Background: In liver transplantation, a minimum graft to patient body weight (BW) ratio is required for graft survival; in liver resection, total liver volume (TLV) calculated from body surface area (BSA) is used to determine the future liver remnant (FLR) volume needed for safe hepatic resection. These two methods of estimating liver volume have not previously been compared. The purpose of this study was to compare FLR volumes standardized to BW versus BSA and to assess their utility in predicting postoperative hepatic dysfunction after hepatic resection.
Methods: Records were reviewed of 68 consecutive noncirrhotic patients who underwent major hepatectomy after portal vein embolization between 1998 and 2006. FLR (cubic centimeter) was measured preoperatively with three-dimensional helical computed tomography; TLV (cubic centimeter) was calculated from the patients' BSA. The relationship between FLR/TLV and FLR/BW (cubic centimeter per kilogram) was examined using linear regression analysis. Receiver operating characteristic (ROC) curve analysis was used to determine FLR/TLV and FLR/BW cutoff values for predicting postoperative hepatic dysfunction (defined as peak bilirubin level>3 mg/dl or prothrombin time>18 s).
Results: Regression analysis revealed that the FLR/TLV and FLR/BW ratios were highly correlated (Pearson correlation coefficient, 0.98). The area under the ROC curve was 0.85 for FLR/TLV and 0.84 for FLR/BW (95% confidence interval, 0.71-0.97). Sixteen of the 68 patients developed postoperative hepatic dysfunction. The ROC curve analysis yielded a cutoff FLR/BW value of <or=0.4, which had a positive predictive value (PPV) of 78% and a negative predictive value (NPV) of 85%. The corresponding FLR/TLV cutoff value of <or=20% had a PPV of 80% and a NPV of 86%.
Conclusions: Based on the strong correlation between the FLR measurements standardized to BW and BSA and their similar ability to predict postoperative hepatic dysfunction, both methods are appropriate for assessing liver volume. In noncirrhotic patients, a FLR/BW ratio of <or=0.4 and FLR/TLV of <or=20% provide equivalent thresholds for performing safe hepatic resection.