MAGIC: Study Design and Rationale for the Phase 2 Clinical Trial of Faricimab for Non-Proliferative Diabetic Retinopathy

Abstract

Purpose

Describe the design and rationale of the MAGIC (ClinicalTrials.gov identifier, NCT05681884) trial assessing change in retinal nonperfusion (RNP) among eyes with non-proliferative diabetic retinopathy (NPDR) in sham and treated patients.

Design

Phase 2 prospective, randomized, multicenter, open-label, clinical trial.

Participants

Adults with NPDR and substantial RNP (>5 disc areas) on ultra-widefield fluorescein angiography images.

Methods

Patients will be randomized (1:1) into 1 of 2 arms: group 1 receiving intravitreal faricimab 6 mg every 4 weeks (Q4W) and group 2 being observed Q16W. At 48 weeks, group 1 will receive faricimab Q16W and group 2 will initiate faricimab Q4W through the final trial endpoint at year 2. The primary endpoint is change in RNP area through week 48. Secondary endpoints include change in RNP area through week 96; percentage of neovascularization, vitreous hemorrhage, DME, and/or PDR development; change in best-corrected visual acuity; change in central subfield thickness; contrast sensitivity (CS) measured using the quantitative contrast sensitivity function (qCSF); detection of apoptosing retinal cells (DARC) analysis; and proportion of subjects with ≥ 2-step improvement in Diabetic Retinopathy Severity Score (DRSS). Safety outcomes include incidence and severity of adverse events.

Main Outcome Measures

Study design rationale.

Results

Recruitment commenced in May 2023 and was completed in May 2024. Group 1 will be dosed Q4W to maximize detection of an impact on RNP progression compared to the untreated group 2. During year 2, group 1 will be dosed Q16W to assess for maintenance of any changes in RNP trajectory achieved during year 1. Group 2 will assess the natural history of RNP progression through year 1, then transition to maximal dosing to evaluate for a potential change in trajectory of RNP progression. Additional assessments, including CS and DARC, will address the unmet need for a better understanding of visual dysfunction among patients with NPDR and investigate the potential role of apoptosis in RNP, respectively.

Conclusion

MAGIC is a randomized clinical trial that assesses RNP progression among eyes with NPDR, including its natural history and the impact of faricimab. Its innovative study design also explores the utility of novel assessments of retinal physiology and function.

Introduction

Diabetic retinopathy (DR) remains one of the most common causes of vision loss globally [1]. Characteristics associated with DR such as loss of pericytes, thickening of endothelial cell basement membranes, and development of microaneurysms contribute to breakdown of the blood-retinal barrier and disruption of retinal vasculature, architecture, and function [2]. This pathophysiologic vascular impairment initially affects capillary beds then progresses to involve both larger arterioles and veins, leading to inadequate perfusion to the metabolically active retina and local ischemia.

Areas of retinal nonperfusion (RNP) are an important biomarker in DR and are present early in the disease course, even among patients with diabetes mellitus (DM) without evidence of clinically apparent DR [3]. As DR severity worsens, RNP has also been observed to increase [3, 4]. Patients with RNP have been reported to have 3.75 times increased odds of having concurrent diabetic macular edema (DME), and prospective trials have identified RNP as a predictor of disease progression to PDR, vision-threatening features, such as vitreous hemorrhage, and DRSS worsening [5, 6]. These findings highlight the importance of RNP when managing patients with DR and even DM more broadly.

The natural progression of RNP is incompletely understood. A post hoc analysis of the phase 3 RIDE/RISE trials quantified macular nonperfusion (MNP) among eyes with DME and found a longitudinal, nonsignificant increase in MNP area among the sham eyes through 2 years which appeared to be blunted with monthly ranibizumab dosing [7]. It is biologically plausible that intravitreal (IVT) anti-vascular endothelial growth factor (VEGF)-A pharmacotherapy could slow RNP development and progression. For example, in a nonhuman primate model, repeated IVT injections of VEGF led to large areas of capillary closure and RNP [8]. Anti-VEGF medications may allow for improved retinal blood flow and reduced local tissue ischemia by decreasing vascular hyperpermeability. Additional proposed mechanisms for anti-VEGF mediated retinal reperfusion include pericyte restoration and basement membrane normalization, which may facilitate retinal microvasculature regrowth [9, 10]. Similar to data from RIDE/RISE, post hoc analysis of the phase 3 VISTA study supported this hypothesis, demonstrating that aflibercept slowed RNP progression through 2 years compared to control and, in some cases, improved perfusion by decreasing zones of RNP [11]. However, not all reports corroborate these findings, with some smaller studies documenting worsening of MNP after anti-VEGF treatment among patients with DME [12, 13]. Similar conflicting literature is also present in the context of peripheral nonperfusion after anti-VEGF treatment, and may be at least partially dependent on the imaging modality utilized [14].

