Dr. Ellen Grant is a radiologist at Boston Children’s Hospital. She spends most days looking over thousands and thousands and thousands of MRI and CT scans.
In the early 2000s, she decided to tackle this increasing data overload. She first turned to the existing image analysis tools that were already in use in the hopes of finding a program that worked. That turned out to be harder than she thought.
Anyone who wants to use research software must be familiar with a terminal and command line—systems that cater to builders rather than users. It’s a barrier that’s kept research from making it to hospitals.
Enter Rudolph Pienaar, who began working with Grant and her colleagues in those early days. He looked at the existing image analysis programs and decided they couldn’t do what Grant needed.
“Reinventing the wheel is a real problem when it comes to computational medical research,” Pienaar says. “People are always writing the same back end—the same code—over and over again. This kind of repetition is a barrier to innovation, and you can only innovate so far if you’re always working on the infrastructure.”
That’s because there’s no standard IT infrastructure or platform to facilitate communication among researchers and medical professionals. Researchers typically run their algorithms on ad hoc technology they’ve cobbled together. And that means health care developers must replicate systems from the ground up.
Pienaar wrote script after script to pull and analyze data. It eventually became an unwieldy mess that—though useful—wasn’t practical. And each iteration was more complicated than the last.
Pienaar and Grant slowly came to realize the challenge wasn’t a missing strip of code, but the whole relationship with researchers, clinicians, and the technology between them. Technology was coming between them when it should have been connecting them.
Both parties needed a standard infrastructure that would support transparent collaboration. An IT stack that wasn’t a bunch of workstations randomly networked together, or even a high-performance computer.
They needed something that thrived on connectivity. They needed a cloud. An open source public cloud that allowed hospital developers to see the source code behind research programs.
They needed the Mass Open Cloud (MOC).
The MOC is part of the Massachusetts Green High Performance Computing Center located west of Boston. It’s an OpenStack® -powered infrastructure that faculty and students from participating institutions can use at no cost. It’s what researchers and hospitals needed to stop reinventing the infrastructural wheel.
With the MOC, Boston Children’s Hospital now has a cloud developed in collaboration with eight universities, four software development companies, two hardware manufacturers, a handful of government entities and a dozen community contributors. This coalition built a platform on the MOC that provides researchers and health care developers with a standard operating environment. It’s an open source tool for research programs, applications and algorithms as well as the hospital applications that use them. Researchers and clinicians can collaborate instantaneously to use and improve code.
It was a good idea that took on its own critical mass and pulled in more people—until it became a great idea and evolved into a revolutionary platform.
And the platform’s name is ChRIS.