Aflibercept for Radiation Maculopathy (ARM Study): Year-2 Extension of a Prospective Clinical Study

Abstract

Purpose:

This work describes the 2-year results of the Aflibercept for Radiation Maculopathy (ARM) randomized clinical study that evaluated intravitreal vascular endothelial growth factor antagonist therapy in radiation maculopathy delivering aflibercept using a second-year collapsed, every-6-weeks, treat-and-adjust interval.

Methods:

Forty patients were enrolled in an institutional review board–approved clinical trial and randomly assigned to aflibercept treatment via 1 of 2 regimens: fixed, every-6-week treatment or variable, treat-and-adjust therapy centered around 6 weeks. All patients had a diagnosis of treated uveal melanoma with documented tumor control, and they had visually compromising radiation maculopathy. At conclusion of year 1, the first 30 patients were offered a collapsed single-arm variable of an every-6-weeks treat-and-adjust aflibercept injection schedule for an additional treatment year.

Results:

Baseline best-corrected visual acuity (BCVA) was 20/63 at ARM study entry 20/62 at the institution of the year-2 extension. At ARM study entry baseline, spectral domain–optical coherence tomography mean central retinal thickness was 432 µm and was 294 µm at the same institution. At the 2-year study’s conclusion, 76.7% (23 of 30) of eyes were better than 20/50, and only 6.7% (2 of 30) ended with a BCVA below 20/200. Final mean BCVA was 20/62 and final mean spectral domain–optical coherence tomography central retinal thickness was 286 µm, but as in year 1, this reduction in number of injections was not statistically significant.

Conclusions:

Aflibercept is effective in treating radiation maculopathy with maintained visual acuity at 2 years but continues to require an ongoing treatment approach to stabilize radiation maculopathy.

Introduction

Treatment of posterior uveal melanoma has advanced from enucleation to globe-conserving therapies, such as radiation, whenever possible. Radiation treatments, including external beam radiation, plaque brachytherapy, proton beam radiation, helium ion radiotherapy, and Gamma Knife (Elekta, Stockholm, Sweden) radiotherapy, ^{1 -9} may result in long-term damage to the retinal vascular endothelial cells, causing radiation maculopathy in more than 75% of eyes undergoing plaque radiotherapy. ^10,11 Currently, uveal melanomas are routinely treated with either brachytherapy (iodine-125, palladium-103, or ruthenium-106) or charged-particle radiotherapy (proton or helium beam). These treatments achieve local tumor control in more than 95% of eyes and avoid enucleation in more than 98% of eyes. These outstanding treatment results have shifted the focus to preserving visual and anatomic function after targeted radiation therapy. Radiation retinopathy, modulated by vascular endothelial growth factor (VEGF), involves apoptosis, cell migration, clotting cascades, increased vascular endothelial permeability, aneurysm formation, telangiectasia, and neovascularization. ¹²

Additional radiation effects on the eye include cataract formation, radiation optic neuropathy, retinal detachment, neovascular glaucoma, nerve-fiber layer infarcts, retinal hemorrhages, and vitreous hemorrhage. ^{11,13 -15}

Radiation damage can be detected and classified through clinical examination, fluorescein angiography, indocyanine green angiography, or optical coherence tomography (OCT). ^{16 -18} High-resolution OCT classification (either swept source or spectral domain) is now the standard in detecting and monitoring radiation-induced macular edema, and it can be graded on a 1 to 5 scale. ¹⁷ This edema grading scale has also been shown to correlate with foveal thickness and visual acuity (VA). ¹⁷ Recent findings using spectral-domain OCT (SD-OCT) to detect intraretinal cystic spaces and photoreceptor loss have demonstrated that these earliest signs of radiation maculopathy manifest as macular edema, ¹⁹ with average onset of OCT-evident edema by 12 months or up to 5 months earlier than clinically detectable radiation maculopathy and as early as 3 months after radiation treatment. ^{5,19 -23} Radiation retinopathy is present in more than 90% of posterior uveal melanomas treated with radiation therapy. The severity of radiation retinopathy appears dose dependent based on macular total radiation dose. Clinically, tumors with a posterior margin within 3 mm of the macula are noted to have early onset and increased severity of radiation retinopathy.