Hypotheses of why some vascular beds in some eyes may respond to anti-VEGF therapeutics are complex and multifaceted. First, it is worth considering and distinguishing between ischemia, characterized by reduction in local tissue delivery of oxygen and micronutrients, and nonperfusion, characterized by anatomic obstruction or loss of retinal blood vessels. Drivers of ischemia may differ between the macula and periphery; for example, macular ischemia has been attributed mainly to pericyte loss, hyperpermeability, and VEGF increase whereas peripheral ischemia has been attributed more to thickening of the basement membrane [14]. Thus, perfusion status and the opportunity for reperfusion of apparently non-perfused retina may also be tissue and region dependent in that cases of paradoxical increases in RNP following anti-VEGF therapy may be attributed to disruption of the normal retinal and choriocapillaris circulation, vasoconstriction, and reduction in capillary density [14]. Much remains incomplete in our understanding of RNP and the effect of anti-VEGF therapy on its genesis and progression.

Although several studies have evaluated RNP, published findings are conflicting and limited by factors including variable methods of quantifying RNP, differing imaging modalities, and varying study designs [14]. The majority of studies available are retrospective, with relatively small sample sizes and limited follow-up [14]. A recent meta-analysis of clinical trial data on the impact of anti-VEGF treatment on RNP demonstrated that treatment may slightly impact the progression of RNP in DR [15]. Critically, however, the review found low certainty of evidence due to the lack of consensus regarding minimally important differences in RNP progression and the differences in outcome measures reported by the studies, which led to the conclusion that additional trials are needed to assess the natural history of RNP and the impact of treatments on RNP progression [15]. There is a need for a large, robust, prospective study, appropriately powered to assess the natural history of RNP and the effects of anti-VEGF therapies on RNP progression. MAGIC aims to address these unmet scientific needs through a 2-year prospective trial. Herein, we describe the rationale and methodology of the ongoing MAGIC trial.

Study Design and Methods

Study Overview

MAGIC is a multicenter, prospective, randomized, open-label, phase 2 clinical trial designed to examine change in area of RNP within and outside the macula over 2 years using ultra-widefield fluorescein angiography (UWFA) within eyes with non-proliferative DR (NPDR) either observed or treated with IVT faricimab 6 mg. This study is being conducted in accordance with the principles of the Declaration of Helsinki and the Health Insurance Portability and Accountability Act. Written informed consent was obtained before initiation of any study procedures, and the study protocol was approved by the Advarra Institutional Review Board (Columbia, MD, USA) on January 6, 2023 (IRB approval Pro00067762), before study start.

The target sample size is 150 patients. Study eyes are randomized 1:1. Group 1 subjects will receive IVT faricimab 6 mg every 4 weeks (Q4W), defined as every 28 days ± 7 days with at least 21 days between injections. Starting at week 48, group 1 subjects will be treated Q16W through week 80 (online suppl. Table S1; for all online suppl. material, see https://doi.org/10.1159/000550491). Group 2 subjects will be seen and observed Q16W. Starting at week 48, subjects in group 2 will be administered IVT faricimab 6 mg Q4W from week 48 through week 92 (online suppl. Table S2). Both groups are to have an end of study visit at week 96 with no treatment administration. Subjects who discontinue study treatment on or before week 16 are eligible for replacement at the site level, with the next eligible subjects. If a subject cannot be replaced at the site level, the subject can be replaced at the study level.

