Influence of Ranibizumab versus laser photocoagulation on radiation retinopathy (RadiRet) - a prospective randomized controlled trial

Radiation retinopathy is an ischemic retinopathy caused by irradiation damage to the retina and choroid. In practice, radiation retinopathy is a common complication following a radiotherapy for intraocular tumors with radiation maculopathy being the leading cause of irreversible vision loss in patients treated for uveal melanoma. Intravitreal injections of anti-VEGF or corticosteroids have been shown to maintain or improve the visual acuity and reduce cystoid macular edema when administered over a long period of time [1‐4].

The clinical appearance of radiation retinopathy mimics some important features of diabetic retinopathy such as exudates, hemorrhages, cotton wool spots, capillary non-perfusion, and the occurrence of macular edema [5‐7].

Due to the clinical and pathophysiological similarities, learnings and techniques from diabetic retinopathy may be transferred. Panretinal laser photocoagulation (PRP) of the ischemic retina is useful in the prophylaxis of proliferative diabetic retinopathy and may stabilize macular edema when intravitreal injections are discontinued [8, 9].

Whether PRP leads to the same effect in radiation retinopathy has not yet been proven in studies.

The aim of the radiation retinopathy (RadiRet) study was to address this gap. Therefore, the RadiRet study was designed to compare monthly ranibizumab injections for a maximum interval of 6 months with focal and PRP in patients diagnosed with radiation retinopathy.

The RadiRet study is hereby the first randomized controlled trial comparing ranibizumab with laser treatment in radiation retinopathy.

The RadiRet study was a therapeutic-exploratory, two-arm, randomized, parallel group, single-masked, active-controlled phase II clinical trial with a follow-up period of 12 months (Fig. 1). It has been registered in the EU Clinical Trials Register, EudraCT Number: 2011-004463-69. Institutional Review Board (IRB)/Ethics Committee approval was obtained. All procedures performed in the study were in accordance with the ethical standards of the institutional research committee and with the 1996 Helsinki declaration. The trial protocol and any amendments were prepared in accordance with the Declaration of Helsinki in the version of October 1996 (48th General Assembly of the World Medical Association, Somerset West, Republic of South Africa). Written informed consent was obtained from all individual participants included in the study.

The primary objective was to investigate the change from baseline in best-corrected visual acuity (BCVA) over 6 months of ranibizumab 0.5 mg (experimental) in comparison to focal laser treatment of the macula and laser treatment of ischemic retina (control) in patients with radiation retinopathy secondary to radiation therapy of uveal melanoma.

The secondary objectives of the RadiRet study were to compare functional and anatomic outcomes between ranibizumab and laser groups over 12 months. The proportion of patients with changes in BCVA after 6 and 12 months and the rate of peripheral ischemia and vitreous hemorrhages were analyzed. As third objective, safety of 0.5 mg ranibizumab was investigated.

Safety endpoints further included local and systemic tumor control, investigation for key arterial thromboembolic events, death, and non-ocular hemorrhage; ocular endpoints comprised intraocular pressure, endophthalmitis, rhegmatogenous retinal detachment, retinal tear, vitreous hemorrhage, lens damage, and signs of ocular inflammation.

The RadiRet study included patients with retinopathy due to radiation of uveal melanoma that presented with radiation maculopathy (visual impairment due to focal or diffuse ME in the irradiated eye that was eligible for laser treatment) and clinical signs of radiation retinopathy, i.e., cotton wool spots, hemorrhages, vascular ischemia. All patients were ≥ 18 years of age, and BCVA was less than 20/32 at the primary visit.

Exclusion criteria were participation in other interventional trials and concomitant conditions in the study eye which in the opinion of the investigator could prevent BCVA improvement, e.g., tumor recurrence, tumor growth underneath the macula, tumor endoresection, and/or previous vitrectomy. Patients with proliferative retinopathies or macular edema due to reasons other than irradiation, e.g., diabetic retinopathy, vein occlusion, or Irvine-Gass syndrome, were excluded as were patients with previous treatment with anti-angiogenic drugs or intravitreal corticosteroids or any other investigational drug within 3 months prior to randomization, or prior laser photocoagulation treatment within 3 months (focal/grid laser) or 6 months (panretinal) prior to study entry. Furthermore, patients with known hypersensitivity against local anesthetics or iodine, patients with anamnestically confirmed stroke or preliminary stages of stroke, or patients with history of myocardial infarction, pregnant or nursing women, and failure to use highly effective contraceptive methods were excluded from the trial.

Assuming a within-group standard deviation of 8.5 letters (calculated based on Finger et al. [1]), a sample size of 27 patients per treatment arm was considered sufficient to detect a clinically relevant difference of 5 letters in the primary variable at a one-sided level of 10% with 80% power (δ/σ ≈ 0.6). Randomization was computer-generated and based on permuted blocks of varying length and stratified by radiation dose to macula and disc. In order to account for stratification and attrition, 27/0.9 = 30 patients were planned to be randomly assigned per treatment arm. Assuming an effect size of 0.6, at least 20 evaluable patients are required to observe an effect in the hypothesized direction with 90% probability. Given the orphan nature of disease, enrollment went expectedly slow. Therefore, and to limit the number of subjects exposed to the trial, the decision was taken to perform a premature final analysis after approximately 50% of the originally planned enrollment target was achieved. At that time, 31 patients were enrolled. This was justified as only 20 evaluable patients are sufficient to observe a treatment effect in the hypothesized direction with a probability of at least 90% (given the expected effect size of 0.6). The study protocol was amended accordingly. The amendment was approved by the local Ethics Committee and competent authorities.

