Deep learning radiology startup Enlitic wins Aussie partner, $10M Series B

San Francisco-based Enlitic is working to apply the advances in deep learning in recent years to medical imaging. Now it's recruited the fastest-growing Australian radiology services provider, Capitol Health, to aid it in that quest with a partnership and by taking the lead on a $10 million Series B financing.

The news comes in the wake of a recent JAMA Internal Medicine study that found that computer software did not enhance the detection of breast cancer in mammograms. That dealt a blow to the logic behind the effort to treat medical imaging as a big data project that can be used to develop computer technology that will exceed the capabilities of an individual radiologist.

But Enlitic's project is designed to be much more sophisticated--applying the deep learning insights of the likes of Google and IBM's Watson, it's intended to mimic the ability of the human brain to recognize patterns in digital images but on a massive scale. The expectation is that the systematic application of such a system to an entire healthcare network will enable faster, more accurate and less expensive patient diagnoses.

In the partnership, Enlitic will apply its deep learning tools to help manage the workflow of hundreds of radiologists across the Capitol Health care network. It will work across all radiology modalities including ultrasound, x-ray, CT, MRI and PET imaging. The expectations for positive outcomes are running quite high.

"Enlitic will forever change the way radiology and all of medicine is practiced," said Capitol's Clinical Director Dr. Anthony Upton in a statement. "In the urban setting, it will improve accuracy and efficiency to deliver optimum patient outcomes. Globally, it will deliver healthcare to populations that have none."

Specifically, Enlitic's adapted deep learning can detect lung cancer nodules in CT chest images 50% more accurately than an expert panel of thoracic radiologists, according to the company. It also reduced the false negative rate. The vast majority of lung cancer patients aren't diagnosed until late stages; it's one of the hardest cancers to detect in medical images. If diagnosed early, patient survival is almost 10 times more likely, Enlitic noted.

In addition, in bone fractures Enlitic has been able to more accurately detect extremity fractures in locations such as the wrist. It was more than three times better than leading radiologists and even more accurate than that as compared to traditional computer vision approach analysis, according to the company.

"We believe that healthcare systems worldwide will embrace technology that delivers more effective results for patients," said Capitol Managing Director John Conidi. "This collaboration is a world-first--something ushering in a new age of data-driven medicine--with outstanding clinical outcomes and efficiency benefiting Capitol Health radiologists, referrers, patients and shareholders."

- here is the announcement