PhD defence of William Mathieu – A Size-Adaptable Flexible 128-Channel Receive-Only RF Coil Array for Brain 3-T MRI
Abstract
The goal of this work has been to make head imaging more equitable, and show that accommodative engineering designs benefit all. In magnetic resonance imaging (MRI), most radio-frequency (RF) coils are designed such that the largest heads fit, however this means that smaller heads are undeserved in terms of sensitivity and ultimately signal-to-noise ratio, as the space between the surface of the head and of some of the coil’s elements is larger. Adding adjustability to coils is a popular approach, however stability and signal-to-noise ratio often suffer, making it self defeating.
A 128-channel RF coil array for head MRI was designed. The array is split up into sub-arrays that are built into semi-flexible plates, which were inspired by the cranial bones of the human head. The plates can be adjusted to fit different heads. The gaps between the plates are bridged by ultra-flexible bowing suture elements which bend outwards when the plates move inwards to accommodate a smaller head, and flatten out for larger heads as the plates move apart.
An average head surface model was segmented into eight 2-mm thick plates 3D-printed in flexible thermoplastic polyurethane (TPU) on which 35-mm diameter elements are threaded. A total of 104 channels are plate bound with the remaining 24 reserved as bowing sutures.
This thesis presents the results of targeted tests that demonstrate the feasibility and effectiveness of the 128-channel design, and the fully realized semi-flexible plate-bound 104-channel coil.
In MRI experiments on phantoms the 104-channel coil outperformed a commercial 64-channel head coil in terms of an average SNR by 1.45-fold and at cortical depths as much as 4.33-fold. It also outperforms in terms of accelerated imaging, specifically at higher R factors. For instance, the maximum g-factor of an R = 8 A-P acceleration was reduced by ≈40%, and in an R = 4 × 4 acceleration, there was a reduction of ≈58%.
The adjustable plate approach presented in this work shows that an equitable coil is possible and has the potential to improve head imaging for all, in terms of imaging speed, patient comfort, and image quality.