NVIDIA Fast-Tracks MONAI Imaging, Clara Suites for Late December Access

NVIDIA's state-of-the-art AI models can help predict oxygen levels for severe COVID-19 patients.
Brad Bergan

The quickening pace of medical technology and healthcare fields is rapidly exceeding humanity's ability to keep up. Like tunnel vision, it can be easy to miss key areas where substantial improvements go unaddressed simply because of the limits of conventional computing and imaging technology.

However, AI and deep-learning applications are closing this gap — creating new solutions where previous blindspots in efficacy, task management, and data science research remained beyond our reach.

Which is why NVIDIA is fast-tracking a new medical open network for AI — called MONAI — an open-source framework and domain-optimized healthcare solution, according to a pre-briefing event Interesting Engineering (IE) attended last week.

The go-to AI and supercomputing firm also unveiled updates on the forthcoming release of NVIDIA Clara — an AI-powered application framework designed to upgrade life sciences and healthcare sectors. Both MONAI and Clara will be available for early access in late December 2020.

And there's more. NVIDIA also unveiled "Inception Alliance for Healthcare" — a new program to accelerate medical startups via NVIDIA's AI toolset and healthcare industry partners.


NVIDIA MONAI in Production
Some of MONAI's pre-trained models include COVID-19 applications. Source: NVIDIA

NVIDIA debuts MONAI, open-network medical AI platform

Initially introduced in April and already integrated into leading institutions of healthcare research, MONAI uses PyTorch to enhance the creation and execution of AI technology for medical imaging and data processing for domain-optimized applications, including life sciences.

Among MONAI's 20 pre-trained models are several new ones developed for the COVID-19 crisis, which include novel training optimizations on NVIDIA DGX A100 GPUs to support an acceleration in turnaround time at 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 — QTIM lab director at the Athinoula A. Martinos Center for Biomedical Imaging at MGH, according to the press release. "Global adoption of MONAI is fostering collaboration across the globe facilitated by federated learning."

NVIDIA MONAI Accelerating Healthcare
MONAI is rapidly becoming the deep-learning industry standard for healthcare. Source: NVIDIA

Linking community-driven science to commercial production

So far, industry adoption of MONAI into the healthcare ecosystem has seen success. Stanford, DKFZ, King's College London, Vanderbilt, and Mass General are among the early adopters of the AI framework for imaging in healthcare. MONAI has and continues to see use in several applications, including industry-shaping imaging competitions, and the first boot camp focused on the framework — held in September 2020, which gathered more than 550 registrants from 40 countries, in addition to undergraduate students.

"MONAI is quickly becoming the go-to deep learning framework for healthcare. Getting from research to production is critical for integration of AI applications into clinical care," said Professor Bennet Landman of Vanderbilt University, in the press release. "NVIDIA's commitment to community-driven science and allowing the academic community to contribute to a framework that is production-ready will allow for further innovation to build enterprise-ready features."

Most Popular
NVIDIA Clara Healthcare Platform
The newest version will help radiologists label 3D CT data, with one-tenth of the normal number of clicks. Source: NVIDIA

NVIDIA Clara to enhance radiologists' imaging capacities

NVIDIA Clara takes the latest advances in AI-assisted annotation, production deployment, and deployment to MONAI users throughout and beyond the healthcare industry.

NVIDIA Clara Imaging
The newest NVIDIA Clara version has an incredibly wide range of applications, from advanced theoretical research to the COVID-19 crisis. Source: NVIDIA

The newest version will help radiologists label complex 3D CT data in one-tenth of the clicks — via a new model called DeepGrow 3D. Instead of the conventional and time-consuming method of segmenting lesion or organ images, one at a time or slice by slice — which may take 250 clicks for larger organs like the liver — NVIDIA Clara users can segment organs of interest with far fewer clicks, according to the press release.

NVIDIA Clara Software Writes Software
Clara's AI-assisted annotation features are integrated with Fovia AI. Source: NVIDIA

NVIDIA Clara's AI-assisted annotation tools — along with the new DeepGrow 3D feature — are integrated with Fovia AI's F.A.S.T. AI Annotation software, which enables users to label training data and enhance radiologists' reading capabilities. Fovia includes the XStream HVR SDK suite, allowing users to analyze DICOM images integrated with major industry PACS viewers.

