The objective of this study was to investigate clinical utility of a graphical method for estimating liver uptake and blood retention of 99mTc-DTPA-galactosyl human serum albumin (99mTc-GSA; DTPA is diethylenetriaminepentaacetic acid) using dynamic single photon emission computed tomography (SPECT) data. When considering the kinetics of 99mTc-GSA, if it is assumed that (1) 99mTc-GSA distributes only between blood and liver, and (2) no metabolism of 99mTc-GSA occurs during the observation period, a plot of liver counts versus cardiac blood pool counts should, theoretically, be a straight line. From the slope and y intercept of a regression line, coefficients for converting count based liver and blood pool data to the per cent injected dose (%ID) can be calculated. The applicability of this method was tested on dynamic SPECT data from 30 patients with liver dysfunction. To validate this method, plasma concentrations (%ID/ml plasma) at 6, 15 and 30 min after the injection were estimated by this method and compared with the measured ones. To investigate the clinical significance of the per cent liver uptake, the value obtained by this method was compared with the results of conventional liver function tests, including serum albumin, the hepaplastin test, prothrombin time and indocyanine green clearance. In every data set, a plot of liver counts to cardiac blood pool counts was fitted well by a straight line (P<0.00001). Estimated plasma concentrations by this method showed good correlation with the measured ones at 6, 15 and 30 min after the injection (r=0.748, 0.838, 0.875, respectively; P<0.0001). The liver uptake determined by this method showed good correlation with the results of conventional hepatic function tests (P<0.002). The graphical method could provide an accurate estimate of %ID of 99mTc-GSA in blood without the need for blood sampling. The liver uptake determined by this method could be a simple but useful quantitative indicator of hepatic function.