Intravitreal bevacizumab for neovascular glaucoma in

Intravitreal bevacizumab for neovascular glaucoma in uveal melanoma treated by proton beam therapy Amir Mahdjoubi, Marie Najean, Stéphanie Lemaitre, Sylvain Dureau, Rémi Dendale, Christine Levy, Livia Lumbroso-Le Rouic, et al. Graefe's Archive for Clinical and Experimental Ophthalmology Incorporating German Journal of Ophthalmology ISSN 0721-832X Graefes Arch Clin Exp Ophthalmol DOI 10.1007/s00417-017-3834-3

1 23

Your article is protected by copyright and all rights are held exclusively by SpringerVerlag GmbH Germany. This e-offprint is for personal use only and shall not be selfarchived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com”.

1 23

Author's personal copy Graefes Arch Clin Exp Ophthalmol https://doi.org/10.1007/s00417-017-3834-3

ONCOLOGY

Intravitreal bevacizumab for neovascular glaucoma in uveal melanoma treated by proton beam therapy Amir Mahdjoubi 1 & Marie Najean 2 & Stéphanie Lemaitre 1 & Sylvain Dureau 2 & Rémi Dendale 3 & Christine Levy 1 & Livia Lumbroso-Le Rouic 1 & Laurence Desjardins 1 & Nathalie Cassoux 1

Received: 2 July 2017 / Revised: 21 September 2017 / Accepted: 16 October 2017 # Springer-Verlag GmbH Germany 2017

Abstract Background To evaluate the efficacy of bevacizumab on reduction of the enucleation rate and control of intraocular pressure (IOP) in neovascular glaucoma (NVG)-complicating proton beam therapy for UM and to identify the determinants of the efficacy of bevacizumab. Methods Retrospective comparative study of patients with rubeosis following proton therapy for uveal melanoma. Patients were divided into two groups: a bevacizumab group and a control group which comprised two subgroups: panretinal photocoagulation (PRP)/cryotherapy and observation subgroups. Bevacizumab was administered by three intravitreal injections at 1-month intervals. A second series of injections was administered when necessary. Data concerning IOP and the secondary enucleation rate were collected and compared between the two groups. Univariate and multivariate analyses were performed to determine predictive factors of response to bevacizumab. Results A total of 169 patients who developed rubeosis following proton therapy between 2006 and 2016 were included: 44 patients in the bevacizumab group and 125 in the control group (38 in the PRP/cryotherapy subgroup and 87 in the observation subgroup). The two groups presented the same baseline characteristics apart from hypertension, retroequatorial site, and proximity of the optic disk, which were more frequent in the control group, while initial retinal

* Amir Mahdjoubi [email protected]

1

Ophthalmology Department, Institut Curie, Paris, France

2

Statistics Department, Institut Curie, Paris, France

3

Radiation Oncology Department, Institut Curie, Paris, France

detachment and larger tumor volume were more frequent in the bevacizumab group. After a mean follow-up of 31 months, IOP was less than 21 mmHg in 54.54% of patients after IVB versus 72.7% before treatment (p = 0.06). Statistical analysis did not reveal any statistically significant reduction of the enucleation rate in the bevacizumab group compared to the observational group, whereas the PRP/cryotherapy group showed better eye retention rate (p = 0.15). No enucleation was performed when IOP was < 21 mmHg before IVB. Multivariate analysis identified initial IOP < 21 mmHg and UM situated away from the macula as predictive factors of good response to bevacizumab. Conclusion Despite the improvement of IOP level, intravitreal bevacizumab (IVB) did not reduce the overall enucleation rate in NVG following proton beam therapy. Nevertheless, this treatment was effective in the early phases of NVG or as preventive treatment. PRP remains a valid treatment for NVG. Keywords Uveal melanoma . Neovascular glaucoma . Bevacizumab . Intravitreal injection . Enucleation . Proton beam therapy

Introduction The management of uveal melanoma constitutes a doublesided problem for specialist ophthalmologists: management of the ocular functional prognosis and management of generally fatal metastatic disease in collaboration with the oncologist. In contrast to enucleation, which was the only available treatment for a long time, local radiotherapy, by proton beam therapy or brachytherapy, ensures local tumor control in about 90% of cases, with preservation of vision in some cases. However, in certain tumor sites or excessively large tumors,

