由 Lisa Chamoff
, Contributing Reporter | October 17, 2018
Kings College London (KCL) is embarking on a project to bring artificial intelligence (AI) solutions to radiology, which researchers say will rethink the practice of radiology and improve imaging in a public health system that serves eight million patients.
Researchers at three major London hospitals that are part of the National Health Service Trust citywide network will partner with computing company NVIDIA, deploying its DGX-2 IA research system and the NVIDIA Clara platform. NVIDIA data scientists and engineers will be on site to work on the project.
Kimberly Powell, vice president of healthcare at NVIDIA, said the DGX-2, with a large memory and 2 petaflops of computing power, allows for faster training of massive 3D volumetric data sets and enhances the use of Niftynet, which trains neural networks for medical imaging.
Quest Imaging Solutions provides all major brands of surgical c-arms (new and refurbished) and carries a large inventory for purchase or rent. With over 20 years in the medical equipment business we can help you fulfill your equipment needs
“It would historically take many weeks to do one training cycle,” Powell told HCB News. “With the DGX-2, you can cut that down to a couple of hours.”
The project will explore how to bring AI from a research setting into a clinical one, eventually providing the ability to, for example, more precisely classify stroke patients and providing the best course of treatment for them, Powell said.
KCL will use NVIDIA’s Clara platform to deploy the applications developed through research.
One piece of the project, which is still in the research phase, is federated learning, which would allow hospitals to leverage each other’s sensitive data without having to share it.
“How can you keep data safe and secure, but still be able to learn from that data?” Powell said.
“This is a huge opportunity to transform patient outcomes by applying the extraordinary capabilities of AI to ultimately make diagnoses earlier and more accurately than in the past,” said professor Sebastien Ourselin, head of the School of Biomedical Engineering and Imaging Sciences at KCL, in a statement. “This partnership will combine our expertise in medical imaging and health records with NVIDIA’s technology to improve patient care across the U.K.”
Powell said it is unique for a project of this scale to be undertaken by a large public health system. The system is also being confronted with a growing shortage of radiologists.
“They see the rate at which imaging is exceeding the number of radiologists going into the field,” Powell said. “They have a huge need to automate quickly.”
The results of the project will have an impact beyond KCL, according to Powell, and they intend to place what is learned out into the open source community, so everyone can benefit from it.
“If you look at the whole practice of radiology, there are going to be thousands of algorithms that will be developed,” Powell said. “We’re going to bring something to the world of radiology, not just this institution.”