Mark R. Barakat, MD
University of Arizona College of Medicine, Phoenix, AZ

From May 1st to May 4th, ophthalmologists gathered in Denver, Colorado, for the 2022 Association for Research in Vision and Ophthalmology Annual Meeting (ARVO22). ARVO22 featured a robust program, with more than 4000 abstracts presented across all primary ophthalmology subspecialties.

Mark R. Barakat, MD, is the director of the Retinal Research Institute with Retinal Consultants of Arizona and a clinical assistant professor of ophthalmology with The University of Arizona College of Medicine in Phoenix. Dr Barakat shares his key takeaways from the latest research on wet age-related macular degeneration (AMD) presented at ARVO22.

Which abstracts presented at ARVO22 in the wet AMD space are most practice-informing and why?

One of the biggest clinical questions right now is how to incorporate faricimab and the Port Delivery System (PDS) with ranibizumab into the treatment regimen for wet AMD.

Updated results were presented on the efficacy, safety, and durability of faricimab after 48 weeks of treatment in the phase 3 TENAYA (ClinicalTrials.gov Identifier: NCT03823287) and LUCERNE (ClinicalTrials.gov Identifier: NCT03823300) trials.1 These trials compared individualized treatment regimens of faricimab with aflibercept in patients with neovascular AMD. Patients in the faricimab subgroups received 4 initial doses of faricimab every 4 weeks and were then transitioned to a dosing schedule of every 8, 12, or 16 weeks according to prespecified criteria for central subfield thickness (CST) and best corrected visual acuity (BCVA) at weeks 20 or 24. Patients randomly assigned to receive aflibercept were dosed every 8 weeks.

Although there was a trend for more variability during the course of the trial for patients who received faricimab every 8 weeks and every 12 weeks, all subgroups demonstrated significant and comparable benefit at the end of the trial. Overall, the trial demonstrated noninferior BCVA and anatomical outcomes for faricimab up to every 16 weeks compared with aflibercept every 8 weeks. At week 48, 45.3% of patients treated with faricimab were on the dosing schedule of every 16 weeks and 78.7% of patients were on the dosing schedule of every 12 weeks.

Another phase 3 study, the ARCHWAY trial (Clinical rials.gov Identifier: NCT03677934), presented a longer follow-up of the PDS with ranibizumab for over 96 weeks.2 Use of the PDS every 24 weeks was noninferior to monthly ranibizumab, with differences in BCVA change from baseline between study arms of 0.4 letters at weeks 60 to 64 and -0.6 letters at weeks 88 to 92. The results support the continued efficacy of the PDS over 2 years with no new safety signals.

A real-world analysis of wet AMD treatment patterns in the United States was also presented at ARVO22. The investigators studied patients newly diagnosed with wet AMD who had not yet received treatment. Of 90,681 eyes, the percentages of eyes that did not receive treatment were 23%, 16%, 15%, and 13% at diagnosis and years 1, 2, and 3, respectively. The mean baseline visual acuity of eyes that received treatment was better than that of untreated eyes (52 letters vs 39 letters, respectively).3

“There are myriad ways in which AI could impact our treatment of patients with AMD, whether in predicting disease activity, anticipating the individualized need for treatment (including the optimal treatment interval), or using at-home imaging to remotely monitor patients.”

Updated data from the ongoing phase 2 AAVIATE study (ClinicalTrials.gov Identifier: NCT04514653) were also presented at ARVO22.4 How close are we to gene therapy becoming a treatment reality for patients with wet AMD?

Gene therapy with a suprachoroidal injection of RGX-314 seems to be well tolerated, with mild adverse events in cohorts 1 and 2, including the inflammation seen in 7 of 30 patients (23%). It is particularly encouraging that this mild inflammation occurred without prior steroid prophylaxis and resolved within days to weeks after treatment with a topical corticosteroid. One of the potential hurdles to making gene therapy a reality is mitigating the immune reaction seen with the viral vectors used. Early efficacy results are also encouraging and highlight the potential benefit of gene therapy for the management of wet AMD. Patients experienced stable vision and anatomy at 6 months after suprachoroidal RGX-314 and saw over a 70% reduction in the number of anti-vascular endothelial growth factor (VEGF) injections needed.
 
Paired with other encouraging data from the surgical subretinal delivery of RGX-314 and pending further pivotal trial results, these findings suggest that gene therapy may well become a clinical reality.4 Retinal gene therapy could have a positive life-changing impact in patients with AMD after just a single administration. In addition to RGX-314, voretigene neparvovec, ADVM-022, GT-005, and HMR59 are other retinal gene therapies being studied.5

Several presentations in the AMD space centered on artificial intelligence (AI) for at-home monitoring. Can AI be used in anti-VEGF selection process and in the detection of recurring intraretinal and subretinal fluid as these could play a significant role in improving patient care?

There are myriad ways in which AI could impact our treatment of patients with AMD, whether in predicting disease activity, anticipating the individualized need for treatment (including the optimal treatment interval), or in at-home imaging to remotely monitor patients.
 
In one study, researchers introduced an AI model trained on optical coherence tomography (OCT) images that could correctly identify active vs inactive neovascular AMD. Using more than 1000 images, the AI model achieved an accuracy of 0.84 (area under the precision recall curve), with a sensitivity of about 95%.6
 
Another AI model used OCT fluid biomarkers with the goal of predicting anti-VEGF treatment in a real-world patient population over 1 year.7 Patients were divided into 2 groups: those who needed more and those who needed less than the average number of injections after loading. The AI model correctly identified the patients in the two groups with an accuracy of 0.74 (area under the curve).

Another abstract focused on using machine learning to help determine the optimal dosing regimen for patients with wet AMD treated with faricimab.8 In an analysis of baseline and imaging features from patients receiving faricimab in the phase 2 AVENUE (ClincialTrials.gov Identifier: NCT02484690) and STAIRWAY (ClinicalTrials.gov Identifier: NCT03038880) trials, several methods showed promise in predicting an optimal patient-specific treatment interval.9,10
 
Lastly, a study reported the performance of the AI-based Notal OCT Analyzer (NOA™) in monitoring retinal fluid volume quantification from repeated self-imaging as well as agreement with office OCT scans at the same visit in patients with wet AMD. A deep learning algorithm analyzed scans obtained by the Notal Vision Home OCT device from 88 eyes in 47 patients with high repeatability (coefficient of variance of 4.1% on a set of scans from 26 eyes) and high agreement with CIRRUS™ HD-OCT scans (concordance correlation coefficient of 0.975 in a subset of 23 eyes).11
 
Home OCT has been shown to be well tolerated by patients with dry AMD and has been validated in both clinical trial and real-world settings, with the caveat that not all patients may have the visual acuity necessary to benefit from this technology. If further validated, home OCT could prove to be another option for monitoring patients with wet AMD, potentially optimizing treat-and-extend strategies and reducing the treatment and visit burden on patients and their caretakers.

The DAZZLE (ClinicaTrials.gov Identifier: NCT04049266) was a negative trial with KSI-301, an investigational anti-VEGF biopolymer conjugate, that failed to meet the primary endpoint.12 What is next for KSI-301?

The DAZZLE trial did, unfortunately, show that KSI-301 failed to meet the primary endpoint of mean change in BCVA from baseline to year 1.12 A total of 557 patients were randomly assigned in a 1 to 1 ratio to receive 5 mg of KSI-301 at variable intervals or 2 mg of aflibercept every 8 weeks. At year 1, the KSI-301 arm gained an average of 1 letter compared with 7 letters in the aflibercept group. Anatomic results mirrored the BCVA in that the KSI-301 group experienced a mean improvement in CST of approximately 92 µm compared with a reduction of approximately 134 µm in the aflibercept group.
    
However, in the subgroup analysis within the KSI-301 arm, the 59% of patients assigned to the dosing interval of every 20 weeks had comparable results with patients in the control group dosed with aflibercept every 8 weeks, both in terms of vision and anatomy. There was increased variability, both in BCVA and CST, in 10% of patients in the every 16 weeks dosing group, and even more so in the 30% of patients assigned to the every 12 weeks dosing group. This variability resulted in lower vision and worse anatomy in those eyes at year 1, affecting the overall endpoint.13
 
The most likely explanation for these findings is that the trial design did not adequately account for patients with a higher need by allowing intervals shorter than every 12 weeks and potentially by having stricter retreatment criteria. It is encouraging that the subgroup treated every 20 weeks achieved a BCVA at year 1 comparable to aflibercept every 8 weeks, with a steady CST curve that did not seem to show any waning anatomic effect at that interval. Data from other ongoing trials with KSI-301, including BEACON (ClinicalTrials.gov Identifier: NCT04592419), DAYLIGHT (ClinicalTrials.gov Identifier: NCT04964089), and GLIMMER (ClinicalTrials.gov Identifier: NCT04603937), will shed further light on this, as the dosing regimens in those trials are more rigid with monthly dosing for wet AMD in DAYLIGHT and a fixed, every 8 weeks dosing for retinal vein occlusion in BEACON.

Reducing the treatment burden on patients with wet AMD is a priority for clinicians. Did any real-world data stand out to you that help inform anti-VEGF dosing in patients with wet AMD?

We now have an abundance of real-world data showing that we, as clinicians, are not achieving the same robust visual gains seen in clinical trials. Although some of this can be ascribed to differences in populations (one being carefully selected for trial participation and one having no such criteria), at least part of this can be correlated with treatment frequency. One of the abstracts at the conference highlighted that eyes that get prompt treatment tend to gain and maintain more vision. It is encouraging to note that a reduced frequency of treatment in recent trials, such as with PDS in ARCHWAY trial and faricimab in TENAYA and LUCERNE trials, resulted in comparable BCVA and anatomic gains compared with rigorous treatment of either ranibizumab every 4 weeks (ARCHWAY) or aflibercept every 8 weeks (TENAYA and LUCERNE).

Key Takeaways

  • Faricimab dosed every 8, 12, or 16 weeks is noninferior to aflibercept dosed every 8 weeks in terms of BCVA and anatomical outcomes.
  • Updated clinical trial data support the continued efficacy of PDS with ranibizumab over 2 years with no new safety signals.
  • Gene therapy with a suprachoroidal injection of RGX-314 seems to be well tolerated. Gene therapy may soon become a clinical reality.
  • KSI-301, an investigational anti-VEGF biopolymer conjugate, failed to meet the primary endpoint of mean change in BCVA from baseline to 1 year.

This Q&A was edited for clarity and length.

Disclosure

Ophthalmics, LLC; Allergan, Inc.; Alimera Sciences, Inc.; Annexon Biosciences; Bausch & Lomb Inc.; Clearside Biomedical, Inc.; EyePoint Pharmaceuticals, Inc.; Kodiak Sciences Inc.; Gemini Therapeutics; Genentech, Inc.; Graybug Vision, Inc.; Gyroscope Therapeutics; Novartis Pharmaceuticals Corporation; NeuBase Therapeutics, Inc.; Ocular Therapeutix, Inc.; Oxurion NV; Palatin Technologies, Inc.; Regeneron Pharmaceuticals, Inc.; REGENXBIO Inc.; ReNeuron Group plc; RIBOMIC Inc.; Stealth BioTherapeutics Corp; and Unity Biotechnology, Inc.

References

1. London N, Guymer R, Demetriades AM, et al. Faricimab in neovascular age-related macular degeneration: updated week 48 efficacy, safety, and durability in the phase 3 TENAYA and LUCERNE trials. Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3710579.

2. Mittra R, Brooks L, Wykoff C, et al. Archway phase 3 trial of the port delivery system with ranibizumab (PDS) for neovascular AMD: end-of-study results. Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3709338.

3. Moshfeghi A, Rahimy E, Boucher N, et al. Real-world management of neovascular age-related macular degeneration in the US. Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3713085.

4. Khanani A. Suprachoroidal delivery of RGX-314 gene therapy for neovascular AMD: the phase II AAVIATE study. Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3713576.

5. Khanani AM, Thomas MJ, Aziz AA, et al. Review of gene therapies for age-related macular degenerationEye (Lond). 2022;36(2):303-311. doi:10.1038/S41433-021-01842-1

6. Hanson RLW, Airody A, O’Dwyer C, et al. EPDev-AI: Early phase development of an ai tool to determine disease activity in nvAMD. Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3713630.

7. Mares V, Bogunovic H, Leingang O, et al. Artificial intelligence to identify conventional treatment patterns in neovascular age-related macular degeneration in a real-world population. Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3714149.

8. Kikuchi Y, Neubert A, Dai J, Quezada Ruiz C. Predicting optimal treatment regimen for patients with neovascular age-related macular degeneration (nAMD) using machine learning. Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3705738.

9. Khanani AM, Patel SS, Ferrone PJ, et al. Efficacy of every four monthly and quarterly dosing of faricimab vs ranibizumab in neovascular age-related macular degeneration: the STAIRWAY phase 2 randomized clinical trialJAMA Ophthalmol. 2020;138(9):964-972. doi:10.1001/jamaophthalmol.2020.2699

10. Sahni J, Dugel PU, Patel SS, et al. Safety and efficacy of different doses and regimens of faricimab vs ranibizumab in neovascular age-related macular degeneration: the AVENUE phase 2 randomized clinical trial. JAMA Ophthalmol. 2020;138(9):955-963. doi:10.1001/jamaophthalmol.2020.2685

11. Elman M, Schechet S. Performance of AI-based Notal OCT analyzer (NOA) in retinal fluid volume quantification from repeated self-imaging with home OCT in eyes with neovascular age-related macular degeneration (nAMD). Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3709993.

12. Regillo C, Ehrlich J, Janer D, et al. Efficacy, durability and safety of KSI-301 antibody biopolymer conjugate in wet AMD – year 1 primary endpoint results from the pivotal DAZZLE study. Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3712506.

13. Bommakanti N and Russell. AMD and anti-VEGF. Abstract presented at: Annual meeting of The Association for Research in Vision and Ophthalmology; May 1-4, 2022; Denver, CO. Session ID 3709338.

Posted by Haymarket’s Clinical Content Hub. The editorial staff of Ophthalmology Advisor had no role in this content’s preparation.

 Reviewed May 2022