Each monthly visit included refraction, BCVA testing using ETDRS charts, intraocular pressure (IOP), slit lamp examination, and OCT assessment of macular thickness and anatomy. Fundus photography and fluorescein angiography using the Heidelberg retina angiograph [Heidelberg Engineering] were performed at the initial visit and at month 6.

From 10/2013 to 12/2015, 46 patients were screened, thirty-two of which could be randomized. (The study was prematurely terminated after 31 included patients.)

All 15 patients who were randomized to receive ranibizumab started treatment. Out of 17 patients who were randomized to the laser group, only 16 were treated. One patient was excluded from the FAS before administration of first treatment as one inclusion criterion was not fulfilled (diagnosis of radiation retinopathy had to be revised).

Patients randomized to the ranibizumab group received a median of 5 injections (range 4 to 6). All laser group patients received 1 focal treatment. For 10 out of the 16 laser group patients, one additional peripheral laser treatment was applied.

Baseline characteristics are depicted in Table 1. None of the patients presented with rubeosis iridis at baseline.

This randomized controlled trial compared the efficacy and safety of intravitreal ranibizumab treatment with laser treatment in radiation retinopathy secondary to radiation of uveal melanoma.

Ranibizumab-treated patients had significantly better BCVA and showed rapid regression of macular edema. On the other hand, the study also showed that the superior efficacy of ranibizumab is contingent to continuous treatment as after cessation of therapy post week 26, visual acuity decreased and central retinal thickness increased back to baseline.

As in this RadiRet study, it has been shown that determining the right treatment frequency with anti-VEGF injections is a key success factor for maintaining the favorable treatment results [1, 2, 11].

Importantly, the RadiRet study also contributes systematic safety information in a comparative setting. The safety profile of ranibizumab in radiation retinopathy suggests safe use, as reported in other studies [1, 11‐13].

Of importance, we did not see any progression of macular ischemia in fluorescein angiography independent of the treatment group during the course of the study. Several studies had raised the question of anti-VEGF-induced macular ischemia progression. However, due to lack of information from fluorescein angiographies in these studies, a macular ischemia in the absence of visual improvement could not be excluded [2, 12].

Other previous studies with VEGF inhibition in radiation retinopathy focussed on radiation maculopathy prevention [11, 13]. These studies demonstrated that the occurrence of radiation maculopathy can be reduced from 50–68% to 33–40% by preventive injections over 24 months.

If radiation maculopathy nevertheless occurs, therapeutic injections become necessary—the topic of the RadiRet study.

RadiRet’s objective was to demonstrate the superiority of intravitreal ranibizumab therapy over central focal laser in terms of visual acuity over 6 months. The termination of the injection therapy after 6 months was intended to demonstrate the effect of anti-VEGF therapy on peripheral ischemia and its complications including proliferation and hemorrhages. In this study, none of the patients in the anti-VEGF group developed bleeding or neovascularization compared to the laser group in which there was a case of bleeding due to posterior vitreous detachment (without statistical significant difference).

Even if the groups appear to be equal with regard to these points, no conclusion should be drawn due to the short follow-up period.

In general, the incidence of proliferative radiation retinopathy described in the literature is very low, especially given the small size of irradiated tumors included into this study [14, 15].

RadiRet study included only a limited number of patients. In view of the rare disease and in addition to the exclusion criterion of a previous vitrectomy, which is routinely performed due to genetic examination of the tumor, enrolment progressed expectedly slowly. Preceding vitrectomy might alter the quantity of intraocular VEGF and improve oxygen diffusion [16, 17] and was therefore excluded from the study.

The RadiRet study employed a pro re nata (PRN) treatment scheme for ranibizumab. As all patients required ranibizumab re-treatments (4–6 injections) after the initiation with three monthly injections, it could be conceivable that results would have been even better with fixed continuous monthly treatment or treat-and-extend regimen.

Despite the small study population, the RadiRet study was able to demonstrate a promising effect. Thus, a continuation of the study would have entailed ethical conflict potential. It remains to be elucidated how long continuation of anti-VEGF treatment remains to be functionally superior over laser treatment.

RadiRet compared the ranibizumab treatment to laser photocoagulation. Since RadiRet showed that ranibizumab was significantly superior to focal laser up to 6 months, a third, combined group with laser treatment of peripheral ischemia and treatment of macular edema with ranibizumab would be interesting to identify differences in injection frequency and long-term visual outcome.

In conclusion, BCVA was clearly superior after ranibizumab compared to focal and peripheral laser treatment as long as the injections continued. After termination of treatment, there was no BCVA difference between ranibizumab and laser therapy. Since we did not see a development into proliferative radiation retinopathy in any of the groups during the study period, ranibizumab and PRP are considered equivalent in this respect.

Intravitreal injections are the therapy standard to achieve visual acuity improvement or stabilization in patients with radiation maculopathy. While intravitreal injections require good patient adherence, PRP is completed after 1–2 treatments. In future studies, photocoagulation of peripheral ischemia should be investigated in addition to intravitreal injections.