Flesh and blood. Bones and guts. As humans we feel nothing more viscerally — in the most literal sense — than our health.
That makes this year’s gathering of the Medical Image Computing and Computer Assisted Interventions Society — MICCAI 2017 — in Quebec City, Canada, one of the best ways to understand how deep learning is improving the lives of people all around us.
The conference brings together leading biomedical scientists, engineers and clinicians to talk about new technologies in medical imaging and computer-assisted intervention, providing an early look at trends poised to sweep through the $6.5 trillion healthcare industry.
This will be the group’s biggest conference yet, with 1,300 attendees. And deep learning — which pairs vast quantities of data with sophisticated neural networks to give computers amazing new capabilities — deserves a lot of the credit, organizers say.
“That’s in no small part due to the increased interest in AI among the health research community,” says MICCAI board member Dr. Simon Duchesene, an associate professor in the Department of Radiology at Université Laval.
AI-Focused Health Research Soars
One sure sign of this: the number of AI-focused papers submitted by the health research community is surging. Of the 800 manuscripts submitted at MICCAI, 60 percent are focused on machine learning. And of those papers, 80 percent use deep learning.
Deep learning is unleashing ideas so futuristic they seem inspired by science fiction. One paper, for example, explores how deep learning can analyze images to help robots perform minimally invasive surgery.
“The advancements we’re seeing now are truly amazing,” Duchesne says. “Thanks to AI, the operating room of the future will soon be embedded into hospitals worldwide.”
Healthcare Startups Booming
A surge in startup activity is another sign of how fast deep learning is being adopted by the healthcare industry. The number of AI and deep learning healthcare startups has grown more than 160 percent in the last five years, analysts estimate.
Deep learning in healthcare is the leading industrial application of AI, according to venture capital tracker CB Insights, raising $1.8 billion across 270 deals since 2012. Just this year, these startups have raised over $132 million across 22 deals.
Startup Arterys — which will be showcasing its technology at MICCAI 2017 — taps into cloud computation and deep learning to help physicians to measure blood flow through the heart’s ventricles. It’s a process that usually takes 45 minutes. Arterys does it in 15 seconds. Theirs was the first cloud-based deep learning application to get FDA clearance to use machine learning in a clinical setting.
Another startup at this week’s conference: Montreal-based Imagia. The AI biotech company is developing imaging biomarkers based on routine medical data to predict cancer patient outcomes and help personalize care. Imagia’s solutions include clinical decision support for image-guided procedures and optimizing drug development clinical trials.
The Sky’s the Limit for AI-Powered Healthcare
Such startups are attacking some of the world’s fastest-growing markets. The annual revenue for medical image analysis in healthcare alone will increase to $1.523 billion worldwide in 2025 from less than $100,000 last year, according to market research firm Tractica.
To learn more about how deep learning and healthcare intersect, watch my recent TedX talk, below.