Big Data, Analytics, and AI in Retina Practice

New tools are harnessing data to increase efficiency, enhance workflow, and ultimately improve patient outcomes.


Retina specialists make critical treatment decisions every day, drawing from a combination of professional training and expertise, personal clinical experience, and whatever valuable information may be at hand. Although each year there is exponential growth in information with the potential to influence such decisions, much of it is unusable, or at least beyond the convenient reach of busy individual practitioners.

Harnessing data to make it usable almost always requires the linkage or large-scale integration of complex information from a range of disparate data sources. An enormous amount of sifting and searching, comparing and inferencing, and other forms of advanced computation are required to develop a tool that is of value to the retina specialist — that is, one that ultimately increases the accuracy and efficiency with which a specialist can make a diagnosis or clinical evaluation, refine or hone a treatment strategy, forecast risks, or perform a delicate procedure.

Here we look at a sampling of innovative data-driven technologies in various forms and stages of development. Each has demonstrated the potential to be of benefit to the retina specialist in one or more ways, which range from directly improving patient care, to facilitating the testing of new research hypotheses, to delivering insights based on the ongoing surveillance, and detailed analysis, of various patient populations.

From Inspiration to Execution

Stories behind the development of data tools designed for retina specialists can be intriguing, particularly when they borrow from sophisticated technology that has been used to solve problems in other fields or industries. A helmet created for use by fighter pilots, for example, was the inspiration behind the Beyeonics One system, a surgical visualization helmet that is in the final stages of development. The helmet designed for fighter pilots enabled pilots to find all relevant data, sorted according to each pilot’s needs, while looking through the cockpit window. Before using the helmet, pilots had been required to look down to obtain information from a control panel, before lifting their gaze to identify targets with the naked eye, all the while maneuvering the plane to aim missiles.

“Just as with the pilot in the cockpit, Beyeonics vision places the surgeon in the center of the OR, and displays all the information the surgeon needs in one place, at the relevant time,” says Anat Loewenstein, MD, director of the Department of Ophthalmology at the Tel Aviv Medical Center in Israel. “When I’m seated in my chair and wearing the helmet, I can turn my head in any direction and still see the high-resolution 3D image of relevant clinical data for each patient. Even when I move my head during surgery to talk to the nurse, or when I make unwanted movements, the image is there with the same sharpness. At the same time, I can still see the OR and my hands throughout the surgery.”

Dr. Loewenstein adds that the Beyeonics system retains high resolution in both low and high magnification at the edge of the field of view, which has been challenging or impossible with traditional optics. With head movements alone, the surgeon can focus or zoom in on patient data — such as preoperative or intraoperative OCT or fluorescein angiography (FA) — that is overlaid and presented in a real-time virtual display.

“The main goal of digital visualization, in my opinion, is not a cool picture or better ergonomics, but better patient outcomes by means of improved informatics,” says Pravin U. Dugel, MD, of Retinal Consultants of Arizona, in Phoenix, Arizona. “We have a lot of information in our clinics, including FA, photos, OCT, and microperimetry. However, currently, when we go to the OR, we have nothing. This is intuitively wrong. The fact that this company is spun out from the military, with digital visualization and informatics for fighter pilots, means that it has the finest such engineering expertise. I am glad to see that this technology is being translated and adapted to the benefit of our patients.”

The inspiration for another data-driven technology designed for surgeons and surgical facilities was the growing pressure that ophthalmologists are under to increase their efficiency, improve their workflow, and cut costs. The Bausch + Lomb Eyetelligence application for Stellaris Elite vision enhancement machines enables surgeons to mine data that has been communicated from more than 50,000 surgical procedures, by more than 250 Stellaris Elite systems, to arrive at adjustments, based on surgical trends and usage patterns, that aim to improve efficiency during surgical procedures.

