Computer Science Students, Professor Find Novel Way to Improve Cancer Treatment by Using Resources

May 26, 2010

launAt UNM, innovation frequently begins with a professor interested in finding a better way to do something. This story begins with Shuang Luan, an assistant professor of Computer Science who finds great satisfaction in finding ways to improve treatment for cancer patients. He is currently exploring ways to make gamma knife radiation treatments as precise as possible. Photo: Shuang Luan

Last fall as he was working with physics graduate student Roy Keyes and Computer Science undergraduate Christian Romano and Dorian Arnold, a colleague and an assistant professor in Computer Science they came up with an inexpensive way to do the complex calculations needed to map radiation treatments by. They simply bought computer time from with Luan’s credit card. 

Planning ways to deliver radiation to tumors in cancer patients takes time. 
Extremely precise calculations that will target the tumor with as little damage to surrounding healthy tissue can take hundreds of hours to do the complex Monte Carlo calculations needed to determine where every proton and electron from the treatment beam is most likely to go. It just is not financially practical for clinics to spend that much time to map the target area of each tumor. So physicians are forced to take shortcuts.

The medical physicists who calculate how much radiation should be used and at what angle it should hit the tumor normally use a model that treats the human body as a vessel of water. But radiation travels through muscle and fat and bone differently and the time it would take to calculate for those differences is counted in the hundreds of hours.

Keyes says the problem comes in the enormous amount of time it takes to maximize the radiation beam so that it hits every part of an irregular target like a tumor, and minimize the radiation exposure to other parts of the body.

“Imagine you want to treat a tumor in the spinal column. It’s important to get the right dose into the tumor, but also very important to avoid the spinal cord as well as organs that might be nearby, such as the lungs or kidneys,” he says. “Because of the different composition of the organs you are concerned about and the potentially large total volume, the calculation could take many days of computer time. If it takes three days, that’s too long.”

No insurance plan can afford to pay for that kind of precision.

But what if clinics didn’t have to buy and maintain sophisticated computers to perform the calculations? What if medical physicists were able to put together treatment calculations that could run in minutes? What if clinics could buy the computer time from Amazon at .10 cents an hour with a credit card on an Amazon account? It is a whole new way to use cloud computing.

Keys made the treatment calculations. Romano figured out how to break the problems into pieces so they could use 200 computer nodes to run the calculations. They used Luan’s credit card and Amazon account.

“In the Computer Sciences lab upstairs they probably have fifty or sixty machines, and a lot of students using them. You cannot just say I’m going to use them all today. But in cloud computing, we just basically type in a credit card number and say give us 200 nodes. And they give it to you in maybe five minutes,” Luan says.

If it all works as they think it will, patients should have fewer side effects from radiation treatments. Physicians should be able to treat cancer more precisely. And insurance companies may find some savings.

Keys and Romano have submitted an invention disclosure to STC.UNM, and STC.UNM has filed a patent application on the invention. STC.UNM is the university’s wholly owned corporation that commercializes intellectual property for UNM research.