Recent OCT angiography (OCTA) studies have documented alterations in foveal vasculature that are associated with uveal melanoma. Further, OCTA may detect vascular alterations secondary to radiation retinopathy prior to SD-OCT detection of radiation maculopathy. The impact of OCTA in the detection and monitoring of radiation maculopathy is undergoing further study. The potential reversibility of damage from macular edema can serve as a critical time point for therapeutic intervention. ²³

Anti-VEGF agents have been proposed for the treatment of radiation-related complications. Studies on eyes harboring choroidal melanomas have demonstrated increases in the levels of VEGF both in aqueous and vitreous samples compared with eyes without tumors. ^24,25 Of note, tumors previously treated with radiation therapy displayed the highest VEGF concentrations. ²⁴ These results have been confirmed and have been shown to correlate with larger basal diameter and tumor height. ^26,27 Multiple investigators reported that administration of bevacizumab, aflibercept, or ranibizumab in radiation retinopathy leads to decreased macular edema, improved or maintained VA, and reduced hemorrhage and retinal edema. ^{4,5,23,28 -30} Although anecdotal evidence in small case series has demonstrated the efficacy of anti-VEGF agents for radiation maculopathy, to date there have been no large-case clinical trials evaluating intravitreal aflibercept treatment over multiple years. ³⁰

At the Miami Ocular Oncology and Retina Institute, improvements in best-corrected VA (BCVA) from bevacizumab usage have been observed in nearly half of the patients with radiation maculopathy, especially when commenced at the first signs of visual compromise. The institute’s efforts lay in maintaining VA and globe salvage for patients suffering visually compromising radiation maculopathy. Hence, intravitreal bevacizumab is routinely administered in patients with OCT evidence of radiation maculopathy and a BCVA of 20/25 or worse. In studies, long-term improvements both in BCVA and SD-OCT have been observed. Recent publications have evaluated both the 5- and 10-year outcomes of patients undergoing intravitreal pharmacotherapy for radiation maculopathy. In spite of these reports, many clinicians have remained skeptical of the benefits of treatment of radiation maculopathy. Some reasons for historically poor outcomes might be the delay in the initiation of treatment or the apathetic conviction to an aggressive treatment regimen. Either or both of these are likely to lead to poor anatomic outcomes.

Therefore, the purpose of this study was to evaluate, in a completed first-year, randomized, prospective clinical trial, the 2-year benefits of intravitreal aflibercept on both VA and SD-OCT grading of maculopathy. This second-year extension collapsed the prior 2 randomized treatment arms into a single arm that used intravitreal aflibercept in a treat-and-adjust regimen for continued treatment of patients with SD-OCT radiation maculopathy associated with a decline in BCVA.

Methods

This institutional review board–approved study used a randomized, prospective, 2-arm clinical trial approach to compare intravitreal aflibercept given in fixed 6-week intervals with treat-and-adjust intervals centered on 6 weeks. Inclusion criteria included clinical diagnosis of uveal melanoma, stable posttreatment tumor control, SD-OCT–graded radiation maculopathy, VA of 20/800 or better, and a measurable decline in BCVA by at least 1 Snellen line. All treated tumors were located in the choroid and had an initial presenting apical thickness of at least 2 mm; radiation dosage for both groups was 85 Gy to the apical height of the melanoma. Exclusion criteria included age younger than 18 years, inability to give informed consent, preexisting non–tumor- or treatment-related macular disease, or media opacity precluding imaging of the tumor and macula. Fine-needle aspiration biopsy for molecular genetic analysis was not an inclusion or exclusion criterion.

All patients were treated and followed within the Miami Ocular Oncology and Retina Institute. Each comprehensive clinical examination included BCVA, dilated fundoscopic examination, SD-OCT, and macular and tumor imaging. Quantitative A- and B-scan echography were obtained at baseline and then every 4 months through the treatment trial. Fluorescein angiography and indocyanine green angiography were obtained at baseline, at the midpoint of the study, and at the study’s conclusion. All examinations were masked to treatment arm by the examiner. All images were evaluated by 2 masked examiners.