“With the Eyetelligence platform, we will have the ability to analyze in large numbers the efficiencies and differences of different port sizes and port designs,” says Sunil Srivastava, MD, a retina specialist at the Cleveland Clinic in Cleveland, Ohio. “With this information, we may be able to identify ideal settings for specific pathologies and obtain real-time feedback for surgeons. In reviewing the data from Bausch’s previous Stellaris platform, from more than 100,000 vitrectomy cases, we found that higher-volume surgeons tended to be more efficient than those with lower volumes. Additionally, the relative efficiencies of 25-gauge and 23-gauge vitrectomy in comparison to 20-gauge was confirmed.”

With Eyetelligence, surgeons can sync their surgical settings across all of the Stellaris Elite machines on which they operate, regardless of location, so that they don’t have to spend time personalizing settings on each system. Surgeons also have a direct connection to Bausch + Lomb for fast access to technical support, using a drop-down menu to specify the type of problem they might be having, and granting Bausch + Lomb remote access to the surgeon’s machine when necessary. New features for feedback and surgical planning are in development.

The goal of creating an adjunct to the surgical microscope that would increase depth of field, enable higher magnification, and provide a convenient real-time display of key data — such as intraocular pressure, cut speed, infusion pressure, and flow rates — was the inspiration behind Alcon’s 3D Ngenuity Visualization System. The system uses a 3D stereoscopic, high-definition digital video camera and workstation to offer magnified stereoscopic images of the surgical field, and it is equipped with digital filters that can be used to customize the surgeon’s view, highlight ocular structures and tissue layers, and adjust light exposure to the patient’s retina. The Ngenuity system allows retinal surgeons to operate while looking at a high-definition 3D screen, instead of bending their necks to look through the eye-piece of a microscope — a feature that is particularly valuable during lengthy vitrectomy surgeries.

“This system allows me to have a more relaxed and ergonomic posture,” says John Kitchens, MD, of the Retina Associates of Kentucky, in Lexington, Kentucky. “I am no longer ‘locked’ into a rigid posture with my neck extended. As quickly as my first day, I noticed considerably less neck fatigue. Instead of being hunched or sitting upright, I am actually able to lean back a bit more in my chair. My posture with Ngenuity has enhanced my ability to do maneuvers such as buckling with the microscope and maneuvering equipment in and out of the cannulas.”

The high-definition screen of the Ngenuity system lets retinal surgeons see 3D images of the back of the eye, with much greater depth and detail than traditional microscopes. Surgeons also may increase magnification while maintaining a wide field of view, as well as use digital filters to customize their view during each procedure. The Datafusion software also allows surgeons to see all the data from their Constellation Visualization System on the Ngenuity screen, in real-time, during surgery — for example, surgeons can track intraocular pressure, flow rates, infusion pressure, and laser power.

A Focus on Analysis

When OCTA technology first became available, it provided only qualitative data — most notably, high-resolution imaging of retinal blood vessels. Retinal specialists undertook the task of interpreting the data, and resulting interpretations were somewhat subjective. Optovue’s Angioanalytics software now offers visit-to-visit change analysis and trend analysis so that in addition to objectively monitoring retinal and vascular changes caused by disease progression, specialists can estimate future change in a patient’s condition based on data from previous visits. The flow area measurement tool also can be used by retina specialists to track changes in choroidal neovascularization (CNV) size, to measure anti-VEGF effectiveness.

“My personal evaluation of Angioanalytics suggests that we now have another way to monitor or determine if additional treatment is necessary by documenting how a drug affects a specific CNV membrane in a specific patient, and how long that effect lasts,” says Paul Tornambe, MD, of Retina Consultants San Diego, in San Diego, California.

When OCTA is used with Angioanalytics, the technologies combine to create color-encoded maps of the vessel densities of the retina or optic nerve, and analyses of areas where there is nonperfusion or abnormal blood vessel growth. Retina specialists can quantify areas of choroidal neovascularization, ischemia, the foveal avascular zone (area, perimeter, and vessel density within a 300-micron radius of the foveal avascular zone) and vessel density of the superficial vascular complex, deep vascular complex, and radial peripapillary capillaries (calculated as percent of total area occupied by OCTA-detected vasculature). Together these tools quantify parameters that are relevant in diabetic eye disease and allow precise visit-to-visit comparisons of changes in these parameters. Such information enables specialists to tailor treatment accordingly — for example, because the extent of ischemia correlates with risk of progression to severe or proliferative diabetic eye disease, specialists may recommend anti-VEGF treatment sooner to reduce the risk of progression, or may follow patients more frequently to monitor changes.

