How Siemens Healthineers Is Streamlining Cancer Therapy with AI

Leading healthcare company accelerates radiation oncology workflow with NVIDIA HGX servers.
by Abdul Hamid Halabi

Cancer incidence rates are on the rise — expected to increase by 63 percent over the next two decades. To meet the growing demand for care, medical technology leaders are turning to AI tools that can help radiation oncologists provide high-quality, individualized treatment faster.

One of the world’s leading healthcare companies, Siemens Healthineers, is using an NVIDIA GPU-based supercomputing infrastructure to develop AI software for generating organ segmentations that enable precision radiation therapy.

Siemens Healthineers’ Sherlock AI supercomputer is powered by NVIDIA HGX 1 and HGX 2 servers loaded with NVIDIA V100 Tensor Core GPUs. The system provides 20 petaflops of performance and is used to run over 500 AI experiments daily.

Both Siemens Healthineers and NVIDIA this week are sharing their latest work in AI for medical imaging at the Society for Imaging Informatics in Medicine annual conference, held outside Denver, Colorado. The event brings together the medical informatics community to share, debate, and address the challenges and opportunities facing medical imaging.

Augmenting Radiation Therapy Workflows

Radiation therapy for cancer patients is a complex workflow that includes modeling the patient, contouring the target and organs at risk, simulating the treatment, planning and delivering the treatment.

One of the most time-consuming tasks in this process is protecting the healthy organs at risk that surround a patient’s tumor and need to be spared from excessive radiation dose. Traditionally, radiation oncologists contour the tumor target volume and organs at risk, deciding how much radiation should be used to treat tumors without damaging neighboring normal tissue.

To help oncologists develop radiation treatment plans faster, Siemens Healthineers uses syngo.via RT Image Suite, a software tool that automatically outlines organs using AI-assisted AutoContouring. Trained on over 4.5 million images using the Sherlock supercomputer, the AI model saves radiation-oncologist time and eases organs-at-risk contouring tasks. In their current research, Siemens Healthineers automatically outlines 28 organs using AI technology.

“AI-assisted AutoContouring helps save time and improve standardization in organ at risk contouring,” said Dr. Fernando Vega, Head of Software and Concept Definition for Radiation Oncology at Siemens Healthineers. “This allows radiation-oncologists to better focus on other crucial aspects of patient care.”

Tapping into Software to Write Software 

Behind this explosion of AI in medical imaging is a new dynamic within the software development paradigm: the advent of software that writes other software.

Traditionally, engineers have written applications from start to finish, a time-consuming process that requires niche computing expertise. Now, with access to powerful compute resources, AI algorithms can leverage training data to learn processes like medical image analysis without every element being explicitly coded by a developer.

Siemens Healthineers, which has been involved in machine learning since the 1990s, is harnessing this AI capability with the Sherlock system. The supercomputer learns from the company’s massive data lake of over 750 million curated images as well as radiology reports and clinical and genomic data. So far, it has led to the development of more than 40 AI-powered applications approved for clinical use.

“We believe that AI is starting a new era in software development, where advanced neural network architectures, large collections of curated data, and massive computational power come together to deliver tremendous performance and high clinical value,” said Dr. Dorin Comaniciu, senior vice president of artificial intelligence and digital innovation at Siemens Healthineers.

Simple and Scalable Infrastructure

Siemens Healthineers’ 20 petaflops Sherlock supercomputer addresses a key computing need in the healthcare industry for an optimized and scalable infrastructure that can be used to develop deep learning tools for imaging and other clinical applications.

The NVIDIA DGX POD reference architecture provides a tested infrastructure for setting up a scalable AI computing system. Through the DGX-Ready Data Center program, NVIDIA and its colocation service providers offer simplified, rapid deployment for customers building and deploying world-class AI data centers for the healthcare industry.

For more on how NVIDIA’s AI platform is enabling advances in medicine and research, see the NVIDIA Healthcare page.