Group 1 participants can receive rescue treatment if indicated after week 48, and group 2 participants can receive rescue treatment if indicated before week 48. At least one rescue criterion must be met prior to rescue therapy administration: proliferative DR (PDR) development or clinically relevant worsening neovascularization of the disc, iris, angle, or elsewhere as confirmed by imaging and review with the medical monitor team, with gonioscopy at investigator discretion if clinically indicated; decrease in best-corrected VA (BCVA) by 5 or more letters from best previous BCVA due to progressive DME; or worsening central subfield DME causing vision loss. Upon meeting rescue criteria, IVT faricimab 6 mg will be given Q4W for the remainder of the trial.

Study Participants and Eligibility Criteria

Men and women of at least 18 years of age were eligible to participate. A single eye from each subject could be enrolled. If both eyes were eligible, the site investigator determined the study eye. Ocular inclusion criteria in the study eye included BCVA ≥20 Early Treatment Diabetic Retinopathy Study (ETDRS) letters (20/400 Snellen equivalent) at a distance of 4 meters, NPDR as confirmed by the site investigator, and substantial RNP, defined as >5 disc areas on UWFA. Pertinent ocular exclusion criteria included any history of treatment with anti-VEGF or any periocular or IVT corticosteroid within 4 months prior to screening, optical coherence tomography (OCT) central subfield thickness measurement >325 µm due to DME, evidence of ocular infection at screening, any pan-retinal photocoagulation (PRP) treatment or macular laser received prior to screening, retinal vein occlusion, cystoid macular edema not attributed to DM, current vitreous hemorrhage obscuring imaging and/or dilated indirect examination, any intraocular surgery within 4 weeks prior to screening, and any intraocular inflammation including scleritis at screening in the study eye. Prohibited medications and therapies are listed in online suppl. Table S3.

Study Status

MAGIC began recruitment May 24, 2023, and finished enrollment on May 1, 2024. MAGIC is currently ongoing as of the date of this publication. Thus, planned interventions and analyses are indicated in the future tense in the following sections.

Anatomic, Functional, and Molecular Assessments

Retinal imaging features will be evaluated with spectral domain-OCT (SD-OCT), OCT-angiography (OCT-A), fundus photography (FP), and UWFA. All SD-OCT (volume scan, IR+OCT, 20° × 20°, high resolution, 97 sections) and OCT-A (volume scan, OCTA, 20° × 20°, high speed preset) images will be obtained using the Heidelberg Spectralis (Heidelberg Engineering, Heidelberg, Germany). All FP and UWFA images will be acquired on Optos ultra-widefield devices (Optos, Dunfermline, Scotland). All SD-OCT, OCT-A, FP, and UWFA images will be graded sequentially by two masked certified graders at the central reading center (CRC; Cole Eye Institute, Cleveland Clinic, Cleveland, OH, USA). Image analysis will be performed using reproducible standardized grading protocols as previously described in the RECOVERY prospective study; UWFA images from MAGIC are being graded by the same independent reading center that graded UWFA images from RECOVERY [16, 17]. The CRC will assess for RNP, central subfield thickness, DRSS, fundus avascular zone (FAZ) area, leakage index (area of leakage divided by the analyzable retinal area), and microaneurysm count (absolute number of microaneurysms within the analyzable area).

Confocal scanning laser ophthalmoscopy will be utilized for detection of apoptosing retinal cells (DARCs). This is a novel imaging technique that allows in vivo human imaging of stressed cells undergoing apoptosis via intravenous administration of fluorescently labelled annexin 5 (ANX776) [18]. It combines the biologic data with an artificial intelligence algorithm to identify disease activity on a cellular level. This assessment will be performed in a subset of the study population. Novai (Novai Ltd., Reading, United Kingdom) will interpret DARC imaging.

Contrast sensitivity (CS) will be measured using the CS function (CSF) via the quantitative CSF (qCSF) method, which utilizes active-learning algorithms to test CS across multiple spatial frequencies on the Manifold Contrast Vision Meter (Adaptive Sensory Technology, San Diego, CA, USA). This novel device uses a computerized Bayesian algorithm to select personalized spatially-filtered optotypes that modulate in contrast and spatial frequency. The patient verbally reports the letters to the experimenter who operates the handheld tablet and records the response. The test estimates a CSF curve based on 25 trials, generating an area under the log CSF curve (AULCSF) which serves as a broad CS metric for analysis [19]. Other metrics generated include low-contrast VA and CS thresholds at multiple spatial frequencies.

