Purpose: A meta-analysis was performed to assess the diagnostic performance of diffusion-weighted imaging (DWI) in liver fibrosis (LF) staging.
Methods: We conducted a comprehensive literature search to identify relevant articles. Diagnostic data were extracted for each METAVIR fibrosis stage (F0-F4). A bivariate binomial model was used to combine sensitivities and specificities. Summary receiver operating characteristics (SROC) curves were performed and areas under SROC curve (AUC) were calculated to indicate diagnostic accuracies. Subgroup analyses were performed between different study characteristics.
Results: Twelve studies met the inclusion criteria for LF ≥F1, 16 for ≥F2, 18 for ≥F3, and 12 for F4. AUCs of DWI were 0.8554, 0.8770, 0.8836, and 0.8596 for ≥F1, ≥F2, ≥F3, and F4, respectively. Subgroup analyses showed that for LF ≥F2 and ≥F3, maximal b values (b max) ≥ 800 s/mm2 performed significantly better than b max < 800 s/mm2. The diagnostic accuracies of 3.0 T and intravoxel incoherent motion (IVIM)-DWI were significantly higher than those of 1.5 T and conventional DWI for diagnosing liver cirrhosis (F4).
Conclusions: DWI is a reliable noninvasive technique with good diagnostic accuracy for LF staging. Using b max ≥ 800 s/mm2, high-field strength (3.0 T) and IVIM-DWI can optimize the diagnostic performance of DWI.
Keywords: Diffusion magnetic resonance imaging; Liver cirrhosis; Meta-analysis; Sensitivity; Specificity.