Improving Diagnosis and Streamlining Referrals

At least one data tool developed in recent years has proven to have surprising indirect value for retina specialists. IDx-DR, which is used by primary care practitioners, rather than retina specialists, enables specialists to spend less time screening and more time treating patients who have actionable disease.

“We know that patient outcomes are improved through early diagnosis of diabetic retinopathy, yet it’s still not happening — only about 15% of Medicare patients with diabetes get the required retinal exam,” says Michael Abramoff, MD, PhD, retina specialist and founder and CEO of IDx. “Our goal is to increase patient access to DR testing by placing an autonomous AI diagnostic system in primary care. This allows a clinical decision in minutes with human oversight, ensuring that patients with DR are caught in the early stages. It also allows for care coordination, so that if patients are found to have DR, they actually get to visit a retinal physician to be treated.”

With IDx-DR, patient referral pathways are expected to change dramatically, because ophthalmologists can now partner with primary care clinics to use IDx-DR to detect diabetic retinopathy, increasing the odds that disease will be caught early. IDx-DR also improves access to screening, by identifying diabetic retinopathy in patients who are not currently seeing an eye care specialist, including individuals with social or economic barriers to eye care, who often do not see specialists until symptoms of vision loss have emerged.

“It was important to me that this autonomous AI system be safe, effective, and equitable,” adds Dr. Abramoff. “It is also more accurate than I am as a retina specialist. The FDA agreed, and granted IDx-DR the first-ever clearance of an autonomous AI tool in any field of medicine in 2018.”

Harnessing the Power of EHRs

In the current age of big data and artificial intelligence, many data experts believe that electronic health records (EHRs) are the single most valuable source of relatively raw health care data available.1 When EHR systems first emerged, they were designed primarily to be of value to CMS and to help meet government regulations, but they did not necessarily offer valuable information to physicians. As newer systems emerged, they gradually became more adept at capturing patient demographics, as well as practice and treatment details, but remained fragmented and yielded data with limited direct application to patient care. During the past decade, however, several newer EHR-based systems have begun to offer significant value to retina specialists.

In 2010, Modernizing Medicine was born out of a chance encounter between tech entrepreneur Daniel Cane and a practicing dermatologist, Michael Sherling, MD, during a routine skin exam. Both experts recognized important unmet needs when it came to available technology for documenting patient visits and streamlining physicians’ workflow. After determining that they could solve important problems in part by teaching practicing physicians how to code, they collaborated on the development of their first EHR system, EMA for dermatology. Since then, they have developed systems for several medical specialties, including ophthalmology.

“Modernizing Medicine’s cofounders had a vision to improve electronic health record systems — one that could save doctors time while collecting structured data at the point-of-care to benefit physicians and healthcare as a whole,” says Michael Rivers, MD, director of EMA Ophthalmology, Modernizing Medicine. “Practicing physicians were hired to work alongside software developers to program medical knowledge into the software, as physicians know best when it comes to what their peers need.”

Dr. Rivers notes that the EMA ophthalmology system comes with built-in retina knowledge, such as retina-specific procedure workflows like office injection and treatment plans. The integrated system allows providers to analyze their own data for valuable insights — for example, it allows them to view outcomes data, by treatment plan, for patients with chronic diabetic macular edema or macular edema secondary to wet AMD, to determine rates of relapse with different medications. It also helps to track patient visits and identifies patients who have missed appointments.

“Having the flexibility and autonomy to analyze your own data in-house provides retina specialists with the ability to identify gaps in treatment and support decision making,” says Dr. Rivers. “This is an analytics solution that integrates with your EHR and practice management technology and makes generating valuable insights and reports virtually seamless for providers. In addition, EMA has built-in MIPS clinical data registries, which are important for certain MIPS measures.”

Modernizing Medicine’s ultimate goal is to empower physicians and staff to make data-backed decisions in all aspects of their practice. Currently, the company is focused on enhancing its analytics capabilities and further developing its patient engagement tools and ambulatory surgical center offerings.