Aqueous humor will be collected for pharmacodynamic analyses, with plans to quantify VEGF-A and Ang-2 through enzyme-linked immunosorbent assay. Large-scale protein analysis will be performed by label-free quantification nano liquid chromatography-tandem mass spectrometry as previously described [20, 21]. Changes in intraocular proteins over time will be assessed relative to retinal biomarkers including assessments of disease severity and responses to faricimab treatment.

Rationale for Dosing and Therapeutic Regimens

Group 1 will be dosed Q4W with IVT faricimab 6 mg in order to maximize the opportunity to detect a meaningful impact on the area of RNP over time compared to the untreated control arm, group 2. At week 48, group 1 will transition to dosing Q16W to assess for maintenance of any changes in the trajectory of RNP. Group 2 will allow for an assessment of the natural history of RNP progression over 48 weeks as there will be no scheduled dosing unless rescue criteria were met. At week 48, group 2 will transition to maximal dosing Q4W to validate the effect of faricimab treatment and to maximize the opportunity to identify a biological impact on RNP progression.

Multiple lines of evidence in retinal vascular diseases managed with anti-VEGF therapy indicate that there is a dose-dependent biological impact of such pharmacotherapy on RNP. For example, through 1 year of the RECOVERY prospective, randomized trial, mean RNP was stable among eyes dosed monthly with aflibercept whereas mean RNP increased significantly among eyes dosed with aflibercept quarterly [22]. Given the superior reduction of RNP progression within eyes dosed monthly, the current study is designed with this similar regimen for the periods of maximal dosing.

Faricimab is chosen as the anti-VEGF therapy for this study because of its additional binding region for angiopoietin (Ang)-2 [23]. Similar to VEGF-A, Ang-2 is a cytokine that is pathologically elevated in DR and clinical evidence has found that targeting both VEGF and Ang-2 is potentially more beneficial than solely targeting VEGF [24].

Only eyes with NPDR are included in order to allow for a sham arm to prospectively assess the natural history of RNP. It would not have been possible to incorporate a sham arm in a study that included eyes with PDR since such eyes generally require treatment. Additionally, proactive treatment of NPDR has been shown to reduce the risk of vision-threatening complication development [25, 26]. For example, the PANORAMA randomized clinical trial found that eyes with moderately-severe to severe NPDR treated with anti-VEGF injections had a significantly higher rate of ≥ 2-step DRSS improvement at week 24 and week 100 [25]. Protocol W also found that PDR or center-involving DME developed less frequently in NPDR eyes that were randomized to quarterly aflibercept than sham at both 2-year (27% reduction) and 4-year (23% reduction) endpoints [26].

Early treatment for this current study on RNP is further supported by evidence that patients with DM have areas of RNP prior to the development of clinical DR [2, 3]. For example, a prospective observational study found that patients with DM without DR had larger foveal avascular zones and areas of capillary nonperfusion adjacent to the fovea than age-matched healthy subjects [3]. Another analysis additionally observed decreases in superficial and deep retinal vessel densities in diabetic eyes compared to healthy eyes [27]. The early onset of RNP highlights the importance of early treatment initiation.

Rationale for Choice of Assessments

UWFA is the imaging modality of choice for RNP area assessment to allow visualization of the greatest possible area of retina. Whereas traditional FA normally captures a 20–50° field of the fundus, UWFA extends beyond the equator and can visualize up to 200° to capture areas of peripheral RNP that may otherwise be missed [28]. Peripheral RNP on UWFA is also of clinical value, as it has been significantly correlated with DME and DR progression [5, 29, 30].

