OBJECTIVE: To evaluate the tradeoffs between temporal and spatial resolution and contrast dosing in dynamic contrast-enhanced magnetic resonance angiography (CE-MRA). METHODS: Bilateral carotid artery aneurysms were created in a swine model. Dynamic CE-MRA using 1 mol/L gadobutrol was performed at 3 T, with high temporal (high-temp), middle temporal (mid-temp), and low temporal (low-temp) resolutions. High temporal CE-MRA was performed twice using 1 mL and 2 mL gadobutrol (2 mL/s). Middle temporal and low-temp sequences were performed once with 2 mL gadobutrol (2 mL/s). The signal-to-noise ratio (SNR) was quantitatively assessed. Blinded reads were used to qualitatively evaluate contrast dose and image quality. RESULTS: The mean SNRs of high-temp, mid-temp, and low-temp resolutions were 56.7, 47.5, and 48.1. There was no significant difference between the 3 sequences with 2 mL gadobutrol. The mean SNR of the high-temp resolution with 2 mL was significantly higher than that with 1 mL (56.7 vs 39.9). In qualitative analysis, the 3 temporal sequences with 2 mL gadobutrol showed no significant differences regarding overall image quality and diagnostic value. High temporal resolution with 2 mL consistently showed the superiority of image quality than that with 1 mL. CONCLUSIONS: High temporal dynamic CE-MRA with 2 mL (0.04 mmol/kg body weight) gadobutrol can produce consistently superior image quality over that with 1 mL (0.02 mmol/kg body weight). For a given contrast dose, the tradeoffs between temporal and spatial resolution will not result in significant differences in image quality in TWIST (time-resolved angiography with interleaved stochastic trajectories). Copyright © 2013 by Lippincott Williams & Wilkins.