Delayed hepatobiliary phase imaging improves bile leak detection on gadoxetate-enhanced MRI
In brief
Delayed MRI raises bile leak detection sensitivity from 27% to 82%
In this retrospective study of 36 MRI examinations, delayed imaging at a median of 65 minutes detected 82% of bile leaks, compared with 27% at conventional timing, while specificity remained 100%. The findings support adding delayed images when standard-timing results are negative or unclear, but the small, single-center study leaves performance in broader practice uncertain.
- Journal
- European radiology (Q1)
- Published
- 15 November 2026
- Study design
- Unclassified
- Evidence level
- Level 5, Expert Opinion (CEBM 5)
- Authors
- Andrea Vanzulli, Jules Grégory, Riccardo Sartoris, Juan Li, Francesco Matteini, Valérie Vilgrain, et al.
- PMID
- 43034939
- DOI
- 10.1007/s00330-026-13017-1
Why clinicians should know about it
- Picked for Radiology, Radiation Oncology, Nuclear Medicine, Medical Physics and Imaging (paper of the day, 17 November 2026): Direct comparison of MRI timing for bile leak detection
Abstract
OBJECTIVES: To compare conventional and delayed (≥ 30 min) hepatobiliary-phase (HBP) gadoxetate-enhanced MRI for the detection of bile leaks. MATERIALS AND METHODS: This retrospective single-center study included all consecutive gadoxetate-enhanced MRI examinations comprising delayed HBP acquisitions performed for suspected biliary leak over 2 years at a tertiary hepatobiliary referral center. Two abdominal radiologists blinded to clinical records independently assessed leak presence at each timepoint. A composite reference standard incorporating endoscopic retrograde cholangiopancreatography, surgical exploration, percutaneous drainage findings, and multidisciplinary assessment was utilized to assign ground-truth leak status. Sensitivity, specificity, positive and negative predictive values (PPV, NPV), accuracy, and inter-reader agreement were calculated. RESULTS: Thirty-five patients (22 men; median age, 56 years; interquartile range (IQR), 48-65) undergoing 36 examinations were included. The most common indications were liver transplantation (n = 13/35, 37%) and cholecystectomy (n = 11/35, 31%). Conventional and delayed HBP were acquired at a median of 21 and 65 min after injection, respectively. Twenty-two of 36 examinations (61%) were ground-truth positive. Inter-reader agreement was perfect for conventional HBP (κ = 1.00) and excellent for delayed HBP (κ = 0.94; 95% CI 0.84-1.00). Conventional HBP demonstrated 27% sensitivity (95% CI 11-50%), 100% specificity (95% CI 77-100%), and 56% accuracy (95% CI 38-72%). Delayed HBP demonstrated 82% sensitivity (95% CI 60-95%), 100% specificity (95% CI 77-100%), and 89% accuracy (95% CI 74-97%). NPV increased from 47% to 78%. Delayed HBP yielded higher sensitivity and overall accuracy than conventional HBP (p < 0.001, exact McNemar tests). CONCLUSION: Delayed HBP improved bile leak detection without loss of specificity. KEY POINTS: Question Does extending gadoxetate hepatobiliary-phase MRI acquisitions beyond the conventional 20-min window improve the detection of active bile leaks? Findings Delayed hepatobiliary-phase imaging raised sensitivity for bile leak detection from 27% to 82% without loss of specificity, highlighting acquisition timing as a key diagnostic determinant. Clinical relevance Delayed gadoxetate hepatobiliary-phase imaging is an effective adjunct for detecting active bile leaks and should be considered when findings at conventional timing are negative or equivocal, with direct implications for clinical management.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.