Last year, I attended a rather odd conference in Berlin where 250 invited startups competed to win €1 Million ($1,157USD), with Steve Wozniak as one of the judges. The startups not only pitched, but exhibited their wares at stalls, and I visited many with interest, asking how they had fared with the pitching — bizarrely, more than a small minority admitted that they had pitched at the event for the first time ever and perhaps they should have practiced before the event!
The finalists were smart city waste solution providers Green City Solutions and population health management solution NavVis and the inaugural CUBE Challenge was won by US-based Enlitic, a digital health and life science startup. The company is developing AI technology with the goal to cover 95% of diagnostic radiology by 2020.
Machine learning in radiology can make image reading faster, more accurate, and provide a second set of eyes as well as triage. With radiologists given larger caseloads coupled with decreasing reimbursement, adding artificial intelligence software to the acquisition and interpretation phases can change the profession’s future for the better.
I attended CUBE Tech Berlin this year. It was a different venue with a different agenda — mostly prearranged networking meetings, presentations, and a smattering of pitching, sans competition. Enlitic's CEO Sally Daub returned to talk about how the company had fared since their win the year before.
She explained that the rationale underpinning their technology:
"Every year, over 300 million diagnostic radiology images are taken in the United States alone. As demand grows for diagnostic services, so does pressure on healthcare providers to operate more efficiently and accurately, at scale."
Enlitic’s technology can interpret a medical image in milliseconds — up to 10,000 times faster than the average radiologist, and their priority has been in detecting and diagnosing early stage lung cancer.
According to Sally, the company was first in the field with the most advanced technology: "We've demonstrated to our partners in blind tests that we can rapidly and accurately identify anomalies in medical images — the only company in the world with this capability."
Enlitic has been hard at work at a number of research initiatives since their creation, including a lung nodule detector and radiology data and testing. During a study with Australia's Capitol Health Ltd., Enlitic deployed a wrist fracture detection system using in-house deep learning models to circle fractures in X-rays, displaying these annotations in the PACS viewer for reading by radiologists. The study measured the accuracy and efficiency of three specialist radiologists, each reading a total of 400 studies, with and without assistance from Enlitic models. The study found that radiologists augmented by Enlitic were 21% faster, 11% more sensitive, and 9% more specific in their reading.
Since the previous year's win, the company has received a lot of inbounds and interest from around the world. This enabled them the opportunity to move forward with commercialization.
Sally was quick to stress that, "When we're using the technology in combination with doctors, the risk is very, very low, but the benefit to patients is very, very high, particularly in regard to second reviews or opinions, triage and pushing the imaging through to specialists, with the data analysis akin to, "If we had thousands of radiologists looking at that data."
"It changes the life of the patient and the life of the radiologists. Images are getting more complex, and radiologists are expected to do more in a shorter period of time. Doctors often only look at the top page of your record rather than a detailed case review." By comparison, use of AI in this instance means that radiologists can spend more quality time with patients.
Notably, Enlitic has achieved much since their 2014 founding, making them a compelling contestant in any pitching competition. But remember, if you have a 1 in 250 chance of winning €1m in a pitching competition — make sure you practice!
CAJ Price at posting:
33.5¢ Sentiment: Buy Disclosure: Held
(20min delay)