Study Year 1

Twenty patients in the fixed every-6-weeks treatment group received intravitreal aflibercept 2.0 mg/0.05 mL using a standardized injection protocol. Patients were seen every 6 weeks, underwent comprehensive ophthalmic evaluation, and were reinjected.

Twenty patients in the treat-and-adjust arm received the first intravitreal aflibercept injection of 2.0 mg/0.05 mL and were scheduled back for visit 2 in 6 weeks for the second intravitreal aflibercept injection. At that visit, and all subsequent visits, the patients were treated with intravitreal aflibercept, and the subsequent visit was adjusted by 1-week steps based on the SD-OCT examination. Improvement of SD-OCT–graded maculopathy allowed a 1-week increase of the follow-up interval, whereas a worse SD-OCT grade mandated a 1-week decrease in the interval before the next treatment.

The study design and year 1 outcomes have been reported. ¹

Study Year 2

The Aflibercept for Radiation Maculopathy (ARM) study extension beginning in year 2 is described as follows.

Thirty consecutive patients were collapsed into a single treat-and-adjust study arm and received the first intravitreal aflibercept injection of 2.0 mg/0.05 mL. They were scheduled for visit 2 in 6 weeks for the second intravitreal aflibercept injection. At that and all subsequent visits, the patients were treated with intravitreal aflibercept, and the subsequent visit was adjusted by 1-week steps based on the SD-OCT examination. Improvement of SD-OCT–graded maculopathy allowed a 1-week increase of the follow-up interval, whereas a worse SD-OCT grade mandated a 1-week decrease in the interval before the next treatment.

Radiation maculopathy SD-OCT grading was performed using modification of the Shields grading schedule: grade 1 was extrafoveal noncystoid edema, grade 2 extrafoveal cystoid edema, grade 3 foveal noncystoid edema, grade 4 foveal cystoid edema mild to moderate, grade 5, foveal cystoid edema severe, and grade 6 foveal cystoid edema severe with subretinal fluid. All SD-OCTs were graded in a masked-to-treatment fashion.

Intravitreal aflibercept injections were given in a standardized protocol that included sterile wire lid speculum, povidone-iodine, and topical proparacaine followed by lidocaine gel application using a cotton-tipped swab. Aflibercept 2.0 mg/0.05 mL was injected through the pars plana inferotemporally unless the injection site would pass through the primary melanoma. In these cases, the injection was performed at least 2 clock hours from the treated uveal melanoma. In patients using topical glaucoma medications, the intravitreal injection was followed immediately by therapeutic paracentesis to minimize the time associated with postinjection intraocular pressure elevations.

Results

Thirty of 30 patients (100%) completed the trial with 2-year follow-up. Patient demographics were similar between the every-6-weeks arm and the treat-and-adjust arm (Table 1). Tumor characteristics between the 2 arms were similar. All patients received standardized 125-iodine brachytherapy with intraoperative echographic localization, and all patients had tumor stabilization as an inclusion criterion. Second-year entry and study results are reported in Table 2. Baseline study entry mean BCVA was 20/63 and was maintained at 20/62 at study conclusion at 60 weeks (1 year) and 20/70 at the 2-year conclusion. At baseline, SD-OCT mean central retinal thickness was 423 µm and improved to 294 µm at 60 weeks (P < .02). Final mean SD-OCT central retinal thickness was 286 µm (Figures 1 and 2). At the study conclusion, 76.7% (23 of 30) of eyes were better than 20/50, and only 6.7% (2 of 30) ended with a VA below 20/200.

Table 1.

Patient Demographics (N = 30).

Demographics	Details	Total
Age, y	40-49	4
	50-59	5
	60-69	10
	70-79	10
	80-81	1
Sex	Female	15
	Male	15
Race	White, non-Hispanic	28
	Hispanic	2
Medical history	Hypertension	19
	Diabetes mellitus	4
Brachytherapy to first Eylea (aflibercept), mean, moa		92.7

^a All patients were previously treated with intravitreal bevacizumab and or triamcinolone acetonide.

