Melanie Cole, MS (Host): Welcome to Better Edge, a Northwestern Medicine podcast for physicians. I'm your host, Melanie Cole. And joining me today is Dr. Paul Bryar. He's a Professor of Ophthalmology and Pathology at Northwestern Medicine, and he's here to highlight diabetic retinopathy screening for us today.

Melanie Cole, MS: Dr. Bryar, thank you so much for being with us. I'd like you to start by giving us a brief overview of diabetic retinopathy, the key pathophysiological mechanisms that underlie the development and progression of this condition.

Paul Bryar, MD: So, diabetic retinopathy is a result of the microvascular complications of diabetes. Specifically, you know, the blood vessels in the retina undergo changes over time in a certain percentage of patients with diabetes. And retinopathy can occur when those small capillaries can leak and cause edema or swelling. It can also progress to what we call proliferative diabetic retinopathy where, because of the decreased oxygen supply to the retina, there's an attempt of the eye to grow new vessels in the retina, which end up doing more harm than good. So, you know, those are the two things that we want to be able to screen for, you know, diabetic retinopathy and diabetic macular edema. And the key with that is, is that if we can find it early enough, you know, 90% of all vision loss related to diabetic retinopathy and macular edema is preventable if caught in time.

Melanie Cole, MS: Well, thank you for that. So, provide an overview for us, Dr. Bryar, if you would, of Northwestern Medicine's diabetic retinopathy screening program. What motivated you to set up this program?

Paul Bryar, MD: Well, the recommendation is, in general for each person with diabetes, to have an annual screening eye exam for diabetic retinopathy and other diabetic eye disease. Nationwide, about 60-65% of patients with diabetes actually get that exam. And that can be much lower, especially in patients where there's healthcare disparities in terms of minorities and patients with decreased access to medical care. That gap, that overall 35-40% of people who have not been screened, that's who we're targeting with this Diabetic Retinopathy Screening Program. We're trying not to replace an eye exam, but to really target those people who have not been screened.

Melanie Cole, MS: Well then, how does this approach at Northwestern Medicine differ from traditional screening methods? What had previously been done and what are you doing now?

Paul Bryar, MD: Traditional screening programs really centered around just trying to increase awareness and education, education to providers, the importance of the eye exam, education to patients that, you know, even if they're seeing well, they need to get an eye exam, and also the other efforts around the electronic health records helping us by practice alerts saying that this patient has been flagged as having diabetes but not having an eye exam in the last year. So, those have been more traditional methods of trying to increase that screening. Those methods have had varying success. So, what we wanted to do is put the emphasis on screening at the point of care in the primary care setting or in endocrinology setting, to give those providers the tools to become aware of the need for screening, and in many cases, do the screening either onsite or just a short time away as a point-of-care test that can be done without an advanced appointment.

Melanie Cole, MS: That's excellent. Dr. Bryar, the screening program was recently expanded to federally qualified health centers in Chicago. I'd love for you to tell us about this expansion and its impact on patients.

Paul Bryar, MD: As I was alluding to earlier, you know, certain populations, that screening rate of patients, with diabetes is low, you know, as low as 35% in certain healthcare settings, minority and lower socioeconomic status, so we really wanted to go into the federally qualified health centers. And we were able to provide them with a retinal camera, which can be used to screen the patients, collaborate with them to create practice alerts, to alert them at the point-of-care and have the images, the retinal images, sent here to Northwestern. We grade them and, with the short turnaround time, give them a final report with a recommendation plan, such as we photograph in a year or refer to a comprehensive ophthalmologist or a retina specialist. So, we've been able to provide these services for them at no cost, and to get the program up and running. That's had tremendous impact on many people because when we do screening exams, probably about one third of all patients have a recommendation of a referral for one reason or another, whether it's for diabetic retinopathy or any other potential eye problem that's detected on the retinal pictures, and that could be glaucoma, macular degeneration or a host of other diseases.

Melanie Cole, MS: Dr. Bryar, standard of care for patients with DM is an annual eye exam or an assessment of retinal imaging via telemedicine. One significant issue with the latter is that presence of poor quality imaging that prevent that adequate assessment. And we saw that during COVID specifically as well. Tell us how AI has been used to address the issue of finding poor quality images.

Paul Bryar, MD: Taking a little step back, how we do the screening is we can do what's called point-of-care screening with a device called a non-mydriatic retinal imaging system. And essentially, it's very similar to what you look into when you get your driver's license vision exam. You know, you just put your chin in there. And these units are very sophisticated. The older units, they would require somebody to manually focus to find the pupil, focus on the retina, and then take an image. Right now, all we really need to do is put the patient on the chin rest. We press start and these units find the pupil, autofocus on the retina, and really the only job of the camera operator is to tell the patient to hold from blinking for about two seconds, and that's it.

