The purpose of this study was to demonstrate the feasibility of steady-state True fast imaging with steady precession (True-FISP) four-dimensional imaging of mouse heart at high resolution and its efficiency for cardiac volumetry. Three-dimensional cine-imaging of control and hypoxic mice was carried out at 4.7 T without magnetization preparation or ECG-triggering. The k-space lines were acquired with the TrueFISP sequence (pulse repetition time/echo time ؍ 4/2 ms) in a repeated sequential manner. Retrospective reordering of raw data allowed the reconstruction of 10 three-dimensional images per cardiac cycle. The acquisition scheme used an alternating radiofrequency phase and sum-of-square reconstruction method. Black-blood three-dimensional images at around 200 m resolution were produced without banding artifact throughout the cardiac cycle. High contrast to noise made it possible to estimate cavity volumes during diastole and systole. Right and left ventricular stroke volume was significantly higher in hypoxic mice vs controls (20.2 ؎ 2 vs 15.1 ؎ 2; P < 0.05, 24.9 ؎ 2 vs 20.4 ؎ 2; P < 0.05, respectively). Owing to the many possibilities offered by transgenic and surgical approaches, numerous models of cardiovascular diseases have been developed in mouse. This provides not only a better understanding of physiology but is also of great use in preclinical trials, e.g., gene therapy or pharmacological treatment.MRI has become a reference noninvasive tool to characterize variations in structure and function of mouse heart. Until now, most of the MR studies on mouse models were performed with two-dimensional white-blood cine-imaging combined with cardiac triggering for MR acquisition synchronization (1-4). Parameters like end-systolic and end-diastolic volumes can be measured in order to evaluate both stroke volume and ejection fraction. Nevertheless, routine MRI techniques suffer from many drawbacks. First of all, two-dimensional imaging results in a limited spatial resolution in the slice dimension. In fact, typical out-ofplane resolutions are in the range of 1 mm, whereas mouse heart does not exceed 10 mm in the long axis. This could result in measurement errors due to a partial volume effect.To overcome this issue, some authors have shown interest in performing four-dimensional (4D) (three-dimensional [3D]-cine) imaging (5-7). Feintuch et al. (6) demonstrated the feasibility of 4D mouse imaging with an isotropic spatial resolution of 200 m. Nevertheless, scan times were very long (between 80 and 120 min) for bright blood imaging. More recently, Bucholz et al. (7) presented impressive images with exceptional resolution (80 m), but injection of a nonconventional liposomal gadolinium contrast agent was required. The other drawback of such experiments is that, in many cases, images are acquired with a bright blood contrast. It is now widely accepted that MR sequences with suppression of the blood signal (black or dark blood sequence) are more appropriate to reduce interobserver variability in the drawing of epicar...