Purpose or Case Report: Free-breathing MRI scans are attractive in pediatric imaging as they reduce the need for sedation and breath-holds. In this work, we evaluate a 3D T1w radial "stack of stars" gradient echo (GRE) acquisition (RAVE-RAdial Volumetric Encoding) in post-contrast abdomen and spine protocols and compare results with conventional Cartesian MRI of similar spatial resolution and volume coverage.
Methods & Materials: Studies were performed on a 3T Siemens Prisma. With radial MRI, data are acquired along k-space spokes during repetitions of the sequence. Consecutive spokes are rotated in-plane by the golden-angle (111.25 deg) to maximize k-space coverage. The center of k-space is oversampled and this feature affords RAVE's robustness to motion. When coupled with compressed sensing and parallel imaging, the radial data further yields dynamic images, (i.e., during contrast passage). We have evaluated free-breathing RAVE in 20 patients referred for non-sedated abdominal and spine MRI exams with contrast. In the abdomen, we have acquired RAVE in both axial and coronal orientations (i.e., liver, enterography). In the spine, we have acquired RAVE in the axial plane. Three radiologists compared RAVE to conventional breath-hold GRE in abdomen scans and free-breathing fast-spin-echo (FSE) images in the spine. A 3-point scale was used to evaluate diagnostic image quality: -1=RAVE is superior, 0=equivalent, +1=RAVE is inferior.
Results: In all cases, RAVE images were considered robust to respiratory, cardiac, and gastrointestinal motion, and no artifacts that impacted diagnostic image quality were noted. In the spine, RAVE was consistently superior to FSE (p<0.01) in delineating nerve tracts and roots against CSF. In the abdomen, RAVE was comparable to breath-hold GRE scans. Fig 1 shows a series of cervical, thoracic, and lumbar spine comparisons between FSE and RAVE in a 14y patient. Note the lack of motion-related artifacts in RAVE, without the use saturation bands. Fig 2 shows coronal RAVE dynamic frames ~15s apart, highlighting contrast uptake, in a 10y patient. A breath-hold post-contrast GRE
image is shown for comparison. Figure 3 shows pre- and post-contrast axial RAVE images in a 14y patient.
Conclusions: Our data demonstrates the potential utility of a free-breathing accelerated 3D T1w RAVE sequence in unsedated pediatric imaging. The technique is particularly useful in patients who are unable to follow breath-hold instructions and suspend respiration, and it is 30-50% faster in scan time than conventional methods.