On this episode of Retina Mentor Moments, John W. Kitchens, MD, sits down with David Eichenbaum, MD, FASRS, for a rapid-fire tour of the durability race in medical retina, covering high-dose aflibercept, angiopoietin-2 (Ang-2) inhibition, tyrosine kinase inhibitors (TKIs), gene therapy, and the pipeline therapies heading into late-stage trials.
Dr. Eichenbaum is a partner and Director of Research at Retina Vitreous Associates of Florida and a Volunteer Collaborative Associate Professor of Ophthalmology at the Morsani College of Medicine, University of South Florida. His site has enrolled in many of the pivotal programs discussed below.
Biosimilars: an economic decision, not a clinical one
Biosimilars offer no clinical advantage, Dr. Eichenbaum explains; at best, they match the reference product. As such, they belong in a practice where there is an economic reason to switch, such as savings for the health system, patient, or practice.
Why 4× the aflibercept dose doesn't provide 4× the durability
Aflibercept 8mg (Eylea HD, Regeneron) delivers four times the molar dose of aflibercept 2mg, yet in the pivotal PULSAR trial it matched, rather than exceeded, the 2mg arm's visual gains while extending intervals: 83% of 8mg patients held intervals of 12 weeks or longer through year 1, and 78% through year 2.1,2
When asked why durability doesn't scale with dose, Dr. Eichenbaum compared biologics’ durability in the eye to cooking an omelet: two egg whites on a hot plate denature into an omelet, but eight egg whites don't make that omelet four times better.
“I think you hit a rate-limiting step in how long a biologic can last in a warm, 98.6° vitreous environment. With a non-conjugated fusion protein, you get more durability with 8mg, but you don't get four times the durability because you're beginning to hit whatever the rate-limiting step is of frying your egg whites in the eye.”
The case for Ang-2: more than molar dose
Dr. Eichenbaum noted that his practice enrolled heavily in the phase 2 and phase 3 faricimab (VABYSMO, Genentech) trials across all indications; he has also worked on Ollin Biosciences' OLN324, a VEGF and Ang-2 bispecific antibody and Merck's TIESPECTUS (MK-8748), a VEGF inhibitor that activates Tie-2 rather than blocking Ang-2 directly.3,4
The open question is how much of faricimab's benefit comes from its higher molar dose of anti-VEGF Fab versus Ang-2 inhibition itself. His read is that Ang-2 contributes:
- OLN324 vs. faricimab: In the head-to-head phase 1b JADE study, OLN324, the higher-potency VEGF/Ang-2 bispecific dried DME faster and more completely than faricimab and flattened wet AMD pigment epithelial detachments (PEDs) faster.3
- Aflibercept 8mg vs. 2mg: Quadrupling the VEGF blockade produced a smaller anatomic step-up than adding (or adding more) Ang-2 inhibition, in his view.
- Faricimab post-hoc analyses: Reduced leakage with faricimab could reflect molar dose, but he believes Ang-2 is also acting on vascular leakage.5
He acknowledges the caveats; these are cross-trial comparisons and early, small studies, “but there's definitely something happening when you put the Ang-2 in there."
Adoption challenges for retina drugs: Price, practice flow, and access
Approval and efficacy doesn’t always guarantee adoption in retina care, as price, practice operations, and access often dictate which drugs actually reach patients.
LYTENAVA
In July 2026, the FDA approved Outlook Therapeutics' LYTENAVA (bevacizumab-vikg) for wet AMD, the first ophthalmic bevacizumab formulation approved in the US.6 Dr. Kitchens put three questions on the table: Will it succeed at the roughly $485-per-dose price under discussion?7 Where does it fit? And what happens to compounded bevacizumab?
Dr. Eichenbaum explained he would gladly use an FDA-approved, GMP-manufactured product in place of repackaged bevacizumab, but the problem is price. A figure near $500 reflects pre-COVID anti-VEGF pricing, he says, and is roughly 10 times what practices pay for repackaged bevacizumab, which he expects to create friction with payers and patients’ out-of-pocket costs.
Moreover, the approved-product advantage also matters less now that low-cost, FDA-approved options exist, including ranibizumab biosimilars such as CIMERLI (ranibizumab-eqrn, Sandoz). He also doesn't expect repackaged bevacizumab to disappear, as off-label uses such as neovascular glaucoma still require a low-cost option.
