​MRI is a powerful imaging modality now widely adopted in everyday clinical practice and neuroscience research. It is the first choice examination for detecting, among other disorders, a number of diseases of the central nervous system (epilepsy, multiple sclerosis, cerebrovascular diseases). However, its use at 7 Tesla and beyond (Ultra-High Field, UHF), with the aim of boosting performance, remains a real challenge because the higher the magnetic field, the greater are the artifacts. At present, UHF MRI applications are mainly in the field of research, and their transfer to clinical applications still requires major technological advances to really take full advantage of them.
 With this in mind, researchers at Baobab (NeuroSpin) have recently published three methodological papers, each focusing on the various pitfalls faced by ultra-high field MRI.

ACCELERATE ANATOMICAL IMAGING (T1) AT HIGH SPATIAL RESOLUTION

Image acquisition at high spatial resolution and even at UHF takes a long time, making it highly sensitive to involuntary movements of the subject in the imager. Effectively reducing examination times is therefore a critical issue. In the first publication, the authors drastically reduced downtime by using a Magnetization Prepared 2 Rapid Acquisition Gradient Echoes (MP2RAGE) sequence, chosen for its high T1 quantification potential, its robustness to magnetic field heterogeneities and its ability to rapidly produce 3D parametric maps. The quality of the obtained raw image deteriorates, but a synthetic reconstruction, based on the NMR signal relaxation model, makes it possible to recover perfect reading quality for a reduction in acquisition time of around 30% in healthy subjects and 15% in patients with multiple sclerosis. A substantial gain without any inconvenience!

ENHANCE THE PERFORMANCE OF FUNCTIONAL IMAGING

Functional MRI (fMRI) sequences ideally require high spatial and temporal resolution combined with complete brain coverage, and again imply very long acquisition times. Non-Cartesian methods, based on the use of SPARKLING, a compressed MRI acquisition algorithm (compressed sensing) designed and developed at NeuroSpin precisely to reduce these acquisition times, are serious candidates for reducing times without compromising image quality (more on SPARKLING). However, imperfections in the static and dynamic fields, induced essentially by the subject's physiological movements, are detrimental to such fMRI sequences, particularly at ultra-high magnetic fields, and the SPARKLING strategy remains very sensitive to heterogeneities in the static magnetic field. In this second publication, the Baobab team used a field camera to stabilize and correct the quality of this type of fMRI acquisition and take full advantage of its performance.

SIMPLY VALIDATE PROTOTYPE RF ANTENNAS

The design and development of radiofrequency (RF) antennas or coils are essential to achieve the expected benefits of UHF MRI. Testing a prototype RF coil on a subject involves extensive checks to guarantee its safety (preliminary electromagnetic simulations and validation on phantoms to determine the specific absorption rate). However, simulation, and even experimentation on phantoms, do not allow to be totally sure that a prototype will perform well in vivo, due to the lack of sufficiently complete models. In the third publication, the authors studied the possibility of testing RF antenna prototypes in vivo, during the development process, without constraining validation and above all without danger to the volunteer. They applied imaging sequences using very low RF power, making it impossible to exceed safety standards under any circumstances. This method allows greater versatility in the antenna development process, making it so much more comfortable for the team's researchers!

Taken together, these results represent significant advances in obtaining particularly detailed anatomical and functional images of the brain at unprecedented resolutions.

Contact : Alexandre Vignaud (alexandre.vignaud@cea.fr)