NVIDIA's Clara's AGX developer kit is already available via select partner developers. Source: NVIDIA

AI models predict oxygen levels in COVID-19 patients

Annotating with AI assistance is crucial for accessing rich radiology datasets — a process recently used to label the public COVID-19 CT dataset, later published by The Cancer Imaging Archive at the U.S. National Institutes of Health. Additionally, this labeled dataset was used during the MICCAI-endorsed COVID-19 Lung CT Lesion Segmentation Challenge.

Research collaborations from 20 hospitals worldwide curated a generalized AI model tailored for COVID-19 patients. For example, the EXAM model predicts required levels of oxygen in patients who've contracted the disease is accessible via the NGC software registry. As of writing, it's undergoing clinical validation at Mount Sinai Health System in New York, along with NIHR Cambridge Biomedical Research Centre in the U.K., Diagnósticos da America SA in Brazil, and the NIH.

"The MONAI software framework provides key components for training and evaluating imaging-based deep learning models, and its open-source approach is fostering a growing community that is contributing exciting advances, such as federated learning," said Professor of biomedical data science, radiology, and medicine Daniel Rubin of Stanford University, according to the press release.

Clara's early-access pathology applications integrate reference pipelines

NVIDIA also plans to widen the release base of NVIDIA Clara to include pathology applications — where greater scales of deployment can choke conventional open-source AI tools. Clara's early-access pathology-intensive version includes reference pipelines for initial training and eventual deployment of AI applications.

"Healthcare data interoperability, model deployment and clinical pathway integration are an increasingly complex and intertwined topic, with significant field-specific expertise," said CTO Jorge Cardoso of the London Medical Imaging and AI Centre for Value-Based Healthcare, in the press release. "Project MONAI, jointly with the rest of the NVIDIA Clara ecosystem, will help deliver improvements to patient care and optimize hospital operations."

NVIDIA Inception Alliance
The NVIDIA Inception program involves more than 1,000 healthcare members. Source: NVIDIA

NVIDIA unveils Inception Alliance healthcare accelerator program

NVIDIA also debuted a new initiative for medical AI startups to find new ways of tracking clinical and research developments, accelerating their performance timeframes via NVIDIA and its healthcare industry partners, according to a press release shared with IE under embargo via email.

Members with the premier status of NVIDIA Inception — the company's accelerator program for more than 6,500 AI and data science startups in 90 countries may now opt-in to the GE Healthcare Edison Developer Program.

This is moving forward via an integration with the GE Healthcare Edison Platform, which adds a global network and new scales for commercial and clinical activities — since GE's platform includes an install base of 4 million imaging, monitoring, and mobile diagnostics units in 160 countries, with 230 million exams and related data, said the press release shared with IE.

NVIDIA Exploding Demand
The demand for AI inferencing in healthcare is exploding. Source: NVIDIA

Nuance AI marketplace helps radiologists save more lives

"Startups are on the forefront of innovation and the GE Healthcare Edison Developer Program provides them access to the world's largest installed base of medical devices and customers," said Vice President and General Manager Karley Yoder of Artificial Intelligence at GE Healthcare, in the press release. "Bringing together the world-class capabilities from industry-leading partners creates a fast-track to accelerate innovation in a connected ecosystem that will help improve the quality of care, lower healthcare costs and deliver better outcomes for patients."

"With Nuance's deep understanding of radiologists' needs and workflow, we are uniquely positioned to help them transform healthcare by harnessing AI. The Nuance AI Marketplace gives radiologists the ability to easily purchase, validate and use AI models within solutions they use every day, so they can work smarter and more efficiently," said Senior Vice President Karen Holzberger of Nuance's Diagnostic Division. "The AI models help radiologists focus their time and expertise on the right case at the right time, alleviate many repetitive, mundane tasks and, ultimately, improve patient care — and save more lives."

"Connecting NVIDIA Inception startup members to the Nuance AI Marketplace is a natural fit — and creates a connection for startups that benefits the entire industry," added Holzberger.

As vaccines like Moderna's — which just received a 100% efficacy rating in treating severe COVID-19 cases — seek FDA approval for emergency use authorization, the expansion of NVIDIA's MONAI, Clara, and Inception Alliance Platforms will help to streamline and optimize our emergence from this year of calamity, into 2021 — while also opening new pathways previously inaccessible for not only medical fields like radiology, but also medical startups in the commercial sector — working together to optimize patient care, and improve the human condition.

message circleSHOW COMMENT (1)chevron