BEOVU
Brolucizumab (Beovu, Novartis) faces a different barrier: strong efficacy, but limited role. “Refrigerator space is finite, new agents keep arriving, and good alternatives exist,” he explained. Veteran retina specialists remember it as "the driest of the dry," but its association with retinal vasculitis and retinal vascular occlusion left lasting concern.8 Yet at a recent Pacific Rim program, Australian colleagues told him that they still use it often and see little vasculitis.
There is precedent for safety improving over time; in early ranibizumab trials, which used a different formulation from the commercial product, intraocular inflammation (IOI) was one of the most common adverse events, but LUCENTIS (Genentech) now has among the lowest IOI rates of any approved anti-VEGF.9-10
SUSVIMO
The Port Delivery System (SUSVIMO, Genentech) shows how operational barriers can stall even a well-regarded product. Dr. Eichenbaum has worked with it for a decade, performing dozens of implants and more than 100 refills, and likes the product in principle.
Yet he has placed fewer than 10 implants since commercial launch, for three reasons:
- Finding the right patient: The ideal candidate needs frequent anti-VEGF injections, and better drugs make those patients rarer.
- Stopping to recognize them: Identifying a candidate and discussing the device adds 10 to 15 minutes of chair time and interrupts clinic flow.
- Securing access: Nearly every case has hit an access problem unless the patient has Medicare with a strong supplemental plan.
While he doesn’t predict that the Port Delivery System will fail, with so many alternatives for wet AMD and DME, it seems to have less of a place than if it would deliver an effective geographic atrophy drug.
These issues can be expanded to other novel products, including ENCELTO (revakinagene taroretcel, Neurotech) for MacTel and the surgically delivered gene therapy ABBV-RGX-314 (surabgene lomparvovec, RegenxBio), as they will likely require similar operational demands.
Developments in TKI therapies
Dr. Eichenbaum's site has participated in both sustained-release TKI programs (DURAVYU and AXPAXLI) from phase 1 onward, and in his opinion, going intracellular and downstream of the VEGF receptor "makes a lot of sense" scientifically.
DURAVYU
Both phase 3 trials of DURAVYU (vorolanib intravitreal insert) in wet AMD missed their primary noninferiority endpoint for visual acuity versus aflibercept 2 mg.12,13 In LUGANO (NCT06668064), EyePoint attributed the miss to 9 of 211 DURAVYU patients who lost 15 or more letters for reasons unrelated to wet AMD.12
Meanwhile, LUCIA (NCT06683742), also missed the primary endpoint, though DURAVYU cut treatment burden by 38%, and 52% of eyes needed no supplemental aflibercept through week 56.13 EyePoint has since restructured and shifted its focus to DME, where its phase 3 COMO (NCT07449936) and CAPRI (NCT07449923) trials are ongoing.13
Ocular Therapeutix: SOL-1, SOL-R, and the single-dose problem
SOL-1 (NCT06223958) met its superiority endpoint: 74.1% of patients given a single dose of AXPAXLI (axitinib intravitreal implant) maintained vision at week 36, versus 55.8% after a single dose of aflibercept 2mg.14,15
The company currently plans to submit an NDA for AXPAXLI in the treatment of wet AMD in Q4 2026.
Dr. Eichenbaum flags two issues:
- Redosing data: Ocular plans to file on SOL-1 alone. If the redosing trial SOL-R (NCT06495918, topline Q1 2027) then misses, the label would cover a single dose, which he calls "the DURYSTA problem."14 If SOL-R meets its noninferiority endpoint against every-6-month aflibercept 8mg and every-8-week aflibercept 2mg, Ocular will have both a superiority and noninferiority trial in hand.
- The adverse event table: With repeat dosing, implant material that patients can see could hurt adoption.
Is SOL-1 a template for future trials?
Dr. Kitchens asked whether more trials will be designed to show superiority in ways that don't map to clinical practice. Dr. Eichenbaum doesn't expect many. SOL-1 was a "boutique" design: it enrolled super-responders who did well on a single aflibercept dose, a durability rarely seen in practice, and rescue waited for a 15-letter loss that clinicians wouldn't tolerate.
Consequently, he is watching the SOL-X extension study (NCT07516132) to see how many of those patients recover vision. As a scientific test of durability, though, he calls it a nice design that showed AXPAXLI does something different from a single dose of aflibercept.
Gene therapy: "one and less," not one and done
Gene therapy chases the dream of a single injection, but Dr. Eichenbaum says no protocol is truly one and done. Every program tolerates some fluid, allows supplemental injections, and leans on topical drops to manage inflammation. The realistic goal is "one, manageable, and less."