Author's personal copy Graefes Arch Clin Exp Ophthalmol

irradiation induces major complications that may require secondary enucleation [1, 2]. Proton beam therapy has the advantage of generating precise and homogeneous irradiation, therefore delivering the high doses necessary for control of UM while sparing adjacent structures, with 5-year local control rates of 94–97% of cases, depending on the series [3]. However, adverse effects can occur during follow-up [4] that may require secondary enucleation in 4.4 to 11% of cases [3, 5–7]. Neovascular glaucoma (NVG) is one of the major causes of secondary enucleation (30–70%) [5–8]. NVG is due to neovascularization in the iris and anterior chamber angle induced by proangiogenic factors, the best known of which is VEGF (vascular endothelial growth factor). The NVG rate after proton beam therapy varies between 12 and 31%, depending on the series, and predominantly occurs during the first 3 years of follow-up. Several risk factors for NVG have been identified: diabetes, very thick tumors, presence of retinal detachment before treatment, and dose received by the anterior segment, macula, and/or optic nerve [6–11]. The advent of intravitreal anti-VEGF injections constitutes significant progress in the management of neovascular glaucoma, allowing improvement of the anatomical and functional prognosis [12]. A large proportion of the publications concerning the management of complications of proton beam therapy for UM have focussed on the treatment of radiation retinopathy or the prevention of NVG, while few publications have assessed curative treatment of NVG [13]. Since 2010, a new management protocol for NVGcomplicating proton therapy for UM, based on intravitreal bevacizumab (IVB), has been adopted at our institution. The present study evaluated the efficacy of this treatment protocol in terms of reduction of the enucleation rate and lowering of intraocular pressure (IOP) and tried to identify predictive factors of the efficacy of IVB.

patients: the panretinal photocoagulation PRP/ cryotherapy subgroup, and the observation subgroup. All patients were treated for UM by proton beam therapy at the Institut Curie proton therapy centre in Orsay, and were subsequently followed at the Institut Curie ophthalmology department in Paris. Patients who developed clinically visible neovascularization of the iris and/or anterior chamber angle following proton beam therapy, with normal or elevated IOP and a minimum follow-up of 6 months after IVB, were included in this study. Patients treated by other treatment modalities (vitrectomy, trabeculectomy, or cyclophotocoagulation) were not excluded in order to preserve the power of the study. Patients with post-proton beam therapy NVG, but treated by IVB or another treatment in another institution were excluded. Patients enucleated for tumor recurrence were also excluded. Written informed consent was obtained from all patients for each step of management, but formal consent was not required for this retrospective study. This study was approved by our institution’s ethics committee and was conducted in accordance with the Declaration of Helsinki.

Proton beam therapy Surgical tantalum clip insertion was performed in all patients to delineate the tumor margins prior to proton therapy at a dose of 60 Gy RBE (relative biological effectiveness) delivered in four fractions according to the previously described procedure [8]. 3D modulation of the eye and the tumor was performed with EYEPLAN software, based on the position of the clips and the ocular fundus diagram. A 2.5 mm safety margin was applied.

Intravitreal bevacizumab injections

Methods Patients This retrospective, non-randomized, interventional, comparative study reviewed all patients treated by proton beam therapy for uveal melanoma, who subsequently developed rubeosis. The patient cohort was divided into two groups: Bevacizumab group: patients treated with IVB between January 2011 and January 2016. Control group: patients not treated by IVB, in whom NVG was diagnosed between January 2006 and January 2016. This group comprised two subgroups of

IVB injections were performed in a dedicated room during the month following the diagnosis of rubeosis according to the following procedure: after instillation of contact anesthesia, antisepsis was performed with 10% povidone–iodine. The anterior chamber paracentesis was performed when the IOP was considered to be elevated on palpation. A 30-gauge needle was used to inject 0.1 ml of 2.5 mg/ml bevacizumab 4 mm from the limbus, away from the tumor or retinal detachment. Each patient received a cycle of three IVB injections at 28day intervals and were reviewed 1 month after the last injection. Another cycle of three IVB injections was indicated in the case of persistent active NVG, defined by persistent rubeosis with IOP > 21 mmHg.


Follow-up Patients treated for uveal melanoma were followed according to the following plan: 1 month after proton therapy, then every 6 months for 2 years and annually thereafter for a total followup of 10 years. Three-monthly follow-up was ensured in the presence of untreated NVG until stabilization in response to pressure-lowering treatment or an indication for secondary enucleation. Patients treated by IVB were reviewed 1 month after each injection, then every 3 months until stabilization of rubeosis and IOP. Each follow-up examination comprised measurement of visual acuity, determination of intraocular pressure, examination of the anterior segment before dilatation, fundoscopy, and B-mode ultrasound. The presence of any systemic complications was recorded for each patient. PRP or cryotherapy was delivered in two to four sessions according to the patient’s availability and access to the fundus. Enucleation was indicated in the case of elevated IOP causing discomfort to the patient, or progression to phthisis bulbi causing an unsightly appearance or interfering with surveillance of the tumor scar. Statistical analysis Demographic and clinical data were extracted from the PROTON3 database and the patient’s electronic charts. Due to the marked difference in terms of follow-up between the two groups, an artificial date of last follow-up was created in order to obtain similar follow-up in the two groups. This method reduced the difference in follow-up between the bevacizumab and control groups, allowing the two groups to be more accurately compared by analysis of censored data. Categorical variables were compared by Chi-square or Fisher’s exact tests. Enucleation-free interval was defined as the time elapsed between NVG diagnosis and enucleation for any cause. Survival curves according to treatment received or IOP/macula distance were computed using the Kaplan–Meier method, and compared using log-rank tests. Univariate and multivariate Cox proportional hazards models were used to determine the clinical variables associated with enucleation. All analyses were performed using R software (version 3.3.2) (http://cran.r-project.org).

Results Between 2006 and 2016, 230 patients developed rubeosis following proton therapy for UM. Among these 230 patients, 47 received IVB at Institut Curie, while 183 were not treated by IVB in our institution.

Sixty-one patients were excluded from the study: 51 had received IVB in another centre, and another ten patients had received intravitreal injections of another drug (ranibizumab, triamcinolone). A total of 169 patients were included in this study: Bevacizumab group: 44 patients, all of whom received at least one cycle of three IVB injections, while a second cycle of 3 IVB injections was necessary for 6 patients. Control group: 125 patients: – –

PRP/cryotherapy subgroup: 38 patients, 27 of whom were treated by PRP and 11 were treated by cryotherapy. Observation subgroup: 87 patients, seven of whom had received other treatments (cyclophotocoagulation, trabeculectomy, or vitrectomy).

Demographic and clinical data Demographic and clinical data for all patient are reported in Tables 1 and 2. The mean age at diagnosis of uveal melanoma was 61.1 years and the mean age at diagnosis of rubeosis was 63.8 years. A male predominance was observed (59.8% of men for 40.2% of women), but the difference was not statistically significant. No significant difference was observed according to the eye affected: right eye (48.5%) and left eye (51.5%). The mean interval between proton beam therapy and the diagnosis of rubeosis was 31.4 months. Comparison of the two groups by Chi-square test did not reveal any significant difference in terms of age, gender, side, ciliary body involvement, initial exteriorization, tumor diameter, tumor thickness, distance from the macula, presence of initial vitreous hemorrhage, metastases and IOP prior to enucleation. However, a significant difference between the two groups was observed for the following parameters: Control group:higher rates of arterial hypertension (p = 0.03, Chi-square test), UM more situated in retroequatorial segment (p = 0.025, Chi-square test) and closer to the optic disk (p = 0.03, Chi-square test). Bevacizumab group:higher rates of retinal detachment (more than one quadrant) (p = 0.003, Chi-square test) and larger UM volume (p = 0.02, Chi-square test).

No systemic complications related to IVB were observed. Survival analysis The median interval between onset of rubeosis and last follow-up was 30.9 months (range: 10.3–70.3) for the

Author's personal copy Graefes Arch Clin Exp Ophthalmol Table 1

Demographic and clinical data for all patients, control group and bevacizumab group

Gender Affected side Diabetes Hypertension Tumor site Retinal detachment

Ciliary body involvement

Vitreous hemorrhage Metastasis

All patients (169)

Control group (125)

Bevacizumab group (44)

P-value*

Male Female Right eye Left eye

101 (59.8%) 68 (40.2%) 82 (48.5%) 87 (51.5%)

75 (60%) 50 (40%) 59 (47,2%) 66 (52,8%)

26 (59,1%) 18 (40,9%) 23 (52,3%) 21 (47,7%)

0.92

No Yes No Yes Anterior to the equator Posterior to the equator No Yes

152 (89.9%) 17 (10.1%) 105 (62.1%) 64 (37.9%) 83 (49,1%) 86 (50.9%) 78 (46.4%) 90 (53.6%)

112 (89.6%) 13 (10.4%) 72 (57.6%) 53 (42.4%) 55 (44%) 70 (56%) 66 (53.2%) 58 (46.8%)

40 (90.9%) 4 (10.1%) 33 (75%) 11 (25%) 28 (63.6%) 16 (36.4%) 12 (27.3%) 32 (72.7%)

0.97

NA No Yes

1 148 (90.8%) 15 (9.2%)

1 111 (89.5%) 13 (10.5%)

37 (94.9%) 2 (5.1%)

NA No Yes NA No Yes

6 152 (93.3%) 11 (6.7%) 6 146 (86.4%) 23 (13.6%)

1 113 (93.4%) 8 (6.6%) 5 110 (88%) 15 (12%)

5 39 (95.3%) 3 (4.7%) 1 36 (81.8%) 8 (18.2%)

0.56

0.041 0.025 0.003

0.49

0.57

0.30

* chi-square used to compare two groups NA: not available

bevacizumab group. As described in the statistical section, follow-up was homogenized between the two groups and a median follow-up of 31 months (range: 1.1–72.2) was adopted in the control group. Kaplan–Meier survival analysis and log-rank test did not reveal any statistically significant difference between the two groups. Furthermore, due to the competing risk between death and enucleation (a deceased patient was no longer at risk of enucleation), competing risk analysis was performed and also failed to demonstrate any statistically significant difference in terms of enucleation rate between the two groups (Fig. 1). Moreover, when the two subgroups of control patients (PRP/cryotherapy subgroup and observation subgroup) were compared to the bevacizumab group, a trend towards better efficacy was observed in the PRP/cryotherapy subgroup (Fig. 2). The overall three-year enucleation rate was 24.8% [95%CI: 10.0%–37.1%] in the bevacizumab group versus 12.2% [95%CI: 0%–22.8%] in the PRP/cryotherapy subgroup and 32.2% [95%CI: 18.8%–43.3%] in the observation subgroup (p = 0.15, log-rank test). Finally, 32 patients (72.7%) in the bevacizumab group had an IOP > 21 mmHg before treatment versus 24 (54.54%) at last follow-up, reflecting the almost statistically significant efficacy of bevacizumab in terms of improvement of IOP (p = 0.06, Chi-square test).

Univariate and multivariate analyses Univariate and multivariate analyses were performed to identify predictive factors of the efficacy of IVB on the enucleation rate according to IOP and various clinical and demographic parameters. No enucleation was performed when IOP before IVB was < 21 mmHg, while all enucleations were performed in patients with IOP > 21 mmHg. Cox proportional hazards models were performed for the following parameters: age at diagnosis of rubeosis, hypertension, diabetes, tumor site, initial retinal detachment, greatest diameter, thickness, distance to the macula, distance to the optic disk, irradiated surface area of the ciliary body, irradiated volume of the lens, irradiated volume of the optic disk, irradiated volume of the macula, and irradiated length of the optic nerve. Age > 65 years and tumors situated away from the macula were identified as predictive factors of good response to IVB (Table 3). Bivariate analysis identified only 2 predictive factors of good response to IVB: low IOP and tumors situated away from the macula, while age > 65 years was not a predictive factor (Fig. 3).

Discussion The aim of this retrospective interventional study was to assess the efficacy of intravitreal bevacizumab therapy for

Author's personal copy Graefes Arch Clin Exp Ophthalmol Table 2

Demographic and clinical data for all patients, control group and bevacizumab group All patients (169) Mean

Median

Age at diagnosis of UM (years) Age at diagnosis of rubeosis (years)

61.1 63.8

Interval from UM diagnosis to rubeosis diagnosis (months) Largest tumor diameter (mm) Tumor thickness (mm)

Control group (125) NA

Mean

(Median)

62 65

62.1 64.9

31.4

23.7

16.6 7.1

16.4 7.5

11 0

Bevacizumab group (44) NA

P-value*

Mean

Median

62 66

58.4 60.8

58 63.5

0.114 0.078

32.1

26

29.5

21.3

0.66

16.3 7

16.2 7.3

17.2 7.4

17.7 8

7 0

NA

4 0

0.12 0.35

Tumor volume (cc)

1

0.9

11

0.9

0.8

7

1.1

1.2

4

0.02

Distance to the optic disk (mm)

3.9

3.4

22

3.5

2.8

11

5.1

3.9

11

0.03

Distance to the macula (mm)

3.6

3

21

3.5

2.8

10

4

3.4

11

0.43

*Chi-square test used to compare two groups NA: not available

rubeosis complicating proton beam therapy for UM by evaluating the enucleation rate and IOP. The group of patients treated by IVB was then compared with a control group comprising PRP/cryotherapy and observation subgroups. NVG is the result of neovascularization in the iris in response to proangiogenic factors, which, when it involves the anterior chamber angle, induces elevated intraocular pressure. Two main pathophysiological mechanisms are involved. The first mechanism is radiation retinopathy, responsible for retinal ischaemia due to vascular occlusion as a result of X-ray-induced histological changes that promote the production of proangiogenic factors [14]. The second mechanism is toxic tumor syndrome, incriminated by several authors, related to the release, in addition to proangiogenic factors, of proinflammatory humoral factors by the irradiated tumor scar, resulting in expansion of exudative retinal detachment, which accentuates retinal ischaemia and the production of proangiogenic factors [15].

VEGF is the best known and most extensively studied of these proangiogenic factors [16]. Several publications have reported high levels of VEGF-A in the aqueous humor, vitreous, and in several ocular tissues of patients with UM in the absence of irradiation, and these levels increased after radiotherapy even in the absence of NVG [16–18]. A statistically significant positive correlation has been demonstrated between high VEGF levels and larger volume UM with extrascleral extension and/or epithelioid or mixed cellular composition [19]. Furthermore, in a series of eyes enucleated for NVG-complicating UM treated by proton beam therapy, high VEGF levels were observed in the tumor, detached retina, ciliary body, and iris, which were probably released from the posterior segment. No neovascularization was observed in the sector of the iris included in the irradiation field [20].

Fig. 1 Kaplan–Meier survival analysis of enucleation in bevacizumab and control groups

Fig. 2 Kaplan–Meier survival analysis of enucleation in bevacizumab group, PRP/cryotherapy and observation subgroups

Author's personal copy Graefes Arch Clin Exp Ophthalmol Table 3 Cox model analysis used to identify predictive factors of the efficacy of IVB on the enucleation rate

Fig. 3 Survival curve according to intraocular pressure and distance to the macula in the bevacizumab group (Cox proportional hazards model)

Hazard ratio

95% CI

P-value

Age at onset of rubeosis 65 years No hypertension Hypertension

1 0.54 1 0.63

[0.28–1.07]

0.08

[0.30–1.3]

