Clinical Trial Designs in Wet AMD: A Brief Review

FDA approval of verteporfin photodynamic therapy (PDT; Visudyne, Novartis) in 1999 marked the era of pharmacotherapy for neovascular conditions of the retina.¹ Soon after, the landscape changed dramatically with the development of biologic therapies that targeted the VEGF family of proteins central to angiogenesis and vascular permeability.^2,3

Pegaptanib (Macugen, Eyetech Pharmaceuticals), an intravitreally administered pegylated aptamer that targets the 165 isoform of VEGF-A, was the first approved in 2004.⁴ It fell out of favor once pivotal trials of newer anti-VEGF-A agents, such as ranibizumab (Lucentis, Genentech/Roche), aflibercept (Eylea, Regeneron), and off-label use of bevacizumab (Avastin, Genentech/Roche), demonstrated more clinically meaningful improvements in vision.

Despite the strides made with standard-of-care intravitreal anti-VEGF-A agents for the treatment of wet AMD, patients’ real-world results fall short of those attained in pivotal, phase 3 registration trials.^5-12 Subsequent treatments, also primarily inhibiting VEGF-A, demonstrate noninferiority to earlier agents but have mostly sought to improve durability without showing superior visual results. This article reviews design aspects of these registration trials, including those of next-generation therapeutic approaches.

MARINA AND ANCHOR: SETTING THE STANDARD

The pivotal superiority trials MARINA and ANCHOR compared ranibizumab with sham and PDT, respectively.^13,14 At 12 months, 95% of those randomized to 0.5 mg ranibizumab lost < 15 ETDRS letters versus 62% in the control group. Similarly, in the ANCHOR trial, 96% of the ranibizumab group lost < 15 ETDRS letters versus 64% in the PDT-treated group at 12 months.^15,16 No approved drug has been shown to be superior to monthly ranibizumab, thus setting the bar for newer therapeutic approaches.

NONINFERIORITY TRIALS

Active control studies seek to show that an investigational treatment is no worse than standard of care (ie, statistically noninferior). The specific noninferior margin and trial design affect the strength of the study’s findings. Regulators require a functional visual acuity endpoint, which can be categorical (percentage of patients with a 15-letter loss or gain from baseline) or continuous (mean change in BCVA from baseline).

The phase 3 VIEW 1 and 2 studies of aflibercept were largely similar to MARINA and ANCHOR.¹⁷ The primary endpoint analysis assessed noninferiority (margin of 10%) of aflibercept versus ranibizumab in the proportion of patients losing < 15 ETDRS letters at month 12. The aflibercept groups were noninferior to ranibizumab (0.5 mg every 4 weeks) and an integrated analysis found no statistically significant change in mean ETDRS letter improvement at 1 year. The trials demonstrated that 2 mg aflibercept dosed every 4 or 8 weeks (after three monthly loading injections) was noninferior to ranibizumab (0.5 mg every 4 weeks).

Monthly and bimonthly (every 8 weeks after 3-month loading) fixed regimens were used in the registration trials for ranibizumab and aflibercept, respectively. These fixed dosing regimens represent on-label standards of comparison for other follow-on therapies exploring similar or extended dosing schedules.

The low cost of repackaged bevacizumab led to its widespread off-label use in wet AMD. The landmark CATT trial established the noninferiority of bevacizumab (1.25 mg every 4 weeks) to ranibizumab (0.5 mg every 4 weeks), validating its use in practice.¹⁸

IMPROVING DURABILITY

A majority of subsequent completed or ongoing trials in wet AMD have been designed to demonstrate noninferiority in mean change in BCVA from baseline using extended dosing intervals of newer therapies compared with fixed dosing with standard-of-care treatment.

Trials of brolucizumab (Beovu, Novartis) and faricimab (Vabysmo, Genentech/Roche) marked the beginning of noninferiority studies of treatment durability for many next-generation therapies, with most employing fixed dosing with aflibercept as a control comparator arm and incorporating designs that tailored the retreatment interval based on protocol-defined disease activity. Key entry criteria largely remained consistent with earlier studies.

