How Elekta Cloud AI Infrastructure Enables Radiation Therapy Breakthroughs

Announced by Microsoft + NVIDIA

Despite the many challenges, some of the most successful examples of moving innovative AI applications into production come from healthcare. At this VB spotlight event, learn how organizations across all industries can follow proven practices and leverage cloud-based AI infrastructure to accelerate their AI initiatives.

Register to watch on-demand for free.

From pilot to production, AI is a challenge for every industry. But healthcare, a highly regulated and high-risk sector, faces particularly complex obstacles. “Dedicated” cloud-based infrastructure optimized for AI is emerging as a key foundation for innovation and operationalization. Leveraging the flexibility of the cloud and high performance computing (HPC), companies across all industries are successfully scaling proofs of concept (PoCs) and pilots to production workloads.

VB Spotlight is with Silvain Beriault, AI Strategy Lead and Principal Research Scientist at Elekta, a top global innovator of precision radiotherapy systems for cancer treatment, and AI Platform and Infrastructure Principal Lead, Microsoft Azure. Collected John K. Lee. They join his VB Consulting Analyst Joe Maglitta, who says the cloud-based AI infrastructure will be a collaboration and global R&D effort aimed at improving and expanding Elekta’s brain imaging and MR-guided radiation therapy worldwide. We discussed how we fostered innovation improvements.

Three major advantages

According to Lee, the benefits of end-to-end, on-demand, cloud-based infrastructure as a service (IaaS) for AI are elasticity, flexibility, and simplicity.

Enterprise AI typically starts with a PoC, Lee said. Get started with a single credit card. As models become more complex and the need for additional computational power increases, the cloud is the perfect place to scale the job. This includes scaling up or increasing the number of interconnected GPUs on a single host to increase server capacity, or scaling out or increasing the number of host instances to improve overall system performance. It is included.

The flexibility of the cloud enables organizations to manage workloads of any size, from large enterprise projects to small, low-power operations. No matter the size of the initiative, a purpose-built cloud infrastructure service delivers value in much less time than building an on-premises AI architecture from scratch, improving TCO and ROI, says Lee. explains Mr.

Regarding simplicity, Lee said pre-tested, pre-integrated and pre-optimized hardware and software stacks, platforms, development environments and tools make it easy for companies to get started. I’m here.

COVID accelerates Elekta’s cloud-based AI journey

Elekta is a medical technology company developing image-guided clinical solutions for the management of brain disorders and improved cancer care. When the COVID pandemic forced researchers out of their labs, company leaders saw an opportunity to accelerate and expand efforts to move AI R&D to the cloud that began years ago.

The division’s head of AI believes a more robust and accessible cloud-based architecture for improving the suite of AI-powered solutions will help improve access to healthcare, including in underserved countries. I knew it would help advance Elekta’s mission to get more.

From a cost analysis perspective, Elekta also knew it would be difficult to estimate current and future needs in terms of high performance computing. They considered the costs and limitations of maintaining an on-premises infrastructure for AI. Balliol points out that the overall cost and complexity goes far beyond buying GPUs and servers.

“Trying to do it yourself can quickly become difficult. With frameworks like Azure and Azure ML, you don’t just have access to GPUs,” he explains. “You get a whole ecosystem to do AI experiments, document AI experiments, and share data between different R&D centers. Common he has ML ops tools.”

The pilot was straightforward. We automated the contouring of organs in MRI images to accelerate the task of outlining treatment targets and creating organs at risk of avoiding radiation exposure.

The ability to scale up and scale down was very important to the project. In the past, “we would launch as many as 10 training experiments in parallel to do model hyperparameter tuning,” he recalls. “We also didn’t train at all because we were just waiting for the data to be curated and ready. was very important to

The company was already using the Azure framework, so it looked to Azure ML as the infrastructure, as well as critical support as the team learned how to start jobs in the cloud using the platform portal and APIs. Did. Microsoft worked with the team to build a highly specialized data infrastructure for their domain, addressing important data security and privacy concerns.

“As of today, we have expanded our automated contouring using an all-cloud-based system. Using this infrastructure, we are able to scale our research efforts to more than 100 organs at multiple tumor sites. Moreover, scaling enabled us to extend beyond simple segmentation to other more complex AI studies in RT, increasing the potential to positively impact patient care in the future. .”

Choosing the right infrastructure partner

Ultimately, Balliol said, by adopting a cloud-based architecture, researchers can focus on their work and develop the best possible AI model, rather than building and “babysitting” AI infrastructure. It is said that it will be possible to develop

Choosing a partner who can provide such a service is important, commented Lee. A strong provider must bring in strong strategic partnerships that help keep its products and services on the leading edge. He said the collaboration between Microsoft and his NVIDIA to develop a foundation for enterprise AI is important for customers like Elekta. But he adds that there are other considerations.

“Remember it’s not just about product offerings and infrastructure. Do they have a whole ecosystem? Do they have a community? They have the right people to help you.” mosquito?”

Register to watch on demand now!

agenda

First-hand experience and advice on how best to accelerate the development, testing, deployment, and operations of AI models and services
The critical role AI infrastructure plays in moving from POCs and pilots to production workloads and applications
How a cloud-based “AI-first approach” and front-line, proven best practices can help you scale AI faster and more effectively across departments or around the world, regardless of industry

speaker