In 2012, Vestrum Health, which provides EHR data analytics for retina practices, was founded by Nick Anderson, MD; John Pollack, MD; and David Williams, MD.

“The idea for the company emerged from our understanding that most data generated by retina specialists was being collected and sold through multiple channels including pharmacies, insurance companies, product manufacturers, drug distributors, and even some EHR companies — without any benefit flowing back to physicians, despite the fact that physicians are doing a great deal of work to create and document this data,” says John Pollack, MD. “We started Vestrum Health to capture the value of physician data for the benefit of retina specialists and, ultimately, their patients.”

Vestrum Health uses proprietary algorithms to provide retina specialists with immediate real-world insights, as well as other services driven by retina specialist requests — such as monthly customized practice performance reports, which can be used to improve practice business efficiency, obtain longitudinal assessments of patient outcomes, and monitor trends in referral patterns.

“Retina specialists today are hungry for information on how their practice compares to others,” says Dr. Pollack. “Vestrum has responded by providing them with monthly benchmarking reports that compare the practice’s key performance indicators with those of other practices, on both regional and national scales. Retina specialists are also increasingly taking advantage of Vestrum’s free clinical trial recruitment services to accelerate the process of accurately identifying patients who meet specific inclusion and exclusion criteria for any particular clinical trial, thus maximizing a practice’s involvement in trials as well as the cost-effectiveness of the trial recruitment process.”

Dr. Pollack adds that practices can access Vestrum Health’s database-centric clinical research support services at no cost. And, in response to demand from retina specialists, Vestrum is currently developing a range of new financial analytic services focused on billing validation, time to payment, and other management insights.

In 2017, Verana Health was formed through a partnership with the American Academy of Ophthalmology (AAO). Mark Blumenkranz, MD, of the Byers Eye Institute at Stanford University, is cofounder and sits on Verana Health’s board of directors alongside fellow ophthalmologists David W. Parke, MD, CEO of AAO, and Steven Schwartz, MD, professor of ophthalmology at the Stein Eye Institute, UCLA. Drawing from AAO’s IRIS Registry, a database with billions of data points from 60 million patients and more than 15,000 ophthalmologists across the United States, Verana Health curates and applies analytics to this clinical data to provide real-world insights that aim to improve patient care.

“Something unique and industry-altering happened 5 years ago with the creation of the IRIS Registry,” says Verana Health’s CEO Miki Kapoor. “Never before has there been a specialty clinical database of this magnitude. The AAO and Dr. Parke were truly forward-thinking in their vision for how this incredible clinical database could be used to derive real-world evidence and enable us to rethink how we do research in medicine today.”

Verana Health harnesses comprehensive EHR data found in the IRIS registry to generate real-world evidence that informs patient care and aims to accelerate medical innovation. The company also provides services that are of particular value to retina specialists — for example, Verana Health has developed a software for trial investigators that facilitates patient recruitment in clinical trials.

“Verana Health’s physician solutions are designed to enable physicians to improve outcomes by enhancing patient care,” says Kapoor. “For retina specialists involved in clinical trials, Verana Trial Connect enables the physician and his or her staff to quickly identify which patients have the highest potential for study recruitment. The solution harnesses information from the trial’s protocol to identify potentially eligible patients, provides helpful next steps in the recruitment process, and tracks patient recruitment status. Trial Connect is particularly useful in the case of rare retinal disease trials for which recruitment has been challenging. With this innovative approach to clinical trial recruitment, the AAO and Verana Health hope to enable physician members to more rapidly bring new therapies to patients who need them most.”

Verana Health and the AAO recently introduced Verana Practice Insights, which enables IRIS Registry participants to understand and benchmark their own clinical care patterns relative to a cohort of their peers. It aggregates population trends in patient diagnoses and interventions to support the adoption of best practices. While it is currently presenting trends related to cataract surgery, in 2020, its capabilities will expand to include retinal conditions.


  1. Panch T, Mattie H, Celi LA. The "inconvenient truth" about AI in healthcare. NPJ Digit Med. 2019;2:77.