Over 1950 Total Lots Up For Auction at Two Locations - NJ Cleansweep 06/13, NJ 06/14

Is 'deep learning' the holy grail of ultrasound innovation?

by Thomas Dworetzky, Contributing Reporter | August 10, 2015
Business Affairs Health IT Medical Devices Population Health Primary Care Ultrasound X-Ray
The major players in a large sector of the radiology market could find themselves disrupted — and possibly out of business — if a revolutionary biotech entrepreneur succeeds in producing the ultimate cheap, small, hand-held ultrasound scanner.

Jonathan Rothberg told MIT Technology Review that he is working on a patented device that will be “as cheap as a stethoscope” and “make doctors 100 times as effective.”

“We want it to work like ‘panorama’ on an iPhone,” Rothberg told the publication. “When I have thousands of these images, I think it will become faster than a human in saying ‘Does this kid have Down syndrome, or a cleft lip?’ And when people are pressed for time it will be superhuman."

He has raised $100 million so far through his Butterfly Network venture to create his smartphone-sized scanner that can be held to a patient's chest to produce a clear, 3-D image.

If he succeeds, it will do to the scanning market what is now happening with the clinical laboratory market, where millions of dollars are going into entrepreneurial projects to make miniature lab tests that are quick, cheap and accurate.

The device, known as a “capacitive micro-machined ultrasound transducer,” or a CMUT, will appear to be a "window", looking at what is inside a person when it is held up to a part of the body, according to Technology Review.

The key to this potentially revolutionary tool is big data. Rothberg revealed to Wired that it will employ so-called “deep learning”, the artificial intelligence approach behind Siri and similar iterative learning systems. They get smarter the more they are used by many people, as it builds their database and allows sophisticated analysis across a huge pool of clinical data.

The device will “transform medical imaging and non-invasive surgery by leveraging advances in semiconductors, deep learning and cloud computing,” Butterfly Network stated in a press release.

The point of this effort, however, is not to put radiologists out of business. "Spending 18 percent of our GDP, or $10,000 per person, per year, on health care is unsustainable and out of touch with the needs of the rest of the world," Rothberg stated.

"Our mission is to democratize health care by launching companies, building devices, and combining advances in biology, semiconductors, and artificial intelligence. We built Butterfly to transform the way we image the body and perform surgery," he said.

Back to HCB News

You Must Be Logged In To Post A Comment