0.21

No diabetes Diabetes Tumor anterior to/on the equator Tumor posterior to the equator No retinal detachment Retinal detachment Proton diameter < median Proton diameter ≥ median

1 0.80 1 1.54 1 0.74 1 0.82

[0.26–2.54]

0.71

[0.80–2.95]

0.19

[0.38–1.43]

0.37

[0.42–1.6]

0.57

Thickness < median Thickness ≥ median

1 1.64

[0.83–3.25]

0.15

Distance to the macula < median Distance to the macula ≥ median

1 0.9 (2.02)

[0.24–1.03]

0.06

Distance to the optic disk < median Distance to the optic disk ≥ median

1 1.15

[0.56–2.33]

0.71

Irradiated surface area of the ciliary body < 50% Irradiated surface area of the ciliary body ≥ 50% Irradiated volume of the lens < 50% Irradiated volume of the lens ≥ 50%

1 0.80 1 0.86

[0.32–2.01]

0.64

[0.42–1.78]

0.69

Irradiated volume of the optic disk < 30% Irradiated volume of the optic disk ≥ 30% Irradiated volume of the macula < 50% Irradiated volume of the macula ≥ 50% Irradiated length of the optic nerve < median Irradiated length of the optic nerve ≥ median

1 1.45 1 1.26 1 1.28

[0.74–2.82]

0.28

[0.646–2.47]

0.5

[0.655–2.5]

0.47


Many publications have reported the use of anti-VEGF in UM. The efficacy and safety of anti-VEGF vary according to the indications. Bevacizumab has not been shown to be effective on control of tumor growth prior to irradiation in several series. In a series of three patients, IVB did not slow the growth of small choroidal lesions with a progressive increase in tumor volume, finally resulting in true UM requiring irradiation [21]. In an in-vitro study, el Filali et al. reported an increase in tumor volume following the use of bevacizumab. This tumor growth appears to be related to the hypoxic conditions induced by bevacizumab, resulting adaptation of tumor cells to ischaemia, and angiogenesis rebound after discontinuation of treatment. This hypothesis is supported by the VEGF peaks observed immediately after stopping bevacizumab [22]. Otherwise, anti-VEGF therapy administered after irradiation of UM to prevent radiation retinopathy was not associated with higher recurrence rates, with variable results in terms of efficacy [13, 23]. Bevacizumab has been shown to be effective in the management of NVG-complicating retinal vein occlusion or diabetes, with better eye-preservation rates by lowering IOP [24]. However, few studies have assessed the efficacy of bevacizumab for the treatment of post-proton therapy NVG. Vasquez et al. successfully used intracameral bevacizumab in one patient, allowing eye preservation and correction of IOP [25]. In a series of four patients treated with one or two IVB, Yeung et al. reported transient regression of rubeosis followed by only one enucleation (and the final IOP was correlated with its initial value)but nevertheless noted a relatively long median interval of 6.2 months between the diagnosis of NVG and initiation of IVB [26]. Finally, in a larger series of 11 patients treated with one to five intravitreal ranibizumab injections, Caujolle et al. reported good results in terms of control of IOP with no enucleations. However, the median follow-up of 11 months was too short to allow definitive evaluation of efficacy [27]. When these treatment modalities failed to control IOP, trabeculectomy appeared to be a good alternative [28]. These results are consistent with those obtained in our study. Despite correction of IOP obtained in almost 20% of patients, bevacizumab failed to reduce the final enucleation rate after 3 years, in spite of the suspensive effect observed during the first year compared to the observation subgroup, while a lower enucleation rate was observed in the PRP/ cryotherapy subgroup. These results are in agreement with those reported by Young et al. The most likely explanation is a transient reduction of VEGF levels without suppression of the retinal or tumor source of production. This hypothesis is supported by the slightly superior efficacy of other treatment modalities (retinal photocoagulation or cryotherapy), which have the advantage of acting directly on retinal ischaemia by inducing permanent destruction of oxygenconsuming retinal tissue.

