While the MRI industry pursues superconducting technology in both short-bore and open magnets, a team of researchers at Stanford University is exploring the use of resistive magnets as an inexpensive alternative. In the process, a sacred cow of this
While the MRI industry pursues superconducting technology in both short-bore and open magnets, a team of researchers at Stanford University is exploring the use of resistive magnets as an inexpensive alternative. In the process, a sacred cow of this imaging modality-homogeneity of the magnetic field-is being put to rest.
The Stanford team has constructed a working prototype consisting of two resistive magnets, which are turned on and off in sequence. The stronger magnet creates a massively inhomogeneous 0.4-tesla field to provide magnetization. It is then turned off in deference to the second magnet, which is also made using resistive technology but composed of six coils so as to create a homogeneous, though much weaker, field of just 23 mtesla. This field supports the acquisition of imaging data.
“It’s a little too early to tell if it will have a real clinical impact, but that is definitely the motivation for some of the work,” said Steven Conolly, senior research associate and leader of the Stanford engineering team working on the prototype.
Funding from several sources, including the Department of Defense and the National Institutes of Health, calls for the development of a clinical scanner that can image the human head, knee, or breast. But the technology has a long way to go. The bore of the current system is hardly big enough to fit a human hand, although the next generation, now under construction, will be bigger. But the images taken so far leave a lot to be desired.
The problem, according to Conolly, is an underpowered center frequency of 1.1 MHz. This low frequency is vulnerable to external noise, which crushes the signal-to-noise ratio and, consequently, sends image quality plummeting. Conolly believes the 4 MHz center frequency planned for the next iteration of the prototype scanner should solve the problem. The goal is to reach “body dominance,” so that any noise showing up in the image comes from the subject being imaged, not the surroundings. Once this is achieved, the signal and noise are both coming from the patient. As a result, the signal goes up in proportion to the noise, and the ratio between the two can be better controlled, leading to good image quality.
Getting to that point, or even testing the idea, is at least a year off, Conolly said. The current prototype must be disassembled and new components, including more powerful magnets, must be added. Funding constraints limit the size of the staff (currently four, including graduate students) and the choice of equipment. The components are often wrested away from previous prototypes or less glamorous technologies. The receiver coil, for example, was taken off a refrigerator.
Whether a commercial system will eventually come out of the project is hard to say, but the discovery process is turning up some interesting nuggets. Because the system uses a very low power magnetic field when acquiring data, the gradients are virtually silent.
“If you’re two feet away, you can’t hear them,” Conolly said. “This wasn’t a motivating force in our choice of the design, but it has turned out to be an advantage.”
Then there is the challenge that goes with building a device whose operation is the antithesis of everything modern. Two decades of clinical MRI have demonstrated that a homogeneous field is absolutely essential to the acquisition of high-quality images. Ask any radiologist reading images made by a poorly tuned scanner. The Stanford design gets around the need for homogeneity, but few people in the MRI community are willing to believe it.
“When I tell people about our system, they think it is insane,” he said. “By their way of thinking, everything we are doing is wrong.”
MRI-Based AI Radiomics Model Offers 'Robust' Prediction of Perineural Invasion in Prostate Cancer
July 26th 2024A model that combines MRI-based deep learning radiomics and clinical factors demonstrated an 84.8 percent ROC AUC and a 92.6 percent precision-recall AUC for predicting perineural invasion in prostate cancer cases.
Breast MRI Study Examines Common Factors with False Negatives and False Positives
July 24th 2024The absence of ipsilateral breast hypervascularity is three times more likely to be associated with false-negative findings on breast MRI and non-mass enhancement lesions have a 4.5-fold likelihood of being linked to false-positive results, according to new research.
Can Polyenergetic Reconstruction Help Resolve Streak Artifacts in Photon Counting CT?
July 22nd 2024New research looking at photon-counting computed tomography (PCCT) demonstrated significantly reduced variation and tracheal air density attenuation with polyenergetic reconstruction in contrast to monoenergetic reconstruction on chest CT.