Even with the more advanced cameras, non-gradable images are an issue. Those can be for a number of reasons. The patient could have blinked, the retina could be out of focus, there could be a media opacity in the way such as a cataract that's blocking our view of the retina. So, all of these various things can cause what we call a non-gradable image. And oftentimes, we wouldn't be able to learn about that until the patient has left, the image is sent, you know, electronically to us to grade, and it's a blurred or ungradeable image. So, that patient then has to be notified to come back to take a picture or to just have an in-person exam.

One of the ways we're working with AI is to analyze the image at the time the photo is taken and the computer can basically give a recommendation, that this is gradable or ungradable. And if it's ungradable right away, the camera operator can just dim the lights, reposition the patient, do whatever needs to be done to take it again. So, we developed an algorithm here which we're refining and, you know, with the intent of pushing it out to these cameras to allow the operators right then and there to retake an ungradable image.

Melanie Cole, MS: Wow. Isn't that amazing? So, tell us a little bit more about how this model has really changed the landscape of what you're doing.

Paul Bryar, MD: In terms of just retinal imaging, it used to be hard to do with expensive cameras. And it was, you know, technically challenging for the camera operator. Now, the technology has evolved where it's actually fairly easy. And we could really train anybody within about 15 minutes, 10, 15 minutes to take these images reliably. So, that could be done by a medical assistant, somebody at the check in, it could be done by the physician, the nurse, the PA, whomever's encountering that patient. That has dramatically changed the landscape.

How AI has changed this is there are now FDA-approved AI algorithms, which will be able to look at an image, and say whether or not there's retinopathy and grade that. So, there are systems that allow us to do that in a manner so that can actually have almost an immediate turnaround time for providers.

There are some issues, you know, with that because while these AI algorithms are approved, some of them are approved to detect retinopathy, I think they're not going to really have widespread adaptation quite yet. Because when we do a retinal screening, about a third of all patients will have some form of retinopathy detected. About 25% will have other eye diseases, glaucoma, macular degeneration, you know, retinal issues that might be detected. Those AI algorithms cannot and are not certified to detect all of those other conditions we talked about. So, we would have to go back and manually regrade all of the AI-driven screens just to make sure we're not missing anything else. Where this naturally goes to is we're now working on teaching these computers using both machine learning and machine teaching to say, "Okay, this is what a glaucoma picture would look like. This is what the macular degeneration," and that the next logical step would be to have these AI algorithms be able to detect more than just diabetic retinopathy.

Melanie Cole, MS: Dr. Bryar, I'd love you to speak for us on how you educate and engage patients in self-management and self-management strategies to optimize their outcomes and elaborate for us, if you would, on the importance of a multidisciplinary collaboration in the care of patients, particularly other healthcare providers involved in the diabetes management itself.

Paul Bryar, MD: So, when I counsel patients with diabetes, I say several things. You know, one, I mean, they're here to see us for their eyes. And I re-emphasize to them that, you know, "Your eye exam was done today, it was normal," or we found some mild changes that are not in fact impacting the vision. But the best way to stay healthy and to preserve the vision is to come back every year for your eye exam and stress the point that even if things aren't perfect, remember the month, what month is this? It's September. Every September is your time to come in for your eye exam. But more importantly, I also emphasize that I'm another voice in their healthcare team to say continue working with your provider who manages your diabetes with you, your regular visits and your diet, your medications, your exercise, and stress that this eye examine, myself, the patient, we're all partners with a whole team of people to effectively manage diabetes as a whole.

Melanie Cole, MS: Such an important topic and really interesting, Dr. Bryar, and so many advances in your field. Looking forward, what do you believe are the most exciting advancements on the horizon for diabetic retinopathy screening? How do you see technology, and particularly AI, which we've been discussing, continuing to shape the future of your field?

Paul Bryar, MD: Yes. We talked about the AI in terms of detecting existing disease. I think on the horizon, what we're looking at beyond that, you know, how do we look at detecting or preventing disease before it is evident in the retina, because we know that the changes we see in the retina are relatively advanced in the disease of diabetes and diabetic retinopathy. And how can we take somebody who has a normal retina, but looking at other factor, whether it's their hemoglobin A1c or the duration of diabetes, their cholesterol, their blood pressure ranges? How do we put all that information together to say that this patient is at a higher risk for developing retinopathy in the next six months or 12 months, even though they have no retinopathy? So, we're looking to leverage all this data that we collect and analyze it to stratify even the normal exams into higher risk in the future or lower risk in the future. So, we can screen them and counsel them and possibly even treat them slightly differently than the low risk patients. That's where I see the advancements going in diabetic retinopathy and diabetes-associated vision loss and prevention of that in the future.

Melanie Cole, MS: Thank you so much, Dr. Bryar, for joining us and I hope you'll join us again and give us updates as some of these advances really take hold. It really is exciting. Thank you again for joining us. To refer your patient or for more information, please visit our website at breakthroughsforphysicians.nm.org/ophthalmology to get connected with one of our providers.

That wraps up this episode of Better Edge, a Northwestern Medicine podcast for physicians. I'm Melanie Cole. Please always remember to subscribe, rate and review this podcast and all the other Northwestern Medicine podcasts.