A Physicist Walks Into a Bar...: How a Random Encounter Helped Track Down Elusive Pain

A decade ago, Yale physicist Kevin Koch was “just hitting the bars” when he struck a random conversation with a fellow graduate student at Gryphon's Pub, a Yale student hangout. As such affairs go, they were soon discussing neuroscience, consciousness, brain imaging and other heady matters. “This student had just come from a lecture by Robert Shulman, a biophysicist and the founder of Yale’s Magnetic Resonance Research Center (MRRC), an important research hub for using magnetic resonance to study the brain,” Koch says. Shulman and his faculty were looking for physics graduate students to work on technical problems in magnetic resonance, and Koch, who spent the previous summer working at GE Healthcare’s imaging lab in Waukesha, was intrigued. “I’ve always been interested in the application of physics in medicine,” Koch says. “I went in, sat down with Shulman, and got hooked. MRI is the forefront of applied physics in medical imaging. There are so many interesting applications.”

Clear Vision: Kevin Koch’s math models helped untangle the effects of magnetic field distortions caused by metal implants.

Koch, 35, is now pursuing those applications at GE Healthcare and his research is already helping doctors like Hollis Potter from Manhattan’s Hospital for Special Surgery. Potter spent 15 years trying to understand elusive pain coming from her patients’ hips, knees and other implants. She began to crack the riddle after Koch’s research at GE untangled the effects of magnetic field distortions caused by metal implants. “I pushed implant designs through mathematical modeling software that I developed in the lab at Yale,” Koch says. “The models allowed us to better understand the impact of the magnetic field distortions.”

Koch’s collaboration with Potter, along with technical partners at Stanford University, allowed GE to develop an innovative magnetic resonance imaging (MRI) system called MAVRIC SL. It allows doctors to see damaged tissue around MR Conditional implants, but also nerve impingement, the integrity of joints and bones fortified with stainless plates and screws, and other medical issues.

When Koch joined the Yale research lab, he knew little about software, but a chance encounter helped out again. “One of my roommates at Yale was a theoretical physicist and a programmer, and he helped get me off the ground,” Koch says. He then taught himself to program in C, a popular general-purpose programming language, and MATLAB, a code used to manipulate numbers and MRI data.

GE hired Koch right after he got his Yale PhD. Potter arranged for Koch to meet with orthopedic surgeons and their patients to “understand how frustrated they were.”

“I didn’t want an ivory tower physicist,” she says. Therefore, “he was thrown full thrust into the clinical translation of the problem. He understood how important it was.”

Koch began by tracking the spatial variation of magnetic fields generated by the implants. “It became apparent that we did not want to follow a conventional x-y-z Cartesian imaging coordinate system anymore,” Koch says. “We collect an arbitrary collection of 3-D volumes and stack them up like a bunch of blocks. We just have to make sure than we add them up properly.”

Koch’s models did not solve the problem, but their results pointed him in the right direction. “Eventually, our group at GE was able to develop a novel MRI acquisition concept that addressed the distortions predicted by the models,” he says.

It took Potter and Koch just over a year to build a working prototype. They took their first image in 2010. Potter says that her hospital has already imaged 3,000 patients with MAVRIC SL. The system does not require a new machine. Hospitals only need to upload new software on existing MRI scanners.

“It gives people an answer for their pain,” Potter said. “From an emotional standpoint of dealing with illness, that’s huge.”