- FDA Proposes Rules for Devices Using Artificial Intelligence
- Providers Need to Embrace Healthcare Artificial Intelligence
ACR intends the platform, called ACR AI-LAB, to help members and other radiologists learn the basics of healthcare AI and use those skill in their practices, explained Bibb Allen Jr., MD, chief medical officer and diagnostic radiologist at Grandview Medical Center in Birmingham, Alabama.
“Empowering radiologists to develop AI algorithms at their own institutions, using their own patient data, to meet their own clinical needs, could potentially accelerate the development of AI for diagnostic radiology,” Allen wrote in a ACR blog post.
The goal of ACR AI-LAB is to move radiology AI out of large institutions with extensive informatics and data science resources into smaller institutions and clinics.
Goal Is Clinical Integration of Radiology AI
“Not only are most algorithms only being used at the institutions where they were developed, at present there are no clear mechanisms to validate these algorithms across multiple institutions or to provide for clinical integration of AI. Many of the algorithms developed thus far have proven brittle in actual clinical use when tried at different institutions,” Allen related.
Radiologists have not widely participated in AI development because they have limited access to computational resources and can be intimidated by the complexity of AI architecture.
“To maximize the development and adoption of AI in clinical practice, radiologists need to be empowered to create AI tools in their own institutions,” he observed.
ACR will unveil ACR AI-LAB at its 2019 annual meeting being held May 18-22 in Washington, DC. At the conference, members will be able to test out the ACR AI-LAB’s cloud-based data and computing tools to annotate images for use in developing algorithms. After the meeting, ACR will provide online access to ACR AI-LAB and sample data from publicly available patient datasets.
Allen related that a standard set of application programming interfaces will be created to ensure that locally developed algorithms can be integrated into electronic health records, picture archiving and communication system, and other platforms.
In addition, ACR and NVIDIA announced that the NIVIDIA Clara AI toolkit is being integrated into the ACR AI-LAB platform. The toolkit includes libraries for data annotation, model training, model adaptation, model federation, and large-scale deployment.
Initial Pilot with Ohio State, Mass General
ACR and NVIDIA conducted an initial pilot with the Ohio State University (OSU) and the Massachusetts General Hospital and Brigham and Women’s Hospital’s Center for Clinical Data Science (CCDS). The pilot helped them define methods needed to enable facilities to work together to refine AI algorithms without sharing potentially sensitive patient data.
Bringing an AI model to patient data can help increase diversity in algorithm training, ease validation of the algorithms, and enable radiologists to learn the steps needed to adapt algorithms to their institutions’ clinical needs, ACR and NVIDIA said in a release.
Using the NVIDIA Clara AI toolkit, OSU was able to import a pre-trained model developed by CCDS. This model was customized to local variables and labeled OSU data for further testing and improvement of the algorithm, all of which took place behind its firewall.
The pilot resulted in an accurate and enhanced cardiac computed tomography angiography model, and the shared approach reduced algorithm training, validation, and testing times.
The architecture used in the pilot program enabled data aggregation, image annotation, image pre-processing and transformation, algorithm transfer and local computing for algorithm improvement.
“Enabling a network of artificial intelligence between hospitals will create more robust algorithms, greater efficiencies and likely lead to better patient outcomes,” said Richard White, MD, chair of the Department of Radiology and Medical Imaging Informatics at the Ohio State University Wexner Medical Center. “This will give us access to high-quality algorithms that will help us accelerate deep learning and machine learning in healthcare.”
NVIDIA Clara powers GE Healthcare’s Edison AI platform and the Nuance AI Marketplace, both of which support the AI-LAB and are solutions for the deployment of AI within the radiology workflow.
GE Healthcare said that by integrating with ACR AI-LAB, it will allow ACR members and other radiologists to develop and deploy their algorithms across hospitals and research centers nationally.
