Abstracts (Continue in Part I)

Abstract

MOTIVATIONThe large purchase and operating expenses of typical, superconducting-based MR imagers limits availability to all but the most wealthy of nations. The large supply of whole-body machines also impedes development of low-cost dedicated systems such as veterinary imagers.We propose a low-cost, permanent magnetbased imager whose design could be readily specialized and be affordable to many "ThirdWorld" countries. INTRODUCTIONOnly a small minority of MR imagers use permanent magnets to create the main magnetic field (B). There are two principal reasons for this: Field strength is limited to 0.3 Tesla (TI;and field homogeneity is difficult to achieve. Our design increases field uniformity and turns the low field strength into a distinct advantage.While debate continues on the subject of optimal field strength, low-field imagers are generally regarded to be a second choice. While signal to noise ratios ( S M ) increase a t least linearly with field strength, there is growing evidence, however, that contrast to noise ratios (CRV) for some applications are better a t low fields (1). DISCUSSIONWe are approaching the design completion of a permanent magnet-based MR imager. The virtual elimination of maintenance for both the magnet and the signal acquisition electronics is the key to our low-cost system.The heart of the design is a C-shaped flux return path and highly machined flat and parallel pole faces. This will be the most 'open' imager available, a primary consideration for interventional MRI. The device can be oriented vertically or horizontally, and allows access to a more comfortable patient from three sides (Figure 1).Accurate imaging requires homogeneous Bfields, and we believe our group has made at least three design advancements in this area. ADVANCEMENTSA large number of carefully placed Neodymium-Iron-Boron (NdFeB) magnet ingots will create a highly uniform field between two flat pole faces, where the object to be imaged is placed. Technology developed at the now defunct Superconducting Super Collider (SSC) in making precise correcting (shim) magnets will be utilized to fabricate our pole faces to a tolerance of one part in ten-thousand. A new method of twisting the iron flux return will effectively invert and then eliminate most of the field inhomogeneities, caused by a lack of symmetry, present in all other marketed imagers of similar design.There is one other advantage present in our design. Due to the low field (0.25T1, and therefore the low resonance frequency of the proton spectra (10 MegaHertz), all signal acquisition and processing can be changed from the conventional analog to a digital architecture. State-of-the-art 12-bit sampling rates are now at 25MHz, or nearly 0.3T equivalently, when the Nyquist limit is considered (2). This dramatic improvement will increase the S/N, while lowering costs. The imager will be virtually free of maintenance and tedious calibration requirements. Ours will be one of the first MR imagers with a digital acquisition system in existence. Specialized radiofrequ...