The MONAI imaging framework is quickly put into production, accelerating AI applications in healthcare

MONAI (Medical Open Network for AI) is an open source framework optimized for the healthcare field. The upcoming NVIDIA Clara application framework is now in production for AI applications in healthcare and life sciences.

Launched in April, MONAI is now adopted by leading healthcare research institutions. It is a PyTorch-based framework that enables AI for medical imaging development through industry-specific data processing, high-performance training workflows, advanced reproducible reference implementations.

As part of the updated Clara offering, MONAI will offer more than 20 pre-trained models, including recently developed models for COVID-19, as well as the latest training optimizations for NVIDIA DGX A100 GPUs. This optimization increases training speed by a factor of six.

“MONAI is becoming the PyTorch of healthcare, paving the way for closer collaboration between data scientists and clinicians,” said Jayashree Kalpathy-Cramer, Ph. With federated learning, the global adoption of MONAI fosters global collaboration.”

MONAI is widely adopted in the healthcare ecosystem. The German Cancer Research Center, King’s College London, Massachusetts General Hospital, Stanford University and Vanderbilt University are all using this AI imaging framework. MONAI is used in everything from industry-leading imaging competitions to the first boot camp for the framework held in September. The boot camp attracted more than 550 registrations from 40 countries, including university undergraduates.

Dr. Bennett Landman of Vanderbilt University said: “MONAI has quickly become the deep learning framework of choice in healthcare. This step from research to production is critical to the implementation of AI applications in clinical care. NVIDIA is committed to For community-driven scientific research, enabling academia to contribute to a production-ready framework. This will help further innovation to build enterprise-ready features.”

new features

NVIDIA Clara brings the latest breakthroughs in AI-assisted annotation, federated learning, and production deployment to the MONAI community.

Its latest version revolutionizes AI-assisted annotation, enabling radiologists to use a new model called DeepGrow 3D. This enables complex 3D CT data to be marked with only one-tenth the number of clicks. Traditional methods of segmenting an organ or lesion by image or slice are time-consuming and can take up to 250 clicks for a large organ like the liver. Today, users can segment with fewer clicks.

NVIDIA Clara’s AI-assisted annotation tools and new DeepGrow 3D capabilities can be combined with Fovia Ai’s FAST AI annotation software to label training data and assist radiologists in reading images. Fovia offers the XStream HDVR SDK kit for viewing DICOM images with the industry leading PACS viewer.

To unlock rich radiology datasets, AI-assisted annotation is key. The technique was also recently used to label a public COVID-19 CT dataset released by the National Institutes of Health Cancer Imaging Archive. This labeled dataset was subsequently used in the MICCAI-certified COVID-19 Lung CT Lesion Segmentation Challenge.

Clara Federated Learning has recently enabled 20 hospitals around the world to conduct collaborative research to develop general AI models for COVID-19 patients. The EXAM model, which predicts the oxygen demand of COVID-19 patients, is available through the NGC software registry and is under development at the Mount Sinai Health System in New York, Diagnósticos da America SA in Brazil, the NIH Cambridge Biomedical Research Center, and the U.S. National Institutes of Health Clinical validation assessments are conducted in the Institute.

“The MONAI software framework provides key components for training and evaluating image-based deep learning models, and its open source approach helps foster a growing community for Contributing to exciting advancements such as federated learning.”

NVIDIA also extends NVIDIA Clara to digital pathology applications. In this area, existing open source tools cannot cope with huge image sizes. Clara for Pathology Early Access includes a reference pipeline for AI application training and deployment.

Jorge Cardoso, Chief Technology Officer, Centre for the Value of Medical Imaging and AI in London, said: “Interoperability of healthcare data, model deployment and clinical pathway integration is an increasingly complex and intertwined topic involving domain-specific expertise. Combined with the rest of the NVIDIA Clara ecosystem, it can help improve patient care and optimize hospital operations.”

The Links:   LM-DA53-21PTW CM200DU-24F CM200DY-12NF