The Startup That’s Bringing AI to Ultrasounds and MRIs

Butterfly Network

That massive MRI machine at the hospital that can look for damage inside your knee? It’s not just enormous. It’s enormously complex and expensive. And operating the thing takes not only time but expertise. The same goes for the ultrasound machine that a provides a look at your unborn baby. But entrepreneur Jonathan Rothberg says it doesn’t have to be this way.

Rothberg just announced $100 million in funding for his three-year-old startup, Butterfly Network, that hopes to create a new handheld medical-imaging device that can make both MRI and ultrasounds significantly cheaper and more efficient. The aim is even to automate much of the medical imaging process, and in a little over a year, if all goes according to plan, the device could be ready for deployment in clinics, retail pharmacies, and in poorer regions of the world.

Jonathan Rothberg. Butterfly Network

It’s not an altogether unexpected endeavor from Rothberg, who has a wealth of experience in the biotech and semiconductor industry. He’s already launched five health startups, two of which—454 and Ion Torrent Systems—turned out to be successful DNA-sequencing companies that he sold for more than $500 million. The latter, Ion Torrent, crammed an entire expensive DNA sequencing machine into a single, cheap semiconductor chip. Similarly, with Butterfly, Rothberg says, he “took a step back and looked at bottlenecks in the healthcare system more broadly.”

Butterfly’s handheld tool, Rothberg says, will walk users through the medical imaging process using on-screen instructions akin to Apple’s Panorama snapshot tool, and it will use ultrasound scanners to create 3D images in real-time. It will then send these to a cloud service, which will work to zoom in on certain identifying characteristics in the images and help automate diagnoses. As Rothberg explains it, the service could look at an ultrasound and notify doctors that a baby has Down syndrome or a cleft lip.

This service will make use of “deep learning,” the same kind of artificial intelligence that’s starting to remake image and voice recognition inside web giants like Google, Facebook, and Microsoft. The more the imaging tool is used, according to Rothberg, the smarter it gets. The idea is to make using the device so easy that any technician or nurse practitioner can take advantage of it—without specific training in the device.

Rothberg also aims to integrate telemedicine into the service, so that a specialist in a remote location can weigh in on the images that the instrument records. That way, he says, there’s potential to use the device in places in the world where there are fewer resources and a lack of people with technical knowledge, like radiologists. So far, Butterfly Networks has provided proof-of-concept that the ultrasound scanner can work, but the timeline to get an actual prototype built is 18 months. And the device represents only the beginning of Rothberg’s aspirations to disrupt healthcare.

Butterfly Network is the first of the ventures to emerge from 4Combinator, a startup incubator launched by Rothberg with the goal to create and nurture companies that look for weak areas in the healthcare system and provide cheap and accessible solutions. Three other companies under 4Combinator have received between $5 million to $20 million in seed funding, and they’re exploring projects related to new treatments in tubular sclerosis and coming up with innovative ways to manufacture drugs.

“This may be an inflection point with our understanding of human biology,” says Rothberg. “In the next ten years, we’ll see diagnostics as well as medicine change. And that will come from both medicine and devices, and the information systems that support them.”