Functional remnant liver volumetry using Gd-EOB-DTPA-enhanced magnetic resonance imaging (MRI) predicts post-hepatectomy liver failure in resection of more than one segment

Background: Gd-EOB-DTPA-enhanced magnetic resonance imaging (EOB-MRI) can be used for evaluating liver functional reserve. We assessed whether functional remnant liver volumetry (FRLV) using EOB-MRI predicted post-hepatectomy liver failure (PHLF) in resection of more than one segment. Methods: We retrospectively analyzed 155 cases of hepatectomy of more than one segment. For assessment of FRLV, signal intensity (SI) of remnant liver was measured in T1-weighted images. Functional remnant liver score was derived by division of SI of liver by SI of muscle (or spleen), resulting in liver-to-muscle ratio (LMR) and liver-to-spleen ratio (LSR). FRLV were calculated by multiplying LMR (or LSR) and remnant liver volume. We investigated preoperative factors predicting PHLF (>= grade B) in study cohort (all cases except for portal vein embolization [PVE], n = 129) and validation cohort (PVE cases, n = 26). Results: In study cohort, PHLF occurred in 5 patients (3.9%). In multivariate analysis, FRLV (LMR) was the most reliable predictor of PHLF (P = 0.013). The cutoff value of FRLV (LMR) predicting PHLF was 615 mL/m(2) (AUC: 0.939). In validation cohort (n = 26), the cutoff value of FRLV (LMR) indicated reliable results, sensitivity (100%), specificity (77.3%), and accuracy (80.8%). Conclusions: FRLV using LMR could precisely predict PHLF of more than one segment, and was useful even in patients who underwent PVE.