The HAWK and HARRIER phase 3 trials investigated 6 mg and 3 mg brolucizumab versus 2 mg aflibercept.¹⁹ A 3-month loading phase was followed by every 12-week dosing for the brolucizumab groups, with an option to decrease to 8-week dosing based on evidence of disease activity. At 2 years, brolucizumab demonstrated noninferiority in mean change in BCVA compared with aflibercept with a similar safety profile. More than half of 6 mg brolucizumab eyes were maintained on dosing every 12 weeks through 48 weeks. Despite better treatment duration, widespread adoption of the drug has been hampered by the risk of occlusive retinal vasculitis and intraocular inflammation.²⁰

The phase 3 TENAYA and LUCERNE trial patients were randomized in a 1:1 ratio to 6 mg faricimab or 2 mg aflibercept.²¹ Faricimab patients were initially dosed with four injections every 4 weeks up to week 12 and then were assigned dosing intervals of every 8, 12, or 16 weeks based on active disease criteria up to week 60. Both trials met the primary endpoint of mean change in BCVA from baseline, with faricimab showing noninferiority to aflibercept. Faricimab treatment was durable, as 80% of patients in each study achieved at least every 12-week dosing and 45% reached the maximum dosing interval of every 16 weeks by year 1.²¹

It remains to be seen how these extended treatment paradigms from the clinical trial setting perform in the real world, where retreatment decisions vary.

New Drugs Under Investigation

The FDA accepted Regeneron’s biologics license application for 8 mg aflibercept based on data from PULSAR that met the primary endpoint of noninferiority in vision gains for both the 12- and 16-week 8 mg aflibercept dosing regimens after initial monthly doses at 48 weeks compared with patients treated with 2 mg aflibercept in an 8-week dosing regimen.²² A majority of the high-dose treatment patients were able to maintain the dosing regimens. The safety profile was similar to that of the approved aflibercept dose and consistent with the agent’s known safety profile.

KSI-301 (Kodiak Sciences) is an anti-VEGF-A antibody biopolymer conjugate with a high molecular weight that is intended to increase residence time in the eye and extend durability. The phase 2b/3 DAZZLE trial randomized patients to either 5 mg KSI-301 on a flexible treatment schedule of 3, 4, or 5 months versus 2 mg aflibercept every 8 weeks following three monthly loading doses. DAZZLE did not meet the primary endpoint of noninferiority of mean change in BCVA from baseline to 12 months. A second trial, DAYLIGHT, is evaluating a more frequent monthly KSI-301 dosing regimen for noninferiority to aflibercept.^23,24

TKIs and gene therapies hold promise for controlling wet AMD in the maintenance phase of therapy with the potential for greater durability.^25-30 However, these therapies are still in clinical trials, and the relative efficacy and safety of these treatments compared with standard fixed, frequent anti-VEGF-A injections have yet to be shown in large scale, pivotal studies.³¹

SUPERIORITY TRIALS

Combination treatment approaches in wet AMD differ with regards to the primary endpoint, as they need to demonstrate superiority of effect on visual function outcomes (with favorable safety) to be considered for regulatory approval. Key entry criteria are mostly consistent with earlier noninferiority studies; however, patients presenting with worse BCVA at baseline are preferred to avoid any potential ceiling effects of the combination treatment, which could hamper the ability to achieve the superiority primary endpoint.

Despite positive phase 2 data, the results of phase 3 trials of the anti-PDGF molecule, pegpleranib (Fovista, Ophthotech), reported that in combination with ranibizumab or aflibercept/bevacizumab, the primary endpoint of superior mean change in BVCA at 12 months was not met in any of the studies. Reasons for the disappointing data may include changes in trial design from the phase 2 study and potential limited pathophysiological role of PDGF in treatment-naïve disease.^32,33

OPT-302 (Opthea Limited), an intravitreally administered VEGF-C and -D ‘trap’ inhibitor biologic, is being investigated in two phase 3 clinical trials in combination with ranibizumab (ShORe) and aflibercept (COAST).^34-36 OPT-302 is given once every 4 or 8 weeks after three monthly loading doses in combination with anti-VEGF-A therapy. Both studies’ primary endpoint is superiority in change in BCVA gains from baseline at 12 months for combination therapy versus anti-VEGF-A monotherapy. The completed phase 2b study of OPT-302 plus ranibizumab achieved the primary endpoint of a statistically significant mean change in BCVA from baseline to week 24 of +14.2 letters—an additional gain of +3.4 letters (P = .0107) over the ranibizumab plus sham control group.³⁷

Phase 2b results also showed that OPT-302 combination therapy had a mean BCVA gain of an additional +5.7 letters over the control group (16.1 vs 10.3 letters) at 24 weeks in a prespecified analysis of treatment-naïve patients with minimally classic and occult lesions. Thus, the phase 3 ShORe and COAST endpoints will be analyzed in a hierarchical fashion, starting with the primary endpoint in these two lesion types (high responders), followed by the total population (including predominantly classic lesions).^34,35

A LOOK AHEAD

As new therapeutic approaches have been developed for the treatment of patients with wet AMD, pivotal registrational trial designs have evolved to assess for further improvements in efficacy or durability of responses over existing standard of care. Our current anti-VEGF-A therapies work well, but we can still do better to gain back more vision, maintain initial visual gains, or decrease the burden for many patients who currently have a real unmet need for better long-term vision outcomes.