Software Platform for MRI Phased Array System Design & Optimization
With the widespread adoption of multi-channel MRI systems using parallel imaging, the design and optimization of receive coil arrays has become imperative, while also being complex. Nevertheless, coil optimization techniques and performance verification are essential steps when designing coil arrays. To address this issue, a novel software platform combines SEMCAD X, the latest generation of 3-D Finite Difference Time-Domain (FDTD) & Finite Integration Technique (FIT) full-wave simulation software, and Musaik™ RF Array Designer software.
The result is the first software of its kind combining all the tools needed to design RF phased arrays in a single user-friendly program. The joint effort from Schmid & Partner Engineering AG (SPEAG), Switzerland, and the National Research Council of Canada (NRC) provides coil designers and radiologists with an essential set of advanced tools needed to simulate, predict and verify array performance, enabling radiologists to perform parallel MRI protocol optimization and quality assurance.
With the urgent need for faster, better, less expensive healthcare technologies, it is anticipated that experts in the global MRI community will recognize the value in the new software platform. Consequently, the combination of Musaik and SEMCAD X lays the groundwork for future breakthroughs in MRI system design.
Figure 1 Simulation of the MRI RF fields in anatomical models in SEMCAD X and analysis of the receive coil performance in Musaik.
Array Coil Design and Performance
Musaik enables the user to import 3-D field simulation results from SEMCAD X (as illustrated in Figure 1), build various virtual array coil configurations, investigate channel compression, and then assess signal-to-noise (SNR) and parallel MRI performance. Furthermore, it allows the user to import experimental data obtained on any MRI scanner, and perform the same SNR and parallel MRI evaluation for accurate performance verification.
Figure 2 Musaik screenshot with (top) an unaccelerated reference image SNR reconstruction and (bottom) g-factor map for a 12-fold (R=4 × 3) 2D acceleration.
Many different view options are provided, including phase analysis, transmit/receive coil functionality and thresholding, which is a desired region of interest (ROI) for region specific performance analysis or small field-of-view (FOV) computations. Two dimensional g-factor maps can be computed, exploring parallel MRI performance while varying the field-of-view and acceleration factor (see Figure 2). Also, complex noise covariance matrices are computed, minimizing the technical knowledge required by users and enabling the assessment of coil isolation during construction troubleshooting.
Hospitals and diagnostic clinics need to be able to analyze and maximize the performance of their MRI systems as easy and painlessly as possible. Parallel MRI has become standard practice for better temporal or spatial resolution and shorter overall exam times. The challenge for hospitals is to get the most out of their MRI scanners; however, this is very difficult because there are many application specific parameters to optimize, such as acceleration factor, field-of-view, etc.
It is the goal of Musaik to assist in understanding the intricacies of parallel MRI. To a radiologist, the concept of MRI protocol optimization would require countless hours of MRI time, altering scan parameters and observing their effects on the resultant image, in a trial and error fashion. This process is tedious and therefore not common practice.
With Musaik, ‘fine tuning’ of MRI protocols can be done offline and in a mathematical way without using valuable scan time, to show how imaging parameters such as field-of-view and acceleration factor can be interplayed to decrease imaging time or increase spatial resolution without significant SNR loss or increased artifacts in a specified region of interest.
With regards to coil design, analysis and optimization, Musaik facilitates the investigation of array coupling through noise correlation matrix calculation and offers 1D and 2D g-factor analysis for evaluating both 2D and 3D MRI acceleration capabilities. It utilizes a SNR/g-factor display to help gauge image quality limitations and is easy to add, combine or remove array channels to assess their contributions.
To sum up Musaik’s image processing and optimization capabilities, the tool utilizes image ratios to assess regional SNR/g-factor gains and incorporates region-of-interest zooming with statistical analysis. The software features a color, grayscale and contour display, and can export in image format for presentations as well as having the capacity to export complex composite datasets for further post-processing.
In conclusion, the new platform offers a unique solution for the design of multi-channel phased array coils, parallel MRI protocol optimization and quality assurance. It is envisaged that the combination of Musaik and SEMCAD X will be key in assessing the limits of new or existing array coils.