Predicting postoperative outcomes of liver resection by magnetic resonance elastography.

Background: Cirrhosis is associated with blood loss during liver resection and postoperative complications. The liver stiffness measurement has recently become available for assessment of liver fibrosis.
Methods: This prospective study was performed to predict postoperative outcomes of liver resection. The liver stiffness measurement was measured prospectively using magnetic resonance elastography for patients who had undergone liver resection for malignancy. We investigated whether the liver stiffness measurement by magnetic resonance elastography is correlated with liver fibrosis and postoperative outcomes.
Results: The median liver stiffness measurement by magnetic resonance elastography in 175 patients was 3.4 (range: 1.5-11.3) kPa, and the pathologic grade of liver fibrosis was significantly correlated with the liver stiffness measurement (r = 0.68, P < .001). The median blood loss during transection per unit area was 4.1 mL/cm ² (range: 0.1-37.0 mL/cm ² ), and the frequency of major complications was 16.0%. The liver stiffness measurement was the only independent prognostic factor for both blood loss (regression coefficient: 1.14, 95% confidence interval: 0.45-1.83, P = .001) and major complications (odds ratio: 2.14, 95% confidence interval: 1.63-2.93, P < .001). Receiver operating characteristic curve analysis indicated a significant correlation between the liver stiffness measurement and major complications with calculated area under the curve of 0.81 (P < .001), and the sensitivity and specificity for prediction of major complications (cutoff value: 5.3 kPa) were 64.3% and 87.8%, respectively. On the other hand, the amount of blood loss was significantly correlated with the frequency of major complications (P = .003).
Conclusion: The liver stiffness measurement by magnetic resonance elastography could be used as a predictive marker for the risk of major complications due to blood loss during liver resection.
(Copyright © 2017 Elsevier Inc. All rights reserved.)