Table 2.

Aflibercept for Radiation Maculopathy Study Results (N = 30 Patients).

Study endpoints	Results
Second-year entry
VA, mean	20/62
SD-OCT CRT, mean, µm	294
Study conclusion, 2 y
VA, mean	20/70
SD-OCT, CRT mean, µm	286
VA > 20/50	23 (76.7%)
VA < 20/200	2 (6.7%)
VA stable within 3 Snellen lines	27 (90%)
VA improved ≥ 3 lines	1 (3.3%)
VA decreased ≥ 3 lines	2 (6.7%)
Injection No.	7.8

Abbreviations: CRT, central retinal thickness; OCT, optical coherence tomography; SD, spectral domain; VA, visual acuity.

Figure 1.

Mean central retinal thickness over a 2-year period. CMT indicates central macular thickness; OCT, optical coherence tomography; sd, spectral-domain.

Figure 2.

Spectral-domain optical coherence tomography of 3 enrolled individuals during their (A) initial visit, (B) at end of year 1, and (C) end of year 2.

One patient developed an aflibercept postinjection inflammatory response in the initial year, but this did not occur again for this patient, nor for any other study injections (1 of 400 injections, 0.25%). No patients developed radiation optic neuropathy, proliferative retinopathy, or neovascular glaucoma. No patients developed endophthalmitis, and no patients developed recurrent and/or metastatic disease or died during the study window.

Conclusions

Ninety percent of brachytherapy-treated eyes present with radiation retinopathy when evaluated with SD-OCT. Radiation retinopathy has previously been treated only when patients developed proliferative retinopathy, vitreous hemorrhage, or tractional retinal detachment. With newer understanding of the early presentation of radiation retinopathy and the impact it may have on anatomic and visual outcome, many investigators have more recently begun earlier treatments including laser photocoagulation, intravitreal steroids, and/or intravitreal anti-VEGFs. ^{4,5,23,28 -33}

Advanced imaging of postradiation-treated eyes has advanced our understanding of radiation maculopathy. The ability to detect early, clinically significant macular edema was enhanced with SD-OCT and swept-source OCT. Recently OCTA has detected vascular changes that are tumor related and vascular changes that are treatment related. ³⁴ The role of OCTA is still being investigated. Further imaging studies using widefield viewing have documented the extensive presence of ischemia and vascular compromise of radiotherapy-treated uveal melanoma. ^35,36

Treatment of these early radiation maculopathy findings were initially limited to laser applications using focal, sectoral, or direct tumor ablation, or combinations of these approaches. Translational consideration of anti-VEGF mirrored early interest in the treatment of other adult vascular diseases including neovascular age-related macular degeneration, vascular occlusive disease (branch retinal vein occlusion/central retinal vein occlusion), and diabetic macular edema. Lacking any Food and Drug Administration–approved indications for treatment, patients with uveal melanoma-related vascular disease were routinely treated in an off-label delivery that, because of cost considerations, initially focused on intravitreal bevacizumab.

Early studies documented a clear benefit to coupling radiation maculopathy treatment with early detection with SD-OCT and short-interval targeted retreatment using SD-OCT guidance for the delivery of intravitreal bevacizumab. Improvements in central macular thickness and VA in patients treated with intravitreal bevacizumab for radiation retinopathy have been reported; however, these patients received injections every month. ^35,36 Additionally, Finger et al have shown that intravitreal bevacizumab provided long-term stability in VA and macular anatomy for patients with radiation retinopathy. ³⁷ In that study, patients received injections as frequently as every 4 weeks. In our center, best-care practices for patients unresponsive to short-course bevacizumab were to transition to treatment with intravitreal triamcinolone acetonide in a combined approach. ³¹ Multiple investigators have noted marked improvement in retinal anatomy but reported variable improvements in VA. ^32,33 Virtually all of these reports were retrospective and nonstandardized, but the increasing recognition of improved outcomes has slowly led to the incorporation of intravitreal injection therapy from a small handful of ocular oncology centers to virtually all centers in the United States.

