No Hype Here: 5 Concrete Examples of AI Transforming Medicine

We justifiably embrace artificial intelligence for everything from autonomous machines to smart cities to the mind-blowing evolution of today’s gaming and VR applications. Another area, however, that is witnessing the transformative impact of AI is life sciences and healthcare. While life sciences and healthcare are respectively unique domains each in their own right, they also enjoy a complementary relationship, with the scientific advances in the former bearing bountiful fruit in the latter.
To share greater insight of the momentous impact of AI and deep learning in healthcare — and of GPUs as the AI delivery mechanism — Advanced HPC in collaboration with our partner NVIDIA, is making available a very compelling IDC whitepaper entitled, From Bench to Bedside: Deep Learning’s Journey in Healthcare. Written by Cynthia Burghard, Research Director with IDC Health Insights, this whitepaper explores the application of deep learning in two key areas of healthcare:
- the augmentation or assistance to physicians and
- predicting the onset of an illness or adverse health event
To receive your free copy, please click: whitepaper.
The following five (5) examples are not merely visionary proclamations of the potential of AI and GPUs in the life sciences and healthcare marketplace, but moreover concrete illustrations of how together they are changing medicine dramatically in the here and now.
Case in point is Memorial Sloan Kettering Cancer Center (MSK). The Memorial Sloan team employs deep learning for tumor detection and segmentation by training high-capacity models and implementing the resulting systems in the clinic. Renowned MSK scientist, Thomas Fuchs, PhD, is using machine learning, a type of AI, to train computers to recognize cancer on digitized pathology slides. According to MSK’s website, the Center’s Department of Pathology is digitizing more than 40,000 pathology slides per month in order to facilitate the effort.
In another example, Google Brain is applying deep learning in TensorFlow, a popular machine learning framework, to tackle challenges in genetics and genomics. In late December of 2017, Google released DeepVariant, a tool that uses deep learning to identify the mutations that an individual inherits from their parents. In an article on the subject, Wired aptly characterized the trend by stating, “DeepVariant is just the front end of a much wider deployment; genomics is about to go deep learning. And once you go deep learning, you don’t go back.” (Megan Molteni, Wired; Dec. 2017.)
NVIDIA in concert with GE Healthcare serves as our third example. Although iterative reconstruction, defined as iterative algorithms used to reconstruct 2D and 3D images, reduces potentially harmful levels of radiation during x-ray examinations, its computation price tag makes it untenable in many healthcare settings. NVIDIA answered the bell with a cost-efficient and highly effective GPU solution that enabled GE Healthcare’s Revolution CT scanners to reduce radiation by an astounding 82 percent, while still achieving the high resolution that medical practitioners need in order to make accurate diagnoses.
NVIDIA GPUs provide another fitting illustration. NVIDIA GPU computing enabled life science researchers at Klaus Schulten’s computational biophysics lab at the University of Illinois at Urbana-Champaign to simulate molecular dynamics at a scale 1,000x larger than otherwise possible. This feat is not merely scientifically impressive, but truly groundbreaking when you consider that this advancement allowed the researchers to create never seen before views of an HIV capsid as well as the first ever simulation of an entire life form, that of the satellite tobacco mosaic virus. (NVIDIA Blog; April, 2018).
For our fifth example, we point to Arterys, a medical imaging software solutions company, that in February of 2018, received 510(k) clearance from the U.S. Food and Drug Administration (FDA) for the Arterys Oncology AI suite. The FDA clearance is significant for a host of reasons, not the least of which is that it validates how AI is measurably improving medical imaging accuracy and consistency. According to a company press release, “With this new technology, radiologists can now easily confirm, evaluate, quantify, and report on the absence or presence of lung nodules and liver lesions along with their key characteristics using a simple web browser . . . with accuracy equal to segmentations performed manually by experienced clinicians.”
The FDA approval of the Arterys suite to facilitate cardiology, lung and liver AI-powered workflows is indicative of the FDA’s heightened commitment to successful deployment and implementation of AI in healthcare, specifically with respect to ensuring that the agency’s regulatory framework and software validation tools are kept flexible and adaptive to AI advancements.
“We’re implementing a new approach to the review of artificial intelligence,” said FDA Commissioner Scott Gottlieb, MD. “AI holds enormous promise for the future of medicine, and we’re actively developing a new regulatory framework to promote innovation in this space and support the use of AI-based technologies,” continued Gottlieb. (FDA chief sees big things for AI in healthcare.)
What might be most interesting about these developments is that we are still in the nascent stages of artificial intelligence manifesting its influence in life sciences and healthcare. Given the burgeoning efforts of NVIDA with its Volta™ GPU platform and a litany of determined medical enterprises forging new paths in AI, emerging medical deep learning ventures may very well be even more transformative than their sentient machine-driven predecessors.
Once again, please remember to download your free copy of the IDC whitepaper, From Bench to Bedside: Deep Learning’s Journey in Healthcare.