Mayo Clinic and IBM score advances in real-time medical imaging

Highlights 

• Showcases leadership in research, medicine and technology
• Applies computationally intensive solutions to real-life diagnostic problems 
• Enables collaborative innovation to drive better patient information and treatment
   

Computed tomography (CT), magnetic resonance imaging (MRI) and other medical imaging techniques have redefined how physicians diagnose and treat life-threatening diseases, greatly reducing the need for invasive surgery. Computers can extract more information than humans can visualize, detecting, for instance, the way the heart is perfusing tissues or identifying textures imperceptible to the eye.

There’s a price, of course: enormous data files, compute-intensive calculations, and often, prohibitive costs that can limit patient access. IBM and Mayo Clinic have launched the Medical Imaging Informatics Innovation Center (MI3C) to help change that picture. A joint collaboration between these respected leaders, the MI3C brings together researchers, clinicians and technology experts on Mayo’s Rochester, Minnesota campus in an environment that fosters innovation.

Already, the pairing is paying off in new solutions for the kinds of diagnostic problems clinicians see every day but can’t begin to resolve on their own.

“This facility allows us to explore projects in medical imaging and radiology that can help provide faster and better information for our physicians, and in turn, improved treatments for our patients,” said Bradley Erickson, M.D., Ph.D., head of Mayo’s Radiology Informatics Lab and MI3C co-director.

Collaborating with a shared vision

IBM and Mayo have been tackling challenges together for years, with a shared vision of information-based medicine.

“MI3C brings together the skills and resources IBM and Mayo Clinic can collectively apply to the medical imaging space,” says Bill Rapp, IBM distinguished engineer and co-director of MI3C. “IBM has world-class research and development teams focused on the fundamental algorithms that drive medical imaging informatics, while Mayo Clinic provides its expertise for exploiting these algorithms in a working radiology environment.”

The algorithms are what keep the team focused: How can we get computers to solve problems more efficiently? To calculate in seconds what now takes hours? To take what humans can imagine, but cannot execute?

Harnessing technology to deliver real results

At the center’s heart are the latest in high-end imaging platforms and computational hardware, including IBM’s breakthrough computing system based on blade technology and the Cell Broadband Engine, jointly developed by IBM, Sony and Toshiba for the computer gaming industry.

The system proved its mettle in a complex image registration challenge, with the computer-enhanced alignment of images from different dates or devices. With aligned images, a radiologist can more easily detect structural changes like the growth or shrinkage of tumors. Leveraging Cell technology and 3D linear algorithms, MI3C reduced registration time for 98 image sets—millions of pixels—from seven hours to under nine minutes. That’s fifty times faster than a traditional processor configuration.

Another algorithm applied to motion correction yielded crisper 2D and 3D images in cases where breathing, heartbeats or patient movement might have compromised an MRI. In addition to these demonstrations, MI3C has numerous potential projects to address, including video swallow studies, tumor growth modeling and coronary artery calcification.

The center’s goal is to provide not only tools with the added diagnostic powers that physicians and clinicians need, but also the opportunity for those users to uncover the information they need much more quickly.

Bringing industry insights to healthcare informatics

The MI3C’s mission is to increase interest and participation in imaging projects from medical imaging manufacturers, software developers and other third parties. With a growing collection of analytics and algorithms, the collaborators are working to transform their learning into improved patient care.

For more information

To learn more about MI3C capabilities or explore project opportunities, contact Bill Rapp at rapp@us.ibm.com or Bradley Erickson, M.D., Ph.D., at bje@mayo.edu, or visit: