
Decades after embracing the potential of AI to diagnose eye disease, University of Iowa retina specialist and IDx President Dr. Michael Abramoff may soon get his bite of the apple. PHOTO DAVE DEWITTE
By Dave DeWitte
[email protected]
The clinical trial results from IDx’s first product candidate aren’t out, yet optimism is high that its IDx-DR device could soon be saving the vision of thousands of Americans each year who suffer from diabetes.
The University of Iowa Office of Research and Economic Development announced Feb. 5 that the Food and Drug Administration (FDA) granted the UI spinout’s IDx-DR a “breakthrough device” designation that will result in an expedited review of the product’s safety and efficacy. The designation is restricted to technologies that provide for more effective treatment or diagnosis of life-threatening diseases and those that cause irreversible damage.
“It’s the first time for us,” said Marie Kerbeshian, director of the UI Research Foundation, which licensed the intellectual property the IDx device uses from UI faculty research. Under the expedited process, the team at IDx in Iowa City and the UI could know in as little as six months if IDx-DR is cleared to go to market.
The system uses artificial intelligence to automate detection of diabetic retinopathy, which causes blindness in some 24,000 Americans each year. With about four hours of training, it can be used by operators with no post-high school education, enabling them to tell a patient within minutes if they have early indications of diabetic retinopathy that require treatment by an eye specialist.
Early diagnosis is critical in diabetic retinopathy, according to IDx President Dr. Michael Abramoff, a retina specialist, because the disease can be managed if detected early, potentially preserving a patient’s eyesight. If the patient goes undiagnosed until they experience signs of vision loss, he said, the damage is already irreversible.
Diabetes is a growing medical problem that afflicts 9.4 percent of Americans, according to the American Diabetes Association, causing an assortment of medical problems from kidney disease to amputations. Diabetes is typically managed through consultation with primary care physicians, who typically refer their patients to opthamologists for monitoring of the potential for eye disease complications because they lack the specialized equipment and knowledge.
Blindness occurs too frequently, Dr. Abramoff said, because 30-50 percent of diabetic patients don’t make it to their eye doctor appointments – often due to a combination of travel and time requirements, and even their own physical limitations.
“Everything is managed well [for diabetics] in primary care except this retinal eye exam,” Dr. Abramoff said. “Between 20,000 and 25,000 people a year go blind – needlessly – from diabetic retinopathy.”
The FDA’s breakthrough device designation set off a chain reaction of events at IDx, which is led by former telecommunications executive Gary Seamans. The most noticeable one has been the hiring of more staff so that IDx can quickly move to market once it receives the green light from the FDA. The company, with 25 employees, is poised to expand from its modest Iowa City headquarters off Highway 1 into a larger location.
“We’re excited and we think the FDA’s excited,” Dr. Abramoff said. “Hopefully, the FDA being excited means they like what they see. We have no way of knowing how long it’s going to take.”
The stakes are high and personal for Dr. Abramoff, a physician-scientist at the UI who became intrigued two decades ago about the potential for computer algorithms to detect eye disease. The Dutch-born doctor authored more than 300 publications, received 10 patents and applied for six others in the field, all while continuing to treat patients at UI Health Care.
Clinical trials involving some 900 patients at primary care clinics scattered across the United States were completed last July. Results of the system’s findings on patients were compared to evaluations of their retinas by a leading international eye institute using high-quality retinal images taken with multiple imaging systems.
“We described how we built the algorithm, how we did the clinical trial, how we never interfered with the clinical analysis, how we got the independent analysis,” Dr. Abramoff said.
IDx-DR typically flags “abnormal” findings in about 10-12 percent of patients screened, who are referred for treatment by eye specialists. The algorithm was developed with a margin of safety which could mean some false positives, but Dr. Abramoff says he’s confident the system is more accurate than he, or any other individual eye doctor, can be in detecting the disease.
When word spread of Dr. Abramoff’s plans to automate diagnosis years ago, he was dubbed “the retinator” by some of his peers.
“It’s from the movie, ‘The Terminator,’” he said. “It’s a bit facetious, but it’s also serious, because am I destroying opthamologists jobs?”
Dr. Abramoff sees it differently. He says diabetic patients who receive a diabetic retinopathy diagnosis using IDx-DR will no longer skip their appointments with eye specialists, and will give those opthalmologists the opportunity to provide the treatment needed.
Still, he understands that the use of automation for such an important diagnosis can be an emotionally charged issue.
“There’s a lot riding on this,” Dr. Abramoff said. “People are excited about putting AI [artificial intelligence] in health care, but nobody has done it – at least where it matters.”
While receiving the results of the clinical trials was a long-anticipated moment for Dr. Abramoff and his team, he said it was ultimately anticlimactic because he realized how much more work had to be done to get the technology into places where it can help patients. His next big moment will come when he’s finally allowed to talk about those results. That will occur on Feb. 22, when he presents the findings to some of the most preeminent scholars in retinal disease at the annual meeting of the Macula Society in Beverly Hills.
A successful market launch of IDx-DR could create momentum for IDx’s broader plans to use AI to detect one of the most common eye diseases, glaucoma, and a range of other diseases affecting other bodily organs.
Ms. Kerbeshian says UI faculty were able to develop the algorithms in part because they had access to decades of patient data from the UI health system that they could use to predict the future of patients from observed conditions.
While companies like IDx could make their founders wealthy, Ms. Kerbeshian said it’s often not the financial potential that drives UI medical faculty members to start companies, but a desire to be more impactful than they can be in treating patients one at a time.
“Through your IP [intellectual property] research, if that extends to becoming a standard of care, it can have such an impact all across the country and the world,” she said.
With the support of strong angel investors, Dr. Abramoff said IDx is fully funded to support the launch of IDx-DR into the market. IDx has also been in conversation with distribution partners such as IBM, which is already working with the company to introduce IDx-DR in Europe.
The company has also begun conversations with investment banks about the possibility of going public to help fund its further development and launch of additional products, he added.