The Inquirer

Nvidia's Project Clara supercomputer looks to speed up detection of deadly diseases

Firm pushes medical imaging supercomputer - based on a ton of its GPUs, of course

Clara
0 Comments

SILICON VALLEY: NVIDIA has unveiled a new healthcare project which aims to improve imaging in healthcare.

Called Project Clara, the new tech was unveiled at the firm's GPU Technology Conference (GTC), in Silicon Valley. It comes in the form of a medical imaging supercomputer, which takes advantage of advancements in computation to renew the capabilities of existing machines that are currently vital to the early detection of deadly conditions, such as cancer and heart disease. 

"There are about three million medical imaging instruments installed around the world. With only a couple of hundred thousand new ones sold each year, it would take decades to update this install base," Nvidia's vice president of Healthcare, Kimberly Powell, told the INQUIRER in an interview at GTC. 

However, she explained that about a decade ago, researchers realised NVIDIA GPUs could provide a more efficient architecture for medical imaging applications and, therefore, could be used to help reduce radiation exposure, improve image quality and produce images in real time.

"The Project Clara virtual medical imaging supercomputer is a platform initiative, taking pieces of GPU containers that already exist, and all the CUDA libraries, meaning we can create imagine apps that rely on this technology," Powell said.

"It's [therefore] helping us reimagine medical imagining. We can now think about ‘super scaling' [because] it uses Kubernetes on GPUs to efficiently scale compute with demand."

Because Project Clara is virtual, it can run many computational instruments simultaneously. It's also remote, meaning it takes advantage of NVIDIA's vGPUs to enable multi-user access. Clara is universal, too, said Powell, so it can perform the computation for any instrument, whether CT, MRI, ultrasound, X-ray or mammography.

As part of the Project Clara announcement, NVIDIA unveiled it's working with a bunch of healthcare partners to push its AI tech in the medical imaging sphere.

One of these is AutoMap, from the MGH Martinos centre. AutoMap can shorten acquisition time of MRI and boosts image quality. Then there's V-Net, a company that can automatically measure anatomy and assess functionality.

"Cinematic rendering pioneered by Elliot Fishman at Johns Hopkins University brings a new level of quality, ultimately saving time for radiologists and improving patient outcomes," added Powell.

Another partner is Subtle Medical, who are working on dozens of applications in medical imaging, and recently won over more than a quarter of a million dollars in Nvidia's Inception program awards.

"New technologies are transforming healthcare," said the firm's founder, Dr Greg Zaharchuk, who is also a radiologist and associate professor in Radiology at Stanford.

"Nvidia's vision for a virtualised imaging supercomputer is an exciting new chapter that will revolutionise our ability to deliver AI-powered healthcare."

Because modern medical imaging applications demand new levels of computing, scale and accessibility, Nvidia believes Clara will help revolutionise medical imaging. µ

INQ Latest