Seeing the Unseen | Breakthroughs in Back-of-the-Eye Therapies

Session Focus

The panel covered how new modalities are reaching the back of the eye, spanning gene therapy for rare inherited disease, encapsulated cell technology, and genomics-driven approaches to common degenerative conditions. Recurring themes were endpoint selection, natural history data, patient stratification and what investors now require.

A Pivotal Readout

Lance Baldo MD opened with results reported that Monday from Beacon’s pivotal phase three study of laruzova in RPGR-mutated X-linked retinitis pigmentosa. The trial showed roughly a 30 percent three-line response rate on low-luminance visual acuity in the high-dose cohort and about 25 percent in the low-dose cohort. He said the results exceeded both his expectations and those he had given his board.

He attributed the outcome to several improvements over earlier attempts rather than any single factor: improved vector design and capsid choice optimized for photoreceptor tropism, codon optimization of the full-length transgene coding for the RPGR protein, and correct patient selection, study design and choice of endpoints.

What Success Looks Like in Clinic

Jason Comander MD, PhD put the science in clinical context with a patient story. A teenager with X-linked RP, treated in an earlier trial with a different company, came in at one month and said, with typical teenage understatement, that he could see better in the dark. Comander told him people had been coming to that clinic with retinitis pigmentosa for 50 years, and he was one of the first to improve.

On the gaps, he identified scale as the central challenge: reaching more patients, more genes and more mutations. Populations below a certain size struggle to attract investment.

He offered two practical counterweights. The rare pediatric disease priority review voucher, currently valued around $200 million, can cross-subsidize substantially. And the field is learning to work at smaller scale and lower cost through approaches like n-of-1 treatments and antisense oligonucleotides.

Encapsulated Cell Technology and a New Endpoint

Richard Small described Neurotech arriving at its indication in reverse. The company had a durable one-and-done encapsulated cell platform expressing a neurotrophic factor that protects photoreceptor health, and went looking for a disease where the platform’s benefit could actually be demonstrated. Macular telangiectasia, an orphan condition, fit both clinically and commercially.

The endpoint problem was the harder part. MacTel progresses so slowly that patients lose roughly one letter of vision a year, which makes a 15-letter endpoint an impossible study that no one would fund. Neurotech instead pursued change in ellipsoid zone, an imaging measure reflecting photoreceptor health.

Small said it took about a year and three or four long sessions with the FDA, and that the agency, particularly Wiley Chambers, became enthusiastic. He noted with pride that Neurotech was the first company to get that primary endpoint through. Without acceptance, there was no way to demonstrate the treatment effect at all.

Genomics and Natural History in Common Disease

Cheng Zhang described Character Biosciences as a clinical-stage company focused on common eye diseases of aging, starting with dry age-related macular degeneration. AMD is prevalent and polygenic, producing substantial heterogeneity across patients.

Unsatisfied with existing datasets, Character ran its own self-sponsored large-scale observational trial in partnership with US ophthalmology centers and private practices. It has enrolled close to 10,000 patients across AMD stages, integrating genetic data with longitudinal clinical progression and ocular imaging.

The goal is understanding not just which genes affect disease incidence but which affect the rate of progression, since that is what a therapy would modulate and what a regulator would assess.

That dataset feeds trial design directly. Registrational endpoints such as growth of geographic atrophy lesions and change in ellipsoid zone loss live in the imaging, so the company can test which biomarkers are most sensitive at which disease stage, then stratify patients by baseline anatomy, lifestyle factors and genetics to pre-specify who is most likely to progress and respond.

A Shared Model for Natural History Data

Comander noted that Mass Eye and Ear has collected data on slowly degenerative retinal conditions for 50 years, the work of a large team rather than any individual.

He highlighted a newer model worth replicating. The Foundation Fighting Blindness Clinical Consortium, led by Jacque Duncan and Rachel Huckfeldt, runs an ongoing natural history study with multiple arms and substudies. A company interested in a particular gene can sponsor slots rather than building a study from scratch. The sites are established, the equipment is in place and the IRB protocol is already approved, which he called a major advantage in both time and expense.

The Investor View

Maha Radhakrishnan MD described a genuine resurgence of ophthalmology-focused companies. Sofinnova evaluates against four pillars, beginning with unmet need, which she stressed is not a binary question. The real question is which specific patient populations benefit more than others.

She praised Character’s longitudinal cohort approach as exactly what allows an investor to understand heterogeneity, identify subpopulations, enrich trials to reach data faster and carve out a clear development path.

She was direct about what no longer impresses. Companies routinely arrive with anti-VEGF programs promising greater durability, and durability alone does not cut it. Existing products are good, biosimilars are good, and treaters already work on a treat-and-extend paradigm, so 16 weeks of durability means nothing on its own.

What she wants to see instead: impact on vision, impact on persistent fluid, impact on the need for rescue medication, and impact on retinal thickness. Her advice to companies confident in a durable asset is to identify the population that genuinely benefits, particularly inadequate responders, rather than pursuing a broad category.

Key Takeaways

1. Endpoint selection can determine whether a program is viable. Slowly progressive diseases require imaging biomarkers, and regulators can be persuaded with sustained engagement.

2. Vector and delivery refinements compound. Beacon’s result came from several incremental improvements rather than one breakthrough.

3. Natural history data is now a competitive asset. Both a company-run study and a shared consortium model were presented as viable routes.

4. Genetics predicts progression, not just risk. That distinction is what enables patient stratification for trials.

5. Durability alone no longer wins investment. Functional and anatomic outcomes matter more than dosing interval.

6. Rare disease economics have workarounds. Priority review vouchers and smaller-scale modalities extend reach to smaller populations.