Role of Intravitreal Antivascular Endothelial Growth Factor Injections for Choroidal Neovascularization due to Choroidal Osteoma

We treated 26 eyes of 25 young patients having a mean age of 30 years with intravitreal vascular endothelial growth factor (VEGF) inhibitor for choroidal new vessel (CNV) formation overlying choroidal osteoma over a mean follow-up of 26 months. Mean number of injections was 2.4 at 6 months, 3.2 at 12 months, and 5.5 at 24 months. CNV was subfoveal in 14 eyes, juxtafoveal in 5, extrafoveal in 5, and peripapillary in 2. By paired comparison, mean decrease from baseline was 119.7 microns at 6 months (n = 15; P = 0.001), 105.3 microns at 1 year (n = 10; P = 0.03), and 157.6 microns at 2 years (n = 7; P = 0.08). BCVA improved by 3.3 lines at 6 months after therapy (n = 26; P < 0.001), 2.8 lines (n = 20; P = 0.01) at 1 year, and 3.1 lines (n = 13; P = 0.049) at 2 years. We conclude that intravitreal anti-VEGF injections improve vision in majority of eyes with CNV from choroidal osteoma.


Introduction
Choroidal osteoma is a rare ossifying choroidal tumor involving mostly young healthy females in the second decade of life [1][2][3][4][5][6]. The benign mass appears as a deep yellowish lesion with distinct geographic borders at the juxtapapillary or macular region, with branching "spider" vessels on its surface. The diagnosis is confirmed with the presence of calcification on ultrasonography and computed tomography. Vision is often compromised by gradual atrophy of the overlying retina [6] or by accumulation of either subretinal fluid or subretinal hemorrhage with or without choroidal neovascularisation (CNV). Frequent exams are recommended for patients with choroidal osteoma for early detection of a subretinal neovascular membrane and potential treatment. Therapies have included laser photocoagulation [7,8], excision of CNV [9], photodynamic therapy (PDT) [10][11][12][13][14], and transpupillary thermal therapy (TTT) [15,16].

Materials and Methods
We reviewed retrospectively the files of subjects having choroidal osteoma who were treated with intravitreal injections of bevacizumab or ranibizumab for active CNV. Intravitreal injections of 0.05 mL or 0.1 mL of either bevacizumab (25 mg/mL) or ranibizumab (10 mg/mL) were administered in the office as 3 initial consecutive doses or based on OCT response depending on physician preference. Intravitreal injection was performed using a 30-gauge needle in a sterile manner after topical anesthesia and povidone instillation in the lower conjunctival sac. Bevacizumab (Avastin, Genentech Inc, San Francisco, CA) aliquots were prepared in the hospital pharmacies of the corresponding institutions. Ranibizumab (Lucentis, Genentech Inc, San Francisco, CA) was also used in some centers. A standardized spreadsheet was used to collect the clinical data. Photodynamic therapy (PDT) with intravenous verteporfin (standard dose 6 mg/m 2 body surface area or half the standard dose) (Visudyne, Novartis AG, Basel, Switzerland) was administered simultaneously in some patients based solely on individual physician preference. Institutional review board/ethics committee approval and patients' signed informed consents were obtained for this study. In addition, this study has been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki for research involving human subjects. The offlabel use of both drugs and their potential risks and benefits was discussed extensively with all patients.
Best corrected visual acuity (BCVA) was assessed using either ETDRS or Snellen charts and listed as logarithm of the minimum angle of resolution (logMar) equivalents. Retreatment was done when there was recurrent activity evaluated by fundus examination, fluorescein angiography (leakage, growth of CNV), or optical coherence tomography. Differences between final and initial BCVA were tested using paired Student -test. Improvement of visual acuity was defined as any fraction of a line of improvement on the ETDRS chart. We did not compare the initial to the posttreatment central foveal thickness because of the different OCT machines among centers as well as because of the need for thickness analysis by gender, race, age, and refractive status [17]. We analyzed only the absolute decrease in central foveal thickness. One patient had bilateral osteoma with CNV arising in one eye and several years later in the fellow eye and hence both eyes were included in the statistical analysis. Significance was set at the 0.05 level. We used SPSS 19 version for statistical calculation (IBM, Armonk, New York, 2010). Literature review till April 2014 (using both PubMed and Google Scholar) was added to ascertain the visual results in this rare disease entity with bevacizumab or ranibizumab therapies [18][19][20][21][22][23][24][25][26][27][28][29]. Collaborators and one of us (AMM) measured the CNV size on digital fluorescein transit films and the osteoma basal diameter on color films in reference to the horizontal disc diameter by using a transparent reticule or ruler on the computer screen.

