Nvidia has been involved with the healthcare sector for more than 10 years, and in the last five years has been working with the research and startup communities in the areas of early detection and improved workflows.

One result is the ability to virtualise existing medical imaging instruments and have the data processed remotely. This means that a 10-year-old scanner, for example, can be used with the latest 3D rendering or AI tools.

In general, earlier and better detection of conditions leads to better outcomes, observed Nvidia founder and chief executive Jensen Huang.

Most ultrasound, MRI, PET and other images look much the same today as they did fifteen years ago, he said, but "we can now reconstruct images better than ever before", for example as lifelike 3D renderings.

Other advantages of modern processing methods is that they allow the required image quality to be achieved with lower doses of radiation and contrast dye.

So Nvidia is working on Project Clara, aimed at developing a "medical imaging supercomputer" that can "virtually upgrade every single medical instrument".

Huang showed a demonstration that took a series of 2D ultrasound images of a heart, reconstructed it in 3D, used deep learning to segment out one of the ventricles, and then displayed it as a moving image along with the calculated volumes and ejection fraction.

Nvidia is collaborating with various hospitals and medical equipment suppliers, he said. These include Massachusetts General Hospital and GE, Powell observed.

"You'll see more of Clara throughout the year," she promised. "It's really a virtualised data centre" that could be in the cloud or on premises.

Nvidia will be working with equipment vendors to obtain the necessary certifications, she added.

One of the problems in medicine is there are only a finite number of radiologists, and some geographical areas are under-served. Bringing technology to bear can make their skills more widely accessible. Furthermore, GPUs are cheap and can be used in the cloud.

"AI has a real role to play in terms of screening," said Powell, suggesting this is an area particularly affected by a lack of radiologists. "Early detection is important, so where can we screen?"

Similarly, genomics is becoming affordable, but adds to the data requiring analysis.

Modern medicine is collecting increasing amounts of digital data about humans, and working out how to integrate and analyse this data will usher in the era of precision medicine, which will transform the pharmaceutical industry and healthcare more generally.

But doing that will require "new computing paradigms," Powell said.

And that, presumably, is where Clara will come in.

Disclosure: The writer attended Nvidia's GPU Technology Conference as a guest of the company.