th percentiles demonstrated higher diagnostic efficacy than the other HA parameters in fibrosis staging, with an AUC value of 0.86 for ≥ F1, 0.87 for ≥ F2, 0.87 for ≥ F3, and 0.86 for ≥ F4.
Conclusions: BOLD-fMRI HA provides increased diagnostic performance in staging HF, especially for the 75th percentiles.
Keywords: Hepatic fibrosis (HF); blood oxygen level-dependent (BOLD); histogram analysis (HA); whole liver.