Rethinking end points in AMD: the power of OCT imaging with Robyn Guymer, AM, MBBS, PhD, FRANZCO, FAHMS

Robyn Guymer, AM, MBBS, PhD, FRANZCO, FAHMS. (Image courtesy of the Centre for Eye Research Australia)

New insights into imaging-based end points could transform how clinical trials are designed for intermediate age-related macular degeneration (AMD). At the Heidelberg 2025 International SPECTRALIS Symposium — And Beyond (ISS), Robyn Guymer, AM, MBBS, PhD, FRANZCO, FAHMS, spoke about the use of the ellipsoid zone (EZ) as a viable end point—marking a potential shift toward earlier intervention.

Guymer is deputy director at Centre for Eye Research Australia (CERA) and head of macular research at CERA. She is also a professor of ophthalmology at the University of Melbourne and senior retinal specialist at The Royal Victorian Eye and Ear Hospital in East Melbourne, Australia. Guymer shared more about her ISS presentation in this interview with Ophthalmology Times Europe.

Please give a brief overview of your presentation, “Using OCT-determined anatomical features as end points in clinical trials in intermediate AMD.”

Robyn Guymer AM, MBBS, PhD, FRANZCO, FAHMS: So I'm going to talk about a new end point for clinical trials, particularly in intermediate AMD, but which currently is approved for geographic atrophy. So this is the EZ line, which is thought to represent the photoreceptor mitochondria. It's been, as we understand it, approved as an end point for phase 3 trials by the FDA. My presentation looks at the use of the loss of the EZ line in a cohort of people with intermediate AMD to see how it functions as an end point.

Not only did we look at a natural history study of intermediate AMD, but also an intervention study, which was with a nanosecond laser, which we found slowed down the progression of AMD in the vast majority of intermediate AMD cases. And we were able to show [this therapy] also slowed down the rate of the loss of the EZ. That's called the LEAD study; it's with a nanosecond laser that aims to slow the progression of intermediate AMD. We published that in 2019. That study showed we slowed the progression of AMD that didn't have reticular pseudodrusen—not everyone, but just those without reticular pseudodrusen. And the LEAD study validated that we were able to save the EZ in those people as well. That's what I'm going to talk about [at the ISS].

Why is intermediate AMD the focus of this specific presentation?

Guymer: Ideally, I think everyone would like to intervene early, before vision is threatened. But at the moment, intervention studies and treatments are for late-stage AMD, geographic atrophy (GA) or neovascular AMD. And the reason why, presumably, pharmaceutical companies choose to intervene there is it's much easier to show a benefit over a reasonable period of time.

If you want to intervene earlier, in intermediate AMD, for example, it could be years before someone progresses. Up until this moment, the end points have been late-stage AMD and growth of late-stage AMD. It's a challenge to intervene early, which is where everyone would like to intervene, and have an end point that you can reach within a reasonable timeframe. The FDA's shift to consider the EZ line as an acceptable regulatory end point means that for the first time, really, we can start to look at intervening in intermediate AMD.

If you really want to have a treatment that targets those people most likely to progress, then you would intervene in intermediate AMD. One in seven people over the age of 50 have drusen. So you don’t really want to intervene in everyone unless, of course, you had a treatment like a supplement, or something very cheap and easy. But anything other than that, you really want to target people at highest risk of progression. So you would target high risk, intermediate AMD.

Are there technical advancements or regulatory changes which make anatomical features feasible end points in today’s ophthalmic landscape?

Guymer: I think, possibly, a bit of all things. So there is a disease called MacTel, or macular telangiectasia type 2b, and it was the subject of a study where the end point was this change in the rate of growth of the EZ line. Which, in that disease, makes sense; patients get a patch of atrophy, it's only one patch and it gets bigger. The regulator in that disease trial accepted a structural end point, rather than a visual acuity end point, which, up until that time, really was the status quo for neovascular AMD trials. Seeing that the regulator accepted that end point for MacTel, I guess people started to look at the dry form of AMD or geographic atrophy.