Discussion
Visual impairment in eyes with choroidal osteoma can be attributable to several factors including subfoveolar location, foveal edema, photoreceptor degeneration [6], serous retinal detachment, and CNV [3]. Without any therapy, choroidal osteoma-associated CNV can result in a progressive and permanent loss of visual acuity. CNV occurs in 31% to 47% of patients followed for 10 years [3,6].
The cause for the development of a CNV in eyes with a choroidal osteoma has not been determined, but choroidal osteomas with overlying hemorrhage or irregular surface appeared at higher risk of developing a CNV [3]. Osteomas, in general, have a high bone turnover reflecting their high metabolic rate and hence may steal blood supply from adjacent tissues, especially overlying retinal pigment epithelium which may upregulate the expression of VEGF. Presumably, attenuation and disruption of the retinal pigment epithelium-Bruch membrane complex overlying the choroidal osteoma allows over years for the development of CNV.
There is no standard treatment for a choroidal osteoma. Various treatments for CNV have been tried, but they do not usually halt visual loss. The results of one study showed that photocoagulation of an extrafoveal classic CNV was successful in closing CNV in 25% of treated eyes [2]. However, photocoagulation can stimulate rapid vascular remodeling and anastomoses in choroidal osteoma [3]. Another study reported that the CNV can be surgically removed, but  Journal of Ophthalmology 7 the postoperative visual acuity was poor at 6/95 (20/320) [4]. PDT has been partially successful in treating CNV in eyes with choroidal osteoma. Earlier studies showed that, 6 months after a single PDT, the metamorphopsia can resolve completely; in one study, the visual acuity was not changed [5] and in another study it improved from 6/60 (20/200) to 6/6 (20/20) [6]. Another study reported that four PDT applications led to closure of the CNV, but the final vision was 6/35 (20/120) [7]. PDT has successfully caused closure of extrafoveal CNV in choroidal osteoma [6]. Laser photocoagulation or PDT in choroidal osteomas with or without CNV may carry the risk of decalcification of choroidal osteoma [10]. Shields et al. [10] proposed that photodynamic therapy could be a therapeutic modality for choroidal neovascularization and induction of decalcification of extrafoveal osteoma to prevent tumor growth into the foveola. However, these results should not be extrapolated to subfoveal choroidal osteoma because decalcification of subfoveal choroidal osteoma could result in worse visual acuity because of loss of retinal pigment epithelium and choroidal perfusion [1,3,6,10]. TTT was effective in obstructing the CNV but the visual outcome was also poor. An earlier report indicates that, at 10 months after one TTT application, vision improved from 6/24 (20/80) to 6/18 (20/60), and the vision was maintained with a scarred CNV [8]. In another report, the final visual acuity was 6/60 (20/200) after three TTT applications [9]. Combination therapy of PDT and anti-VEGF injections reduced the number of anti-VEGF injections, hence reducing the risk of vision-threatening complications. The reduction in the number of injections was marginal in the current series ( Table 2) and there was little difference between the use of PDT or its omission, although a direct comparison could not be done because of the small number in the current series as well as difference in protocols in administration of both anti-VEGF agents and PDT.
Ahmadieh and Vafi [18] reported the dramatic response of a juxtafoveal CNV associated with choroidal osteoma to a single intravitreal injection of bevacizumab in a 19-yearold female with visual improvement from 6/60 (20/200) to 6/7.5 (20/25) and resolution of metamorphopsia with the treatment effect persisting during a 9-month followup period. Subsequently, the positive effects of intravitreal anti-VEGF injections were confirmed in 10 cases with CNV associated with choroidal osteoma (Table 4). Kubota-Taniai et al. [25] described the long-term effect of 2 intravitreal bevacizumab injections (4 months apart) in complete angiographic regression of CNV for a period of 4 years in a 12year-old girl with visual improvement from 6/30 (20/100) to 6/9 (20/30). We noted similar response in 8 out of 26 eyes needing only 1 to 2 injections with maintenance of visual improvement. It is possible that small osteomas or osteoma that do not grow do not need further injections. This issue was not studied in the current protocol. The growth of the tumor during followup was not assessed also and it could be that growing tumors secrete more VEGF and require more injections. The young age of the patients with osteoma may partly explain the good response to anti-VEGF therapies. A single case had retinal pigment epithelial tear after anti-VEGF injection [29].

Conclusions
The inherent limitations of this study include its small number, retrospective nature, limited follow-up, lack of a standard therapeutic protocol, lack of a control group, and difference in OCT machines among centers. However, our results suggest that intravitreal bevacizumab or ranibizumab might be an effective therapeutic option for choroidal osteoma-associated CNV that is causing deterioration in vision, particularly when the CNV is juxtafoveolar or subfoveolar in location. In eyes where the CNV is not subfoveolar, adjunctive therapies such as laser photocoagulation or PDT could be considered. Further studies with longer follow-up are indicated to confirm the long-term efficacy of bevacizumab or ranibizumab monotherapy in choroidal osteomas.