On univariate analysis, IOP < 21 mmHg before initiation of bevacizumab was found to be a predictive factor of good response, suggesting enhanced efficacy of anti-VEGF therapy when it is initiated at the early phase of rubeosis before the neovascular membrane invades the anterior chamber angle, or even as preventive treatment. Comparison between the two groups demonstrated statistical differences in terms of retinal detachment (RD), tumor volume, location, and distance from the optic disk, which constitute risk factors for NVG. While the presence of a larger tumor volume in the bevacizumab group may lead to NVG, mostly as a result of toxic tumor syndrome, radiation retinopathy could play a role in the pathogenesis of NVG in the control group, in which most tumors were located close to the optic disk. In order to evaluate the impact of this factor, multivariate analysis was performed to assess how these factors may impact the efficacy of IVB. RD, distance from the optic disk, or tumor volume were not identified as reliable predictive factors of the efficacy of IVB on the enucleation rate. However, these findings alone are not wholly sufficient and cannot thoroughly overcome this intrinsic bias of a retrospective study. Several publications have studied preventive treatment of NVG by acting on the sources of production of VEGF prior to the onset of rubeosis. Argon laser photocoagulation is a classical treatment for ischaemic retinopathy. In a series of 14 patients, photocoagulation induced regression of preretinal neovascularization in 92% of patients with proliferative radiation retinopathy [29], while in a larger series, panretinal photocoagulation induced regression of neovascularization in only 66% of cases [30]. However, these results do not concern NVG-complicating irradiated UM, which is due to more complex pathophysiological mechanisms. Transpupillary thermotherapy (TTT), which has not been demonstrated to be as effective as curative treatment for UM, was associated with a decreased incidence of retinal detachment and an accelerated reattachment when applied after irradiation [31]. However, the incidence and severity of NVG and enucleation rates vary depending on the series [32, 33]. Two surgical techniques are used to prevent the development of NVG: drainage of subretinal fluid, and tumor resection. Although good results have been reported for treatment of retinal detachment alone, with no tumor resection procedure after 1 year of follow-up [34], several authors prefer to perform tumor scar resection in order to prevent toxic tumor syndrome. Endoresection via pars plana vitrectomy allows combined treatment of the retinal detachment and tumor resection. It is indicated for tumors > 5 mm and < 15 mm in diameter. Cassoux et al. reported a reduction of NVG and enucleation rates compared to the control arm and the TTT arm [33]. Konstantinidis et al. reported similar results for transscleral resection of the tumor scar [15], but this technique is more complicated and


does not allow drainage of subretinal fluid. In a large-scale study comparing drainage of subretinal fluid and endoresection, Seibel et al. reported a lower 5-year NVG rate in the group treated for retinal detachment, and an even lower NVG rate in the endoresection group compared to the control group. However, the final enucleation rate was comparable for these two groups and 50% lower than the enucleation rate observed in the control group. Several studies have evaluated the preventive use of antiVEGF therapy. Systemic administration of bevacizumab allowed reattachment of the retina in two patients [35]. However, in a prospective study in patients treated by brachytherapy, IVB did not accelerate resorption of subretinal fluid compared to the control group, in contrast with triamcinolone [36]. In another retrospective study, IVB was administered every 2 months for 6 months after proton beam therapy and then every 3 months until reattachment of the retina, followed by photocoagulation of zones of retinal ischaemia. Reattachment of the retina was obtained more rapidly in the IVB group compared to the control group, with a significantly lower NVG rate [37]. The findings of this last study were consistent with those of the present study, which showed that IVB generally did not improve eye preservation rates compared to observational attitude, while PRP or cryotherapy were more effective. These results highlight the importance of PRP in the treatment of radiation retinopathy, as it allows permanent control of VEGF leakage from the ischaemic retina. However, in the presence of a large RD, PRP cannot be performed until spontaneous reattachment of the retina, which can be a long process, allowing the appearance of rubeosis. In the bevacizumab group, we noted a relatively lower rate of enucleation within the first year and, furthermore, the absence of enucleation when IVB was administered at earlier stages of rubeosis. This is a particularly interesting point, as it highlights the role that bevacizumab could play in patients with a large RD by neutralizing VEGF and halting progression of rubeosis until PRP becomes feasible, which eliminates one of VEGF’s sources. In this study, we chose to compare patients treated by IVB with the overall population of patients who developed rubeosis since 2006. Anti-VEGF therapy was first used in treatment of this severe complication of UM treated by radiotherapy at the end of the last decade, and bevacizumab was introduced in our centre for this indication in 2010. Consequently, the majority of patients with rubeosis since this date have been treated by IVB, and only a few patients were managed by observation with no active treatment. Furthermore, due to the characteristics of NVG, it would be particularly difficult to conduct a prospective study, as these patients often present a fragile psychological state and require close tumor scar and liver monitoring and frequently experience painful eye, making randomization and control for all

clinical features more difficult in contrast to studies on NVG prevention strategies. The injection protocol used in our study consisted of three injections at 1-month intervals, with a second series of injections in the case of partial response. More frequent monthly injections are difficult to perform because many patients live a long way from our centre, and bevacizumab is not available in retail pharmacies, while the other anti-VEGF drugs are not reimbursed for this indication. To the best of our knowledge, this study reports the largest cohort with sufficient follow-up. The main limitation of our study concerns its retrospective design, as the populations compared were not followed simultaneously and patient characteristics were not strictly homogeneous. However, rubeosis was only controlled when IVB was initiated early in the course of the disease, before new vessels had reached the anterior chamber angle and before onset of NVG, or possibly when it was administered more frequently or even continuously. Another alternative consists of using bevacizumab as preventive treatment in patients presenting risk factors for NVG in association with other preventive treatment. Furthermore, macular UM was identified as a risk factor for poor efficacy of IVB, which can probably be explained by the severity of NVG-complicating macular UM because of the increased sensitivity of this region to radiation-induced ischaemia related to its high oxygen consumption. Fortunately, UM in this localisation are often diagnosed earlier, which makes it possible, in addition to ensuring a better prognosis for the patient, to avoid treating larger tumor volume, which adds another risk factor for the development of NVG to radiation maculopathy. Finally, while tumor scar resection is one of the main ways to avoid toxic tumor syndrome, PRP constitutes the standard treatment for radiation retinopathy and still has a central place in prevention and treatment strategy for NVG-complicating radiotherapy of UM. Funding No funding was received for this research. Compliance with ethical standards Conflict of interest All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest (such as honoraria; educational grants; participation in speakers’ bureaus; membership, employment, consultancies, stock ownership, or other equity interest; and expert testimony or patent-licensing arrangements), or non-financial interest (such as personal or professional relationships, affiliations, knowledge, or beliefs) in the subject matter or materials discussed in this manuscript. Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its subsequent amendments or comparable ethical standards.

