Radiation damage to the retina (radiation retinopathy) or optic nerve (radiation optic neuropathy) is usually a delayed onset, slowly progressive occlusive vasculopathy that can cause significant visual morbidity following plaque radiotherapy for choroidal melanoma (CM).1,2 Previous studies have shown microvascular changes demonstrated on optical coherence tomography angiography (OCTA) in the foveal,3–6 parafoveal,7–9 and peripapillary region,10 often preceding the onset of clinically evident radiation retinopathy (CERR) and optical coherence tomography (OCT)-evident cystoid macula edema (CME). These microvascular changes also correlate with radiation dose to these regions10 and final visual acuity.5,10 Previous studies have employed 840 nm spectral domain OCTA (SD-OCTA) with a smaller, more limited scan width of 3 × 3 mm3–7 and 4.5 × 4.5 mm10 allowing determination of foveal changes but little beyond that. Compared with SD-OCTA, swept source OCTA (SS-OCTA) employs a tunable swept laser centered between 1040 and 1060 nm, with less sensitivity roll-off and greater signal penetration than SD-OCTA. Furthermore, SS-OCTA demonstrates faster A-scan acquisition, resulting in a broader, more informative scan of 15 × 9 mm.11
Herein, in this single-center study, we explore quantitative semiautomated analysis of total widefield SS-OCTA (15 × 9 mm) to evaluate retinal microvascular alterations in 105 eyes following plaque radiotherapy of CM.
We reviewed medical records of all patients with CM treated with I-125 plaque radiotherapy [70 Gray (Gy) planned tumor apex dose] and imaged with wide-field 15 x 9 mm SS-OCTA at any time following plaque radiotherapy, from March 1, 2018 to August 31, 2018 at the Ocular Oncology Service of Wills Eye Hospital (Philadelphia, PA). All scans were individually reviewed by investigators (L.S.L. and D.A.C.) for segmentation errors and artifacts. Scans with insufficient image quality (secondary to poor signal strength, segmentation, or