Magnetic Resonance Imaging (MRI) is a very versatile and non-invasive technology providing unique contrast between biological tissues with very powerful diagnostics capabilities. The use of high field (>3T) scanners yields increased signal-to-noise ratio (SNR) thereby enabling higher resolution images. For that reason, many laboratories throughout the world (around 50 sites) have pushed in that direction by trying to develop the tools necessary to exploit the full potential of 7T scanners. To obtain images, nuclear spins first need to be excited by means of electromagnetic waves. As the magnetic field strength increases, the frequency of those waves likewise increases and leads by interference to zones of shades in the brain, very detrimental to medical diagnosis. To fight this problem, the community has developed the so-called parallel transmission (pTx) technology. It consists of placing around the head of the subject several RF transmitters that can be controlled independently in amplitude and phase. Again by interference and by suitable optimization algorithms, the excitation can be tailored to be homogeneous over the full brain to return optimal SNR and contrast. Nearly 15 years have passed since the advent of this technology and despite the tremendous progress and demonstrations done in the community, this technique has not been transposed to clinical routine. In spite of its potential, the main limitation indeed has been the calibration step needed for each subject, taking time (thus money) and expertise. Before one could hope to exploit the technology for a useful purpose, roughly 15 minutes were necessary and were deemed an unacceptable burden, with furthermore potential technical or human errors slowing even more the procedure. The UniPAT Proof Of Concept project came to birth after a demonstration was made that such procedure in fact could be skipped by using"universal solutions", i.e. designed to be robust versus intersubject variability. The demonstration, however, was made on a home-made coil so that the solutions could not be deployed to other sites for a broader applicability and with potential commercial outcome. The project hence consisted of carrying out the demonstration this time on available commercial coils, with key MRI sequences, and to disseminate the potential of the approach.
The project went beyond the initial plan. For bigger impact, first a database of 10 maps of electromagnetic and static fields was built over 10 healthy subjects by using the commercially available 8Tx/8Rx Rapidbiomedical coil available on site. The maps were then used offline to design non selective and selective pulses to mitigate the RF inhomogeneity problem at 7T in T1-weighted (MPRAGE) and GRE2D sequences. The results showed great mitigation of the problem with minor impact on performance compared to the solutions tailored to the subjects. Seduced by those results, Siemens encouraged to repeat the demonstration on the Nova 8Tx/32Rx coil, which is their workhorse coil. Not having the coil on site, the instigators of UniPAT collaborated with the German Center of Neurodegenerative Diseases (DZNE) in Bonn to carry out the same scans at their site. A database of 10 subjects thus again was built and the robust solutions were tested on an additional set of 10 volunteers. The results were even more successful given the higher coil performance. Later, similar "calibration-free", universal, pulses were designed for refocusing purposes, highly required for T2-weighted sequences. The latter sequences are particularly sensitive to RF field inhomogeneities and simply yield non-exploitable images unless proper measures are taken. Through the use of a sophisticated analysis, such solutions could be determined yet with great simplicity, versatility, and they were validated again in Bonn on 5 subjects. Finally, because scanner tuning variability may exist across sites, the pulses were sent to a couple of additional laboratories having the same set up (University of Minnesota and University of Maastricht) to blindly try the plug and play solutions. The universal pulses again greatly outperformed the standard approaches, demonstrating the robustness of the technique. The work above has led to 2 publications (in open access) and 7 abstracts. A powerful toolbox has been programmed to design solutions virtually covering all 3D MRI applications. Incidentally, this work has also pushed the safety studies in order to boost RF pulse performance while guaranteeing subject's safety.
The outcomes of the project have been presented at several conferences/workshops and at some coil manufacturer's site. An external sub-contractor was also hired to study the commercial value of the corresponding patent and invention. In the end, dealing with the main scanner manufacturers was advised and a few business-oriented discussions have taken place with the leading one. In the meantime, the same company however has developed its own tools to address the calibration problem mentioned above. Many efforts have been made in that regard to speed up the procedure so that universal pulses are for the moment not considered a priority. Nevertheless, no feedback from users is available yet, so that time will tell whether the calibration-free pulses developped in UniPAT have a commercial future. It remains indeed that many academic sites across the world have expressed an interest for acquiring such solutions. Downloadable products with expirable licenses are currently being elaborated to maximize dissemination and exposure.