He is bullish that first-generation intravitreal gene therapies work and expects at least one or two to prove effective without unacceptable safety baggage. Since current protocols require an anti-VEGF run-in, he doesn’t see them reaching first line soon. Instead, he expects them to move up the algorithm gradually if supplemental injections stay low.
Economics will also shape adoption. If reimbursement in 2030 looks like it does today, gene therapies priced in the high five to six figures will likely go to patients with good coverage, while those with managed care or poor access will likely see them pushed far down step-therapy algorithms.
Retina therapies targeting IL-6
Having been involved in Genentech's IL-6 program for uveitic macular edema (UME, VAMIKIBART) and Kodiak Science's IL-6/VEGF bispecific protein KSI-101 (tabirafusp alfa) trials, Dr. Eichenbaum predicted that macular edema secondary to inflammation (MESI) is a sleeping giant.
The phase 3 PEAK (NCT06990399) and PINNACLE (NCT06996080) trials evaluating KSI-101 for MESI are currently enrolling participants,16 and he noted that it has been easy to recruit patients because the population is large and underrecognized, and no intravitreal biologic is approved for it.
“If something is swollen and they're not a wet AMD or diabetic patient, and they don't have an RVO, everything else is MESI.”
If KSI-101 shows IOI rates similar to aflibercept and no occlusive vasculitis beyond what an inflamed population would be expected to show, he predicts a major indication. Kodiak's first PEAK readout is expected in December 2026, with the combined PEAK and PINNACLE analysis in Q2 2027.17
Roche recently announced that it had submitted a BLA for VAMIKIBART in the treatment of UME, which the FDA accepted for review, based on two completed phase 3 trials: MEERKAT (NCT05642312) and SANDCAT (NCT05642325). The FDA is expected to make a decision by July 2027.18
WNT and Tie-2: testing truly new biology
"I don't think we're cutting the mustard with just different flavors of VEGF inhibition," Dr. Eichenbaum says. Faricimab introduced the first new biology, Ang-2 inhibition, but left open how much its effect comes from molarity versus Ang-2.
Merck's REMIGROMIG (MK-3000, formerly RESTORET), a tri-specific WNT pathway agonist, offers a purer test. In the pivotal phase 2b/3 BRUNELLO trial (NCT06571045) in DME, it was noninferior to monthly ranibizumab at week 52, a control that Dr. Eichenbaum explains is effective, safe, well-established, and hard to beat.19 Of note, Merck reported that the REMIGROMIG arm had higher rates of proliferative diabetic retinopathy, vitreous hemorrhage, and treatment discontinuation. A second pivotal trial, BAROLO (NCT06957080) is ongoing.19
In wet AMD, Dr. Eichenbaum is more excited about Merck's TIESPECTUS, now in two pivotal phase 2b/3 trials (MALBEC [NCT07440225] and TORRONTES [NCT07496567]) against aflibercept.20
Other WNT agonists are in development for DME, including:
- Surrozen: SZN-8141, a bispecific WNT agonist/anti-VEGF, and SZN-8143, a trispecific WNT agonist/anti-VEGF/anti-IL-6. Both are in preclinical studies.21
- Neuvasq: NVQ401 and NVQ501, which activate WNT signaling through a different receptor complex (Gpr124/Lrp6). Dr. Eichenbaum says preclinical data suggest they may be more potent than the Merck or Surrozen molecules.22
The Empire Forum: Where retina meets investment
As a co-chair for the Empire Forum, Dr. Eichenbaum explained that the inaugural one-day event is open to all attendees and designed specifically to get physicians, industry, and the investment community talking to each other. The event is taking place Thursday November 5 at the Sheraton New York Times Square Hotel, where the OSN New York & Retina 2026 event kicks off the day after.
“If you want to be in the room where it happens with investment, industry, and futurism, think about the Empire Forum.”
Key takeaways
- Durability has limits: Quadrupling the aflibercept dose doesn’t quadruple durability, and Ang-2 inhibition may add benefits beyond molar dose.
- Adoption isn’t guaranteed: Price, clinic flow, and access limit can cause obstacles in practice implementation for approved therapies.
- The TKI race has split: DURAVYU missed its primary endpoint in both phase 3 wet AMD trials, while AXPAXLI met superiority in SOL-1 and awaits redosing data.
- Gene therapy will be “one and less,” not one and done: Expect gradual uptake, shaped by run-in requirements and reimbursement.
- New mechanisms are arriving: IL-6 inhibition for inflammatory macular edema and WNT and Tie-2 agents for DME and wet AMD are entering late-stage testing.
