OBJECTIVE. Posthepatectomy liver failure (PHLF) is one of the most serious complications after hepatectomy. The objective of the present study is to assess the potential diagnostic ability of 99mTc-galactosyl human serum albumin (GSA) scintigraphy to predict PHLF as defined by the International Study Group of Liver Surgery (ISGLS). MATERIALS AND METHODS. Data from 100 patients who underwent 99mTc-GSA scintigraphy and subsequent hepatectomy were retrospectively analyzed. The blood clearance ratio (HH15), hepatic uptake ratio (LHL15), and maximal removal rate (Rmax) of 99mTc-GSA (GSA-Rmax) were calculated as scintigraphic parameters for the total liver. In addition to the ratio of preoperatively estimated remnant liver (ERL) counts to total liver counts (rERL-GSA), the ratio of actual remnant liver (ARL) counts to total liver counts (rARL-GSA), determined by applying a more accurate resection line with reference to both pre- and postoperative CT, was obtained from SPECT images. Functional remnant liver parameters of ERL-LHL15 (LHL15 of the estimated remnant liver), ERL-Rmax (maximal removal rate of estimated remnant liver counts), ARL-LHL15 (LHL15 of the actual remnant liver), and ARL-Rmax (maximal removal rate of actual remnant liver counts) were calculated using these values. ROC analysis was performed to evaluate the ability of these parameters to predict PHLF. Multivariate analysis was performed to identify independent predictors of PHLF. RESULTS. PHLF occurred in 33 patients. Each of the ARL parameters had a significantly higher diagnostic performance compared with the corresponding ERL parameter (AUC values: for rARL-GSA vs rERL-GSA, 0.77 vs 0.62 [p = 0.0004]
for ARL-LHL15 vs ERL-LHL15, 0.79 vs 0.64 [p = 0.0005]
and for ARL-Rmax vs ERL-Rmax, 0.78 vs 0.66 [p = 0.0003]). According to multivariate analysis, each of three ARL parameters was identified as an independent predictor of PHLF (p &lt
0.0001 for all). CONCLUSION. Technetium-99m�labeled GSA scintigraphy is useful for predicting PHLF, particularly for applying an accurate resection line on GSA-SPECT images.