Initiatives to facilitate image sharing and the use of artificial intelligence are emerging as radiologists gather to discuss technology advancements.
At least two large initiatives, boasting of many partners, are being announced in conjunction with the start of the annual meeting and conference of the Radiological Society of North America, which convenes at Chicago’s McCormick Place in Chicago.
The partnerships begin to work at the conundrum of how to use the growing numbers of digital radiological images to improve care and expand the use of artificial intelligence across multiple organizations through cloud-based approaches.
At the 2019 RSNA conference, NVIDIA is expanding its Clara image-sharing model in a concept it calls NVIDIA Clara Federated Learning (Clara FL), which aims to support data sharing while enabling healthcare organizations to retain control of their images.
NVIDIA says the new approach will serve as a distributed, collaborative learning technique. It will run on the company’s recently developed EGX intelligent edge computing platform.
The system aims to gain access to huge volumes of data to train artificial intelligence models while protecting patient privacy. Clara FL runs on qualified edge servers from global system manufacturers, performing as a distributed client system that’s able to perform deep learning training locally while collaborating to train a more accurate global model.
Participating healthcare organizations include the American College of Radiology, the Massachusetts General Hospital and Brigham and Women’s Hospital’s Center for Clinical Data Science, and UCLA Health. NVIDIA says these early participants “are pioneering the technology. They aim to develop personalized AI for their doctors, patients and facilities where medical data, applications and devices are on the rise and patient privacy must be preserved.”
Hospitals that participate in the project label their own patient data using the NVIDIA Clara AI-Assisted Annotation SDK integrated into medical viewers like 3D slicer, MITK, Fovia and Philips Intellispace Discovery. Using pre-trained models and transfer learning techniques, NVIDIA AI assists radiologists in labeling, reducing the time for complex 3D studies from hours to minutes.
Meanwhile, in another announcement timed for the RSNA conference, Change Healthcare and Google Cloud are working together to build out what they’re calling a “next-generation enterprise imaging initiative.” Change Healthcare is hosting the initiative on the Google Cloud Platform.
The project aims to build and implement a cloud-native enterprise imaging network, and it’s getting a further boost from the addition of four health systems as development partners. They include Bronson Healthcare, Community Health Systems Professional Services Corp., Montefiore Nyack Hospital, and University of Wisconsin School of Medicine and Public Health and UW Health, Madison, Wis.
Each of these partners will work with Change Healthcare to help accelerate development of the solution by implementing the platform as it is built and providing ongoing, real-world feedback. Collectively, the partners manage 124 hospitals with an annual imaging volume of more than 5.6 million studies.
The partners will migrate more than 66 million studies to the Change Healthcare Enterprise Imaging Network cloud, with more than 2.8 petabytes being handled by the network. Customer implementations are expected to go live in the first half of 2020 as a fully managed software-as-a-service (SaaS) solution from Change Healthcare.
The imaging network will include an image archive, image viewer, imaging analytics and Change Healthcare’s AI Orchestration Services, a vendor neutral platform that streamlines integration of multiple AI algorithms across multiple vendors.
“Providers aren’t realizing the full benefits in improved care coordination, cost realization and reduced infrastructure complexity that true cloud-native solutions can provide,” says Tomer Levy, general manage of cloud solutions at Change Healthcare. “From the time we first partnered with Google Cloud, we’ve focused on building a solution that doesn’t simply replicate traditional on-premise systems, but delivers everything providers expect in an enterprise imaging service, plus clinical and operational capabilities that are only available through a true cloud-native SaaS platform.”