Author's personal copy Graefes Arch Clin Exp Ophthalmol Informed consent For this type of study formal consent is not required. 18.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

Damato B (2012) Progress in the management of patients with uveal melanoma. The 2012 Ashton Lecture. Eye (Lond) 26:1157– 1172. https://doi.org/10.1038/eye.2012.126 Hawkins BS (2011) Collaborative ocular melanoma study randomized trial of I-125 brachytherapy. Clin Trials 8:661–673. https://doi. org/10.1177/1740774511419684 Desjardins L, Lumbroso-Le Rouic L, Levy-Gabriel C et al (2012) Treatment of uveal melanoma by accelerated proton beam. Dev Ophthalmol 49:41–57. https://doi.org/10.1159/000328257 Groenewald C, Konstantinidis L, Damato B (2013) Effects of radiotherapy on uveal melanomas and adjacent tissues. Eye (Lond) 27:163–171. https://doi.org/10.1038/eye.2012.249 Damato B, Kacperek A, Chopra M et al (2005) Proton beam radiotherapy of choroidal melanoma: the Liverpool-Clatterbridge experience. Int J Radiat Oncol Biol Phys 62:1405–1411. https://doi.org/ 10.1016/j.ijrobp.2005.01.016 Egger E, Zografos L, Schalenbourg A et al (2003) Eye retention after proton beam radiotherapy for uveal melanoma. Int J Radiat Oncol Biol Phys 55:867–880 Seibel I, Riechardt AI, Heufelder J et al (2017) Adjuvant ab interno tumor treatment after proton beam irradiation. Am J Ophthalmol. https://doi.org/10.1016/j.ajo.2017.03.027 Dendale R, Lumbroso-Le Rouic L, Noel G et al (2006) Proton beam radiotherapy for uveal melanoma: results of Curie InstitutOrsay proton therapy center (ICPO). Int J Radiat Oncol Biol Phys 65:780–787. https://doi.org/10.1016/j.ijrobp.2006.01.020 Conway RM, Poothullil AM, Daftari IK et al (2006) Estimates of ocular and visual retention following treatment of extra-large uveal melanomas by proton beam radiotherapy. Arch Ophthalmol 124: 838–843. https://doi.org/10.1001/archopht.124.6.838 Foss A, Whelehan I, Hungerford J et al (1997) Predictive factors for the development of rubeosis following proton beam radiotherapy for uveal melanoma. Br J Ophthalmol 81:748–754 Mishra KK, Daftari IK, Weinberg V et al (2013) Risk factors for neovascular glaucoma after proton beam therapy of uveal melanoma: a detailed analysis of tumor and dose-volume parameters. Int J Radiat Oncol Biol Phys 87:330–336. https://doi.org/10.1016/j. ijrobp.2013.05.051 Aref AA (2016) Current management of glaucoma and vascular occlusive disease. Curr Opin Ophthalmol 27:140–145. https://doi. org/10.1097/ICU.0000000000000239 Reichstein D (2015) Current treatments and preventive strategies for radiation retinopathy. Curr Opin Ophthalmol 26:157–166. https://doi.org/10.1097/ICU.0000000000000141 Archer DB, Amoaku WM, Gardiner TA (1991) Radiation retinopathy–clinical, histopathological, ultrastructural and experimental correlations. Eye (Lond) 5(Pt 2):239–251. https://doi.org/10.1038/ eye.1991.39 Konstantinidis L, Groenewald C, Coupland SE, Damato B (2014) Trans-scleral local resection of toxic choroidal melanoma after proton beam radiotherapy. Br J Ophthalmol 98:775–779. https://doi. org/10.1136/bjophthalmol-2013-304501 Ly LV, Bronkhorst IHG, van Beelen E et al (2010) Inflammatory cytokines in eyes with uveal melanoma and relation with macrophage infiltration. Invest Ophthalmol Vis Sci 51:5445–5451. https://doi.org/10.1167/iovs.10-5526 Boyd SR, Tan D, Bunce C et al (2002) Vascular endothelial growth factor is elevated in ocular fluids of eyes harbouring uveal

19.

20.

21.

22. 23.

24.

25.

26.

27.

28.

29. 30.

31.

32.

33.

