In the current study we investigate the accuracy and precision of hepatic perfusion measurements based on the Johnson and Wilson model with the adiabatic approximation. VX2 carcinoma cells were implanted into the livers of New Zealand white rabbits. Simultaneous dynamic contrast-enhanced computed tomography (DCE-CT) and radiolabeled microsphere studies were performed under steady-state normo-, hyper- and hypo-capnia. The hepatic arterial blood flows (H(A)BF) obtained using both techniques were compared with ANOVA. The precision was assessed by the coefficient of variation (CV). Under normo-capnia the microsphere H(A)BF were 51.9 +/- 4.2, 40.7 +/- 4.9 and 99.7 +/- 6.0 ml min(-1) (100 g)(-1) while DCE-CT H(A)BF were 50.0 +/- 5.7, 37.1 +/- 4.5 and 99.8 +/- 6.8 ml min(-1) (100 g)(-1) in normal tissue, tumor core and rim, respectively. There were no significant differences between H(A)BF measurements obtained with both techniques (P > 0.05). Furthermore, a strong correlation was observed between H(A)BF values from both techniques: slope of 0.92 +/- 0.05, intercept of 4.62 +/- 2.69 ml min(-1) (100 g)(-1) and R(2) = 0.81 +/- 0.05 (P < 0.05). The Bland-Altman plot comparing DCE-CT and microsphere H(A)BF measurements gives a mean difference of -0.13 ml min(-1) (100 g)(-1), which is not significantly different from zero. DCE-CT H(A)BF is precise, with CV of 5.7, 24.9 and 1.4% in the normal tissue, tumor core and rim, respectively. Non-invasive measurement of H(A)BF with DCE-CT is accurate and precise. DCE-CT can be an important extension of CT to assess hepatic function besides morphology in liver diseases.