Recently, a vascular-space-occupancy (VASO) MRI technique was developed for quantitative assessment of cerebral blood volume (CBV). This method uses the T(1)-shortening effect of gadolinium diethylenetriamine pentaacetic acid (Gd-DTPA) with imaging parameters chosen that null the precontrast blood magnetization but allow the postcontrast blood magnetization to recover to equilibrium. A key advantage of VASO CBV estimation is that it provides a straightforward procedure for converting MR signals to absolute physiologic values. However, as with other T(1)-based steady-state approaches, several important factors need to be considered that influence the accuracy of CBV values obtained with VASO MRI. Here, the transverse relaxation (T(2)/T(2) (*)) effect in VASO MRI was investigated using multiecho spin-echo and gradient-echo experiments, resulting in underestimation of CBV by 14.9% +/- 1.1% and 16.0% +/- 2.5% for spin echo (TE = 10 ms) and gradient echo (TE = 6 ms), respectively. In addition, the influence of contrast agent clearance was studied by acquiring multiple postcontrast VASO images at 2.2-min intervals, which showed that the concentration of Gd-DTPA in the first 14 min (single dose) was sufficient for the blood magnetization to fully recover to equilibrium. Finally, the effect of vascular Gd-DTPA leakage was assessed for scalp tissue, and signal extrapolation as a function of postinjection time was demonstrated to be useful in minimizing the associated errors. Specific recommendations for VASO MRI acquisition and processing strategies are provided.