Dr. Nicholas Dwork, an assistant professor in the Department of Biomedical Informatics at the University of Colorado School of Medicine, has filed a provisional patent for a technology that could improve the speed of 3D magnetic resonance imaging (MRI) scans. The present invention can lead to faster results, increase the clinical utility of MRI, and ultimately improve patient care.
Generating images of internal organs and tissues when a patient enters an MRI machine’s magnetic tube involves complex mathematics and engineering. Any movement of the patient can corrupt the image, and the scan can take an hour or more. During this time, the machine makes loud noises as radio waves bounce off the body structure to create images. Because the field of view—the portion of space being imaged—is usually rectangular, radio waves also bounce off areas outside the body. This is where Dwork’s technology comes into play.
Assume you don’t know anything about the image you want to make. In this case, you will need the maximum amount of information to produce an image, which means collecting the maximum amount of data. How can we reduce the amount of data required? If we had more information about the picture we were making, we wouldn’t need as much data. “
Nicholas Dwork, PhD, Assistant Professor, Department of Biomedical Informatics, University of Colorado School of Medicine
His method tunes the sampling pattern produced by the machine’s magnetic field.
“They’re called pulse train maps,” Dwork explained. “Think of an MRI as a musical instrument. Like musical instruments, they can play different songs. You can think of it as a musical score for an MRI machine.”
He worked with faculty and staff in the Department of Radiology to develop the programming application and will continue to refine the technology through his own research.
“Mainly, it means combining this particular method of speeding up MRI with other techniques to make the scans extremely fast,” he said. “These other techniques include so-called fractional Fourier sampling, parallel imaging, compressed sensing and deep learning.”
Dwork estimates that his method of MRI scanning can reduce time by about 25 percent, allowing doctors to get results faster and reducing the time patients spend in the MRI tube. He envisions the faster scans being used for multiple purposes, potentially extending the use of MRI to very young children and pregnant women. He hopes to study its use in twin transfusion syndrome, a rare condition in which twin fetuses share the same placenta and blood vessel network. Fast, accurate MRI can guide surgeons in cauterizing blood vessels in the placenta to address the condition and improve fetal outcomes.
Dwork has focused his entire career on applying advanced mathematics to medical problems, and he says his work would not have been possible without the support he received at CU. As a member of the CU Boulder Department of Applied Mathematics Affiliated Faculty Program, he introduces medical problems to the mathematics department while also bringing novel mathematical insights to the CU Anschutz Medical Campus. He also worked with CU Innovations, CU Anschutz’s technology commercialization and venture development office, to market the invention to imaging companies and conduct research in a clinical setting.
“CU Anschutz is the perfect place for the type of work I love to do, applying advanced mathematics to solve medical problems,” he said. “I am extremely grateful to the Department of Biomedical Informatics for giving me this professional opportunity and providing me with so many resources to maximize my likelihood of success.”
University of Colorado Anschutz