34.

melanoma: identification of a potential therapeutic window. Br J Ophthalmol 86:448–452 Missotten GSO, Notting IC, Schlingemann RO et al (2006) Vascular endothelial growth factor a in eyes with uveal melanoma. Arch Ophthalmol 124:1428–1434. https://doi.org/10.1001/ archopht.124.10.1428 Yang M, Kuang X, Pan Y et al (2014) Clinicopathological characteristics of vascular endothelial growth factor expression in uveal melanoma: a meta-analysis. Mol Clin Oncol 2:363–368. https://doi. org/10.3892/mco.2014.247 Boyd SR, Gittos A, Richter M et al (2006) Proton beam therapy and iris neovascularisation in uveal melanoma. Eye (Lond) 20:832– 836. https://doi.org/10.1038/sj.eye.6702072 Lima BR, Schoenfield LR, Singh AD (2011) The impact of intravitreal bevacizumab therapy on choroidal melanoma. Am J Ophthalmol 151:323.e2–328.e2. https://doi.org/10.1016/j.ajo. 2010.08.040 el Filali M, Ly LV, Luyten GPM et al (2012) Bevacizumab and intraocular tumors: an intriguing paradox. Mol Vis 18:2454–2467 Shah SU, Shields CL, Bianciotto CG et al (2014) Intravitreal bevacizumab at 4-month intervals for prevention of macular edema after plaque radiotherapy of uveal melanoma. Ophthalmology 121: 269–275. https://doi.org/10.1016/j.ophtha.2013.08.039 Andrés-Guerrero V, Perucho-González L, García-Feijoo J et al (2017) Current perspectives on the use of anti-VEGF drugs as adjuvant therapy in glaucoma. Adv Ther 34:378–395. https://doi.org/ 10.1007/s12325-016-0461-z Vásquez LM, Somani S, Altomare F, Simpson ER (2009) Intracameral bevacizumab in the treatment of neovascular glaucoma and exudative retinal detachment after brachytherapy in choroidal melanoma. Can J Ophthalmol 44:106–107. https://doi.org/10. 3129/i08-171 Yeung SN, Paton KE, Waite C, Maberley DA (2010) Intravitreal bevacizumab for iris neovascularization following proton beam irradiation for choroidal melanoma. Can J Ophthalmol 45:269–273. https://doi.org/10.3129/i09-259 Caujolle JP, Maschi C, Freton A et al (2012) Treatment of neovascular glaucoma after proton therapy for uveal melanomas with ranibizumab injection: preliminary results. Ophthalmic Res 47:57–60 Riechardt AI, Cordini D, Rehak M et al (2016) Trabeculectomy in patients with uveal melanoma after proton beam therapy. Graefes Arch Clin Exp Ophthalmol 254:1379–1385. https://doi.org/10. 1007/s00417-016-3310-5 Kinyoun JL, Lawrence BS (1960) Barlow WE (1996) proliferative radiation retinopathy. Arch Ophthalmol 114:1097–1100 Bianciotto C, Shields CL, Pirondini C et al (2010) Proliferative radiation retinopathy after plaque radiotherapy for uveal melanoma. Ophthalmology 117:1005–1012. https://doi.org/10.1016/j.ophtha. 2009.10.015 Char DH, Bove R, Phillips TL (2003) Laser and proton radiation to reduce uveal melanoma-associated exudative retinal detachments. Am J Ophthalmol 136:180–182 Desjardins L, Lumbroso-Le Rouic L, Levy-Gabriel C et al (2006) Combined proton beam radiotherapy and transpupillary thermotherapy for large uveal melanomas: a randomized study of 151 patients. Ophthalmic Res 38:255–260. https://doi.org/10.1159/ 000094834 Cassoux N, Cayette S, Plancher C et al (2013) Choroidal melanoma: does endoresection prevent neovascular glaucoma in patient treated with proton beam irradiation? Retina 33:1441–1447. https://doi.org/10.1097/IAE.0b013e31827f65c8 Gibran SK, Kapoor KG (2009) Management of exudative retinal detachment in choroidal melanoma. Clin Experiment Ophthalmol 37:654–659. https://doi.org/10.1111/j.1442-9071.2009.02127.x

Author's personal copy Graefes Arch Clin Exp Ophthalmol 35.

36.

Newman H, Finger PT, Chin KJ, Pavlick AC (2011) Systemic bevacizumab (Avastin) for exudative retinal detachment secondary to choroidal melanoma. Eur J Ophthalmol 21:796–801. https://doi. org/10.5301/EJO.2011.6477 Parrozzani R, Pilotto E, Dario A et al (2013) Intravitreal triamcinolone versus intravitreal bevacizumab in the treatment of exudative retinal detachment secondary to posterior uveal melanoma. Am J

37.

Ophthalmol 155:127–133.e2. https://doi.org/10.1016/j.ajo.2012. 06.026 Mantel I, Schalenbourg A, Bergin C et al (2014) Prophylactic use of bevacizumab to avoid anterior segment neovascularization following proton therapy for uveal melanoma. Am J Ophthalmol 158: 693.e2–701.e2. https://doi.org/10.1016/j.ajo.2014.07.002