The lack of a standardized SD-OCT grading plan, the variability in treatment intervals, and the limited number of injections were cited by many investigators as problematic in achieving best anatomic and visual results for patients with raditation retinopathy. These problems were first addressed in our aflibercept for radiation retinopathy randomized clinical study. In that study, 40 patients were randomly assigned to receive intravitreal aflibercept in either a fixed, every-6-week dosing strategy or a treat-and-adjust strategy focused around 6 weeks. The 1-year ARM report was significant for maintenance of VA, improvement in macular anatomy, and lack of treatment morbidity. In this study, both treatment arms required significant numbers of intravitreal aflibercept injections during the first year of the study and only minimal reduction in injection numbers were seen between the 2 study arms, requiring approximately 1 injection every 6.2 weeks.

This article reports the two-year data from our randomized clinical trial for the use of aflibercept in the treatment of SD-OCT–documented, visually significant radiation maculopathy. This study is important for several reasons: First, for the year 1 ARM study, it used a prospective, randomized, clinical trial approach (with well-balanced patient entry criteria), and compared the 2 most common intravitreal anti-VEGF treatment approaches with either fixed interval (every 6 weeks) or treat-and-adjust (centered on every 6 weeks) injections. Finally, this second-year extension treated all patients using a treat-and-adjust schedule with defined treatment criteria, study evaluation, and follow-up and reports the outcomes at 2 years from the initial study entry for all 30 patients (100%).

It was hoped that the treat-and-adjust arm would allow for fewer intravitreal injections during the 1-year study window, but this treatment reduction (8.4 treat and adjust vs 9.0 every 6 weeks) was not found to be either statistically or clinically significant, and although all patients were collapsed into a treat-and-adjust strategy, only modest reductions in year-2 injection frequency was noted (8.4 injections in year 1 to 7.8 in year 2). Nonetheless, the outcomes for this 2-year extension trial mirrored those of our center and others, with notable improvement both in VA and anatomy at 1 year that is maintained through year 2.

Long-term data are available for treatment over both 5- and 10-year windows that suggest the possibility of long-term anatomic and VA stability. These prior reports have been downplayed in the absence of standardized evaluations and treatment approaches. Many investigators have failed to achieve improved outcomes for their patients, and we believe this is multifactorial but hinges on delayed initial treatment coupled with inadequate treatment regimens that extend patient follow-up beyond the efficacy windows of anti-VEGF therapy. These outcomes correlate well with poor anatomic and visual outcomes for patients treated with anti-VEGF for neovascular age-related macular degeneration or diabetic macular edema with nontargeted extended follow-up.

The best comparator of both VA and secondary enucleation outcomes postbrachytherapy remains the medium tumor trial arm of the Collaborative Ocular Melanoma Study. In that study, mean VA at 36 months was 20/200, and approximately 15% of treated eyes were enucleated. The format of multicentered, randomized clinical trials will remain the criterion standard for level 1 clinical decision making but for an “orphan” disease like uveal melanoma, clinical care will be affected by small, well-designed, investigator-sponsored trials.

This study documents the importance of early treatment but suggests that treatment may be delayed (with close postbrachytherapy follow-up) until radiation maculopathy is noted. Other investigators have also documented ongoing, adjunctive use of anti-VEGF therapy beginning immediately after radiotherapy and continuing in an ongoing manner. One excellent randomized study clearly documented the decline in VA and anatomy after cessation of that study and anti-VEGF treatment. These data, coupled with our 2-year ARM extension, clearly establish the need for early, SD-OCT–targeted, frequent intravitreal treatment to maintain best visual and anatomic outcomes. Finally, no patient required enucleation for complications of radiation treatment of uveal melanoma, and we believe that this was also a benefit of anti-VEGF treatment. These data, along with advances in other areas of anti-VEGF therapy, hopefully will provide a framework for treating patients with radiation maculopathy in both an early-detection framework (SD-OCT) and an aggressive treatment schedule (treat-and-adjust injection schedule centered on every 6 weeks.

Limitations of this study include that this is a single-center report and that only 1 additional year has been reported thus far. We look forward to long-term follow-up to document further sustained anatomic and visual benefits for these complex patients.