OCT-A is chosen to supplement UWFA images because of the well-recognized challenges of identifying and quantifying areas of RNP with FA within eyes following anti-VEGF therapy, as such dramatic reduction in leakage can create substantial changes in contrast and imaging features that make zones of RNP more difficult to detect. Therefore, OCT-A may allow more reliable RNP assessment [31, 32]. Unlike FA, OCT-A can more precisely identify RNP because it is not affected by choroidal fluorescence or dye leakage [33]. OCT-A also offers the advantage of allowing more quantitative measurements, including vessel density and foveal avascular zone areas across different horizontal regions of the retina [34]. However, OCT-A is limited in its ability to assess the mid and far peripheral retina and has been reported to miss detection of blood vessels with flow below a certain velocity threshold, allowing low-flow vasculature to be missed [33, 35]; both OCT-A and UWFA will be utilized due to their complementary ability to assess RNP.

Although high-contrast BCVA remains the standard method for assessment of visual function in the majority of retinal diseases, many patients with DR have visual dysfunction not fully captured by BCVA [36] and may retain “normal” VA with advanced stages of DR [37]. CS testing is included to address an unmet need for additional functional assessments that are able to quantify visual dysfunction among patients with preserved high-contrast VA. CS testing produces a description of the minimum amount of contrast required to differentiate objects from each other and may be a better representative of real-world visual function [38]. Studies have found CS to be a more sensitive test than VA in identifying visual function change, as eyes with DR have been found to have reduced CSF but normal VA [37]. For example, a prospective study compared CS thresholds among 24 asymptomatic patients with type 1 DM without DR and age-matched nondiabetic control subjects. Despite no visual or systemic symptoms within the DM group, the DM eyes were found to have significant losses on dynamic and static CS tests [39]. In MAGIC, the larger sample size will strengthen CS analyses. Additionally, rather than utilizing time-consuming traditional methods (such as the Pelli-Robson chart which has been reported to have low test-retest reliability) [40], MAGIC will utilize qCSF as a more efficient (approximately 3 times faster), sensitive, and reliable test for CS [19, 41, 42].

A sub-study using DARC will be included as a hypothesis-generating analysis regarding the potential role of cellular apoptosis among eyes with RNP and DR. A larger study could be considered further assessing apoptosis in DR progression, depending on the results obtained in the current work.

Aqueous humor will be collected to comprehensively study the protein signature in both the treatment-naïve as well as the post-faricimab treated states, and to evaluate these changes longitudinally. Comprehensive protein analysis would lead to better characterization of DR disease states, and possibly inform novel drug targets.

Screening and Randomization

Prospective participants are first consented for screening, then interviewed to obtain demographic information and medical history including ocular history and current medications. Intraocular pressure of both eyes, ophthalmic examination including dilated ophthalmoscopy and slit lamp examination, BCVA and manifest refraction, FP, UWFA, SD-OCT, OCT-A, blood pressure, and urine pregnancy test (for women of childbearing potential) are also done within 7 days of randomization. Color fundus photograph, UWFA, SD-OCT, and OCT-A images are sent to the CRC.

Once informed consent is completed and all eligibility criteria are met, the study eye is randomized into group 1 or group 2. Eyes are randomized using a centralized list that uses a block randomization scheme with a block size of 6, ensuring equal allocation to the treatment and control groups within each block. The randomization sequence is generated using a computer-based central randomization scheme by Medrio Randomization and Trial Supply Management (Medrio Inc, San Francisco, CA, USA).

Study Outcomes and Rationale

The primary objective of the study is to examine the change in RNP area within and outside the macula over 48 weeks on UWFA images. To accomplish this, group 1 (maximal Q4W therapy) will be directly compared against group 2 (observation) over the first 48 weeks. Secondary objectives include safety and tolerability of 6 mg faricimab, efficacy of 6 mg faricimab compared to observation from baseline through week 48 and week 96 through assessment of visual and anatomic parameters, change in vision from baseline to week 48 and week 96, natural history of RNP through CSF and DARC imaging, and change in total RNP through week 96 using UWFA and OCT-A.

Safety Assessments

To ensure the safety of all patients throughout the study, several safety assessments will be included such as regular ophthalmologic assessment, adverse event monitoring, and protocol-specified laboratory safety tests. Additional safety imaging assessments will be permitted at the principal investigator’s discretion.

