Federated learning heralds AI benefits for radiology
Radiology workflows are set for a speed and efficiency boost when the UK’s first AI platform for NHS hospitals emerges from a new development agreement between King’s College London and graphics processing unit (GPU) pioneer NVIDIA.
With the assurance of patient data privacy afforded by a federated infrastructure, the platform will enhance the benefits of AI in medical imaging, eventually allowing specialists to train computers to automate one of the most time-consuming aspects of radiology interpretation.
Clinicians from major London hospitals onsite at King’s College Hospital, Guy’s and St Thomas’ will join researchers and engineers from King’s and NVIDIA with the aim of accelerating the discovery of data strategies, resolving targeted AI problems and speeding up deployment in clinics.
This will be the first time that federated learning has been applied to algorithm development in the NHS, ensuring the privacy of patient data while allowing AI models to be built on data from different patient demographics and clinical attributes.
AI algorithms will be developed at multiple sites – ultimately, in different NHS trusts across the UK – using data from each hospital without the need for it to travel beyond its own domain.
The collaboration is part of King’s London Medical Imaging & AI Centre for Value-Based Healthcare, an ongoing project intended to transform 12 clinical pathways in oncology, cardiology and neurology, as well as improve diagnoses and patient care in the NHS.
WHAT'S THE IMPACT?
In the first phase of the project, King’s will implement NVIDIA’s 2-petaflop GPU-powered DGX-2 supercomputers plus the NVIDIA Clara AI toolkit, which has its own imaging technologies and is a
key part of the NVIDIA Clara developer platform. Intelligent workflows can be built on the platform, which consists of libraries for data and image processing, AI model processing and visualization.
The collaborative project will allow the NHS to share and analyse data on an unprecedented scale, bringing together a critical mass of industry and university partners, according to Professor Sebastien Ourselin, head of the School of Biomedical Engineering & Imaging Sciences at King’s College London.
“This centre marks a significant chapter in the future of AI-enabled hospitals, and the infrastructure is an essential part of building new AI tools which will benefit patients and the healthcare system as a whole,” he said.
WHAT'S THE TREND?
As models are developed at individual trusts, the data gathered will give more accurate and representative insight into patients from each particular are. The NHS will also be able to combine these trust-specific models to build a larger, demographically richer model overall.
The work could lead to breakthroughs in classifying stroke and neurological impairments, determining the underlying causes of cancers and recommending the best treatments for patients.
“Together with King’s College London, we’re working to push the envelope in AI for healthcare,” said Jaap Zuiderveld, vice president for EMEA at NVIDIA. “DGX-2 systems with the NVIDIA Clara platform will enable the project to scale and drive breakthroughs in radiology, and ultimately help improve patient outcomes within the NHS.”
ON THE RECORD
“The NVIDIA DGX-2 AI system’s large memory and massive computing power make it possible for us to tackle training of large, 3D datasets in minutes instead of days, while keeping the data secure on the premises of the hospital,” said Professor Ourselin.