In addition, it is probably the emergence of artificial intelligence that allows us to actually measure this EZ line through algorithms. It's not easily possible on a large scale for humans to measure this line, or the absence of the line. Now we have an ability to—in a feasible and sort of cost effective way—measure this with an artificial intelligence algorithm. And we have had some treatment that seems to impact this line. I think it's a culmination of everything coming together.

What’s the biggest challenge in clinical trial design? Is it keeping up, clinically, with the pace of innovation? Is it adapting to quick changes in the landscape?

Guymer: Herein lies the very big problem. The biggest challenge is not the clinicians, but pharma and regulatory authorities. They are not budging. The regulatory authorities want specific end points, and pharmaceutical companies only want to run trials in which the end point is a regulatory-approved end point. What has to shift is for pharma to be prepared to run a phase 2 study to, potentially, an end point that isn't a regulatory-approved end point, just to gain confidence into which drug to take forward into phase 3.

At the moment, [most companies are] not prepared to do a longer study. They're not prepared to have an end point in phase 2 that's not the same end point in phase 3. I would argue that there's plenty of ways to be agile: to do phase 2 studies which look at saving the EZ, [external limiting membrane] ELM, [outer nuclear layer] ONL thickness or drusen volume, but none of them are approved end points. So, pharma is not interested...In a chronic disease like AMD, glaucoma or diabetes, companies need to change their mindset, to be prepared to either do a longer trial or have their phase 2 study reflect their best guess at the best end point that will give them confidence that their drug works. This is where we are at the moment.

And certainly, I'm trying. I started a study with Roche and Genentech called the HONU study, which is looking to try and validate an incident end point, so that you don't have to look at the slopes of the growth. Companies have to believe in themselves that stopping nascent GA is as good as stopping GA. Now, the FDA has not approved that yet. But if you want to do a trial starting early and finishing early, then really, you only have to convince yourself that your drug works. You don't have to convince the FDA.

How can you do this without companies being agile and taking risks? I think, if the clinician could show that drusen volume reduced and function did not reduce, or there was no progression to atrophy, we'd be very happy with reduced drusen volume—but that's currently not an approved end point.

How are modern imaging tools changing the way that clinicians understand the structure of the eye?

Guymer: I think OCT, and now better-resolution OCT, clearly shows us much more than what we used to get from a color fundus photo. The ability to see the whole 10 layers of the retina, and in the sphere of age-related macular degeneration, to be able to see the outer retina and the changes much more granularly than in a color fundus photo allows you to monitor progression in much finer steps. So thus, if you have an intervention, it might be possible to start at step one and end at step two or three, rather than being required to go all the way to show that you're not going to lose vision from the late stage of disease. High-resolution OCT will allow, and does allow, a more granular ability to see progression.

I think in the future, with the higher resolution OCT, we'll get to see more, and perhaps it will allow us to understand the aetiology and the actual pathological processes better than we do now. So not just only see what goes missing, but what precedes what, and what the layers look like. For example, we still can't have a very good look at the Bruch's membrane, where the high-resolution OCTs might allow that. Also, the retinal pigment epithelium (RPE). We only sort of grade "Is the RPE present or not?" Clearly, it is sort of sick. But we don't really have a marker of a dysfunctional RPE, structurally. The hope would be, perhaps with better resolution in the future, we might be able to gauge that this RPE is sick and not working well.

What do you hope audiences take away from your presentation?

Guymer: The exciting thing that's really changed in the last 12 to 24 months is that people are thinking about intervening early in intermediate AMD. For the first time, we have the possibility of being able to design a feasible trial that gets to an end point which the FDA, at least, has seemingly agreed to. It opens up the possibility that we may have treatments for earlier in the disease. I think what pharma can take away, if they're in the audience, is because we now have ways that we can intervene that seem to work, we just need to work out how to get those interventions underway earlier, and make them safer, perhaps. But it's now at least feasible. You can offer pharma and biotech a possible pathway for their drug to be tested.