Gradient-Free Quantitative MRI using a Combination of B1-Selective Excitation and Fingerprinting

Project Summary Magnetic Resonance Imaging (MRI) is one of the most important medical imaging modalities because of its ability to detect and characterize lesions throughout the body. However, access to MRI is severely limited by its expensive hardware, complex siting requirements and typically-qualitative images, which require highly skilled radiologists to interpret. This project proposes a fundamentally new way to encode MRI that could enable sub- stantially cheaper and more flexible quantitative MRI scanners. Today the overwhelming majority of MRI scans are encoded using two primary methods: B0 gradients and parallel imaging using an array of receiver coils. B0 gradients take up a significant fraction of the bore diam- eter; are loud and induce peripheral nerve stimulation, compromising patient comfort; they have relatively long switching times due to the high inductance of the coils; they require bulky cooling systems and customized am- plifiers; they are expensive, representing 25-30% of the cost of a clinical scanner; and they must be carefully designed and customized to a scanner's B0 magnet. B0 gradient encoding also suffers from spatial errors due to concomitant terms, which increase with decreasing B0 field strength and will limit the performance of emerging portable and low-cost MRI systems. Parallel imaging enables scan acceleration by differentiating signals across large spatial distances, but cannot encode complete images on its own. While some have proposed a third class of encoding methods using radiofrequency transmit (B1+) gradients, none of the methods described to date have been translated into clinical use because of practical limits on their performance, stringent hardware requirements and lack of flexibility in image contrast. This project will develop and validate a fourth, fundamentally new way to encode MRI based on parallel transmission using B1+-selective pulses produced by wireless RF coil units with on-coil amplifiers that perform RF transmission and reception, combined with an acquisition and reconstruction process inspired by MR Finger- printing (MRF). This new method, Selective Encoding through Nutation and Fingerprinting (SENF), completely eliminates the need for B0 gradients and is compatible with a wide range of magnet designs and flexible ac- quisition strategies. Unlike previous B1+ imaging methods, SENF places no strict spatial variation requirements on the RF gradient fields, which enables flexible system design, and the same coils can be used for spatial en- coding and signal reception. Furthermore, instead of suffering from errors due to complex spin dynamics during RF encoding, SENF leverages those dynamics to its advantage to differentiate quantitative tissue parameters. Successful completion of this project will enable a new generation of cheaper, more accessible, more modular, and lower-maintenance MRI scanners with quantitative outputs that can be more directly related to disease and tissue states.