Objective: A critical problem in cardiovascular MRI in small rodents is adjusting the sequence acquisition to the high heart and respiratory rates. The aim of this study was to compare a retrospectively self-gated fast low angle shot navigator (RSG-FLASH) sequence with a conventional prospectively triggered (PT-FLASH) sequence for cine imaging of the ascending aorta in mice at 9.4 T.
Material and methods: Ten C57/BL6 mice were examined with a horizontal bore 9.4 Tesla MRI animal scanner using a dedicated 2 × 2 phased-array surface coil. We acquired a RSG-FLASH sequence (RSG-FLASH sequences (repetition time (TR) / echo time (TE) = 6.5/2.5 ms, flip angle (FA) = 10 degrees, field of view (FOV) = 2 × 2 cm, matrix = 384 × 384, slice thickness = 1 mm, 25 movie frames) perpendicular to the ascending aorta using the IntraGate technique. At the same position, we performed a PT-FLASH sequence (TR/TE = 6.5/2.1 ms, FA = 10 degrees, FOV = 2 × 2 cm, matrix = 384 × 384, slice thickness = 1 mm) in which the maximum number of movie frames had to be adjusted to the interval between two R-peaks (RR interval) of the electrocardiogram (ECG) with: number of frames = RR interval / TR." Cross-sectional vessel areas at end-systole (AES) and end-diastole (AED) were measured to determine the aortic strain (ΔA = (AES-AED)/AED). Two blinded readers rated the sequences for presence of flow and trigger artifacts and their influence on the depiction of the blood/vessel-wall interface. Irregularities in displaying the cardiac cycle and the overall suitability of the sequence for aortic strain evaluation were assessed using a 5-level ordinal scale. Statistical differences were analyzed using Student t test and Wilcoxon signed rank test (P < 0.05). Intra- and interobserver variability was evaluated using Bland-Altman analyses.
Results: No significant differences were noted between techniques regarding the measured vessel areas (AED: P = 0.07, AES: P = 0.34), ΔA: P = 0.1). Similarly, there were no significant differences in heart (P = 0.06) and respiratory (P = 0.24) rates. The acquisition time for RSG-FLASH sequence was significantly shorter (P = 0.04). Significantly fewer flow and trigger artifacts were noted by both readers with the RSG-FLASH sequence. Likewise, both readers considered the RSG-FLASH sequence to be superior for depiction of the blood/vessel-wall interface. The RSG-FLASH sequence was also rated superior regarding irregularities in displaying the cardiac cycle and in terms of overall suitability for evaluation of AED, AES, and aortic strain (P < 0.05 each).
Conclusion: RSG-FLASH is preferable for cine imaging of the aorta. It provides the same quantitative data as PT-FLASH cine imaging but is less prone to flow and trigger artifacts. RSG-FLASH permits more homogeneous depiction of the cardiac cycle and is faster than the PT-FLASH sequence. PT-FLASH is more prone to misregistration of the respiratory cycle or the ECG by the external monitoring device used for acquisition. This effect may be even more pronounced in animals with disease models that are less stable in terms of heart and respiration rate during anesthesia.