All adverse events, including serious adverse events and adverse events of special interest, will be recorded on the adverse event electronic case report form and reported to the sponsor. Adverse events of special interest include cases of potential drug-induced liver injury characterized by an elevated alanine transaminase (ALT) or aspartate transaminase (AST) in combination with either an elevated bilirubin or clinical jaundice, data related to a suspected transmission of an infectious agent by the study drug, and any sight-threatening adverse events (decrease of ≥30 letters in BCVA score lasting more than 1 h, requirement of surgical or medical intervention to prevent permanent loss of sight, or association with severe intraocular inflammation).

Individual occurrences of these events will be evaluated and documented by the study sites. Investigators will use correct medical terminology when reporting adverse events. Safety will be assessed through descriptive summary of ocular and nonocular adverse events, deaths, and ocular assessments. The adverse event severity grading scale for the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE) will be used for assessing adverse event severity [43]. Clinically significant laboratory abnormalities and clinically significant vital sign abnormalities will be reported as adverse events and evaluated as part of the adverse event assessments.

Statistical Methods

The study aims to enroll a total sample size of approximately 150 patients, which was determined to provide more than 80% power to demonstrate a difference in mean change in RNP of approximately 59.1 mm². This was calculated using results from the RECOVERY trial [42], in which subjects with DR were treated Q4W or Q12W with IVT 2 mg aflibercept through 52 weeks: Q4W subjects experienced a mean change in RNP of −12.0 mm² (SD = 129) and Q12W subjects experienced a mean change of 66.8 mm² (SD = 100), resulting in a mean difference of 78.8 mm². Since the RECOVERY trial included subjects in a more severe disease state than MAGIC, MAGIC was predicted to have a smaller response in the treatment group. Thus, the mean difference within MAGIC was estimated to be approximately 75% of that observed in RECOVERY. Considering a mean difference of 59.1 mm² in MAGIC, for 80% power, a sample size of 124 was calculated. To account for loss to follow-up of roughly 20%, the sample size was increased to 150.

Data collected at all follow-ups will be graphically assessed, with statistical analysis conducted at 48- and 96-week time points. The analysis population for the efficacy analyses will consist of all randomized subjects with data available at the given time points, with subjects grouped according to their assigned treatment (intention-to-treat analysis). Significance of all analyses will be determined at p < 0.05.

An intention-to-treat analysis will be conducted for both primary and secondary endpoints. The primary endpoint, defined as the change in area of RNP from baseline through week 48, will be analyzed using multivariable linear regression. The study group will be included in the model as the independent variable with disease severity (defined by baseline BCVA) as a covariate. If baseline characteristics such as age and sex are found to be imbalanced between the two study groups, these variables will be included as covariates in the model to account for potential confounding. Each secondary efficacy will be analyzed in a similar manner with details provided in online suppl. Table S4.

The safety analysis population will include all randomized subjects who received at least one dose of study drug, with subjects grouped according to treatment received. The incidence of all treatment-related adverse events, serious treatment-related adverse events, and adverse events of special interest will be summarized separately for each study group. If the incidence of adverse events is sufficiently large to allow for logistic regression, a logistic regression will be performed. In this model, the adverse event will serve as the dependent variable, the study group as the independent variable, and age, sex, and disease severity as covariates to control for potential confounders, if necessary.

Discussion

MAGIC is a phase 2 trial that has enrolled patients with NPDR to evaluate the natural history of RNP and the impact of faricimab on its progression. The current description of the study design provides the rationale for the unique features of MAGIC, which are informed by the unmet need for a more complete understanding of RNP and associated changes in retinal physiology and function.

Natural history studies are foundational for understanding a disease process and helping advance new therapeutic opportunities. Although studies have considered the presence of RNP in eyes with DR and correlated RNP with various biomarkers, there remains an incomplete understanding regarding the natural history of RNP in DR. Notably, the RIDE/RISE, VISTA, AFFINITY, and RECOVERY studies have reported outcomes of RNP changes longitudinally both in the absence of and following the initiation of anti-VEGF pharmaceutical treatment. However, possibly related to differences in methodology, imaging modality, and study design, results have been in some cases conflicting.

