In considering what an ideal MRI system might be, it is helpful to evaluate the user needs. No system is ideal for everything and most MRI system designs represent a balance of both requirements and constraints. This paper analyses the opportunities and the limitations offered by some selected recent technical trends in MRI systems and examines a strategy to reach the definition of what might be an optimal MRI system design.
In an ideal process, the design of a MRI system should start from the user needs, and the technical design should follow. The primary user needs have to be weighted not only with what is technically feasible but also with other aspects (e.g. initial cost, cost of maintenance, reliability, etc.). Undoubtedly, speed and image quality parameters such as SNR, CNR, resolution and spatial accuracy are important criteria in defining the performance of an MRI system. While speed and image quality parameters can be traded off against each other, safety has to be guaranteed at all times and without compromise. These are the initial conditions and the boundary conditions that should guide us in the development of a potentially ideal MRI system. For instance, speed in MRI can be achieved using strong gradients, parallel imaging or a combination of both [2],[3],[4]. Parallel image techniques rely on spatial information provided by phased array coils. Parallel imaging has become a key feature in modern MRI systems because it enables to replace gradient encoding with spatial localization based on a priori knowledge of the multi-element coil sensitivity. During the last two decades, coils with an increasing number of elements have been developed. Such high density element arrays can be very advantageous in some applications for SNR and highly accelerated imaging but there is a limit to the useful number of elements which is dependent on the particular clinical need. It has been shown that parallel imaging performance in a phased array coil tends to increase with increasing B0 beyond a certain threshold [5]. Though high B0 fields may seem attractive for reaching high SNR and improved parallel imaging performance, the dielectric penetration effects tend to compromise the intensity and contrast uniformity of the images and the “equal fidelity” SNR grows more slowly than the predicted linear dependency with B0 [1]. The meaning of “equal fidelity” SNR (efSNR) in this context is the SNR normalized by the pixel bandwidth as a function of B0 in order to compare images with equivalent spatial fidelity. Examination of the relationship of pixel bandwidth to SNR shows that efSNR grows according B01/2 as compared to the commonly assumed linear relationship. This is an important consideration when comparing SNR of different systems operating at different static field strengths. Some clinical applications (e.g. brain imaging) require speed, SNR and high resolution, therefore strong gradients are mandated. However, there is a physiological limit to the gradient performance and it is the limit of peripheral nerve stimulation [6]. Irrespective of what the gradient hardware is capable of (in terms of gradient strength and slew rate) the usable performance is constrained by the limits on peripheral nerve stimulation. This is just one example of the constraints to consider. The ideal MRI system is therefore the result of an optimal choice among clinical needs, safety constrains and economic viability. Hence, it is unlikely that there is a single MRI system design that is ideal for all clinical needs but rather an optimal MRI scanner for each application. Understanding the true user needs is a key point in developing the appropriate trade-off. As a final remark, an ideal MRI system would provide the necessary information for a medical professional to make a diagnosis. Ideally, it would provide only the information required and in a manner that is quantitative, definitive and robust against human errors [7].
[1] Harvey PR, “Stronger magnets, higher gradients…is that all?, ESMRMB, Vienna, 2016.
[2] Pruessmann KP, Weiger M, Scheidegger MB, Boesiger P., SENSE: sensitivity encoding for fast MRI. Magn Reson Med 42 (p952), 1999.
[3] Sodickson DK, Manning WJ, Simultaneous acquisition of spatial harmonics (SMASH): fast imaging with radiofrequency coil arrays. Magn Reson Med. 38 (p.591), 1997.
[4] Griswold M, Jakob P, Heidemann R, Nittka M, Jellus V, Wang J, KieferB, Haase A, GeneRalized autocalibrating partially parallel acquisitions(GRAPPA). Magn Reson Med 47 (p.1202), 2002.
[5] Wiesinger F, Boesiger P, Pruessmann KP, Electrodynamics and ultimate SNR in parallel MR imaging, Magn Reson Med 52 (p.953), 2004.
[6] Chronik BA & Rutt K, Simple Linear Formulation for Magnetostimulation Specific to MRI Gradient Coils, Magn Reson Med 45 (p.916), 2001.
[7] Collier DC, Burnett SSC, Amin M, Bilton S,Brooks C, Ryan A, Roniger D, Trann D, Starkschall G, Assessment of consistency in contouring of normal-tissue anatomic structures, J Applied Clinical Medical Physics 4 (p.17), 2003.