A post hoc analysis of the phase 3 registrational RIDE and RISE (A Study of Ranibizumab Injection in Subject with Clinically Significant Macular Edema with Center Involvement Secondary to Diabetes Mellitus) studies quantified change in area of MNP among 666 patients through 3 years on FA [44]. All eyes had vision loss due to DME with no history of treatment within 3 months of enrollment and were randomized to either IVT ranibizumab Q4W or sham. RNP was assessed using posterior-pole FA imaging without UWF assessment. Although both arms exhibited increase in MNP over 24 months, MNP increased significantly faster in the sham arm with the percentage of patients with no MNP decreasing from 74% to 57%. At 24 months, initiation of ranibizumab 0.5 mg every month corresponded with a reversal of MNP progression and an increase in the percentage of patients with no MNP from 57% to 63% at 36 months. This was the first large, prospectively collected, standardized dataset to illustrate that VEGF blockade may impact MNP development.

Progression of RNP among eyes receiving treatment was further addressed in the VISTA (Longitudinal Retinal Perfusion Status in Eyes with Diabetic Macular Edema Receiving Intravitreal Aflibercept or Laser) trial. A total of 466 patients were randomized to laser, IVT aflibercept Q4W, or IVT aflibercept Q8W after 5 monthly doses and RNP was assessed using posterior-pole FA imaging without UWF assessment. Post hoc analysis revealed that proportionally more eyes experienced improvement in RNP with Q4W/Q8W aflibercept, and proportionally more eyes experienced worsening in RNP within the laser control group; for example, improvement in perfusion was observed in 15% of control patients compared to 40–45% of treated patients and worsening in perfusion was observed in 25% of control patients compared to 9% of treated patients over 2 years [11].

In a smaller study, AFFINITY (Efficacy of Intravitreal Aflibercept Injection for Improvement of Retinal Nonperfusion in Diabetic Retinopathy) prospectively analyzed RNP changes on UWFA among 38 eyes with DR and RNP over 1 year [45]. Twenty fellow eyes that met inclusion criteria did not receive treatments and were analyzed as controls. Through 1 year, no significant increase in the RNP index (total area of RNP divided by total retinal area) was reported among control eyes. Among treated eyes, RNP index directly correlated to treatment initiation and discontinuation; nonperfusion and vascular leakage decreased 6 months after baseline following 6 monthly anti-VEGF injections, then returned to baseline levels following another 6 months of discontinued treatment. Although this prospectively identified an association between treatment administration and RNP, the sample size was limited.

Progression of RNP among eyes receiving treatment was also addressed in RECOVERY (Intravitreal Aflibercept for Retinal Nonperfusion in Proliferative Diabetic Retinopathy). This trial analyzed RNP on UWFA among 40 eyes with PDR that were randomized to either aflibercept Q4W or every Q16W through 1 year, then crossed over from year 1 through year 2. From baseline through year 1, eyes which received aflibercept Q4W had on average a stable amount of RNP compared to the Q16W arm which experienced progressive RNP. However, after crossover from year 1 to year 2, both arms experienced substantial increases in RNP during year 2 [20]. This suggests that frequent, continuous VEGF inhibition may be needed to maximize any beneficial impact on RNP progression. However, the RECOVERY trial investigated PDR, which is a more severe disease state than NPDR and may behave differently in terms of RNP. The RECOVERY trial was also limited to 40 eyes, with only 33 completing year 2.

Both RIDE/RISE and VISTA considered the natural history of RNP and the impact of anti-VEGF therapy but did so post hoc and did not evaluate RNP areas in the far periphery due to the imaging modalities employed. AFFINITY and RECOVERY prospectively utilized UWFA to analyze RNP areas in the peripheral retina, but both were small studies. The findings from AFFINITY, which demonstrated no significant growth in RNP through 1 year without treatment, also conflict with those from RIDE/RISE that reported an increase among the sham eyes over 2 years. Larger, prospective longitudinal studies of the natural history of RNP are needed. The ongoing phase 2 MAGIC trial addresses this unmet need and simultaneously includes treated eyes in the design for direct comparison. The utilization of UWFA and OCT-A will allow evaluation of both the posterior pole and the peripheral retinal circulation. In its unique study design, MAGIC aims to help bring clarity to unanswered questions while investigating novel assessments for RNP that may guide future study designs.

Acknowledgments

MAGIC study group: Charles C. Wykoff, MD, PhD, Dante Pieramici, MD, Justis P. Ehlers, MD, Luis A. Lesmes, PhD, Lasse Jørgensen Cehofski, MD, PhD, SriniVas R. Sadda, MD, Michael Ip, MD, David M. Brown, MD, Avery W. Zhou, MD, Liisa M. Baumann, BS, Gail M. Teagle, BS, Ryan A. Sahraravand, MD, Jeremiah Brown, Jr., MD, Daniel Virgil Alfaro, MD, Sarah E. Holy, MD, Michael J. Borne, MD, Robert W. Wong, MD, William A. Pearce, MD, Matthew A. Cunningham, MD, Effie Z. Rahman, MD, Margaret Chang, MD, Stephen Hypes, DO, Abdish R. Bhavsar, MD, Jesse McCann, MD, PhD, Phillip Ferrone, MD, John Thordsen, MD, Elizabeth Richter, MD, PhD, Kenneth C. Fan, MD, Hasenin Al-Khersan, MD, Cassandra Cone, BS, Deneva Zamora, BS, Varun Chaudhary, MD.

Statement of Ethics

This study is being conducted in accordance with the principles of the Declaration of Helsinki and the Health Insurance Portability and Accountability Act. Written informed consent was obtained before initiation of any study procedures, and the study protocol was approved by the Advarra Institutional Review Board (Columbia, MD, USA) on January 6, 2023 (IRB approval Pro00067762) before study start. The study's clinical trial registration number is identifier NCT05681884 registered with ClinicalTrials.gov. Participant registration took place from May 24, 2023, to May 1, 2024. All adult participants provided written informed consent to participate in this study. This protocol was created in accordance with the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) Statement. The SPIRIT 2025 Checklist is provided in online supplemental material.

Conflict of Interest Statement

Institutional grant from Genentech: Jeremiah Brown, Dante Pieramici, Robert Wong, David Brown, Lasse Cehofski, Michael Ip, Justis Ehlers, Varun Chaudhary, Charles Wykoff; Consulting fees from Genentech: Dante Pieramici, Matthew Cunningham, Margaret Chang, Lasse Cehofski, Michael Ip, SriniVas Sadda, Justis Ehlers, Varun Chaudhary, Hasenin Al-khersan, Charles Wykoff; Payment or honoraria for lectures, presentations, speakers bureaus, manuscript writing or educational events from Genentech: Jeremiah Brown, Matthew Cunningham, Margaret Chang, Lasse Cehofski, SriniVas Sadda; support for attending meetings and/or travel from Genentech: Matthew Cunningham, SriniVas Sadda; participation on a Data Safety Monitoring Board or Advisory Board for Genentech: Jeremiah Brown, Matthew Cunningham, Margaret Chang, Varun Chaudhary; Board member, stock options, and employment in Adaptive Sensory Technology, Inc: Luis A Lesmes.

Funding Sources

Support provided primarily by Genentech Inc, South San Francisco, CA. Funding also was provided in part by the Texas Retina Research Foundation (TRRF), Houston Texas. Funders had no role in preparing, reviewing, or approving the manuscript; and will have no role in the design, data collection, data analysis, and reporting of this study.

Author Contributions

A.W.Z. made substantial contributions to conception and design of the work; drafted the work and reviewed it critically; gave final approval of the version to be published; and agrees to be accountable for all aspects of the work. R.A.S., L.M.B., G.M.T., J.B., D.P., S.E.H., M.J.B., R.W.W., M.A.C., W.A.P., E.Z.R., M.C., A.R.B., D.M.B., A.D.V., K.C.F., L.J.C., M.I., S.R.S, L.A.L., J.P.E., V.C., H.A.K., and C.C.W. made substantial contributions to conception or design of the work; reviewed the work critically; gave final approval of the version to be published; and agree to be accountable for all aspects of the work.

Funding Statement

Support provided primarily by Genentech Inc, South San Francisco, CA. Funding also was provided in part by the Texas Retina Research Foundation (TRRF), Houston Texas. Funders had no role in preparing, reviewing, or approving the manuscript; and will have no role in the design, data collection, data analysis, and reporting of this study.

Data Availability Statement

The data collected in the course of this study are not publicly available due to their containing information that could compromise the privacy of research participants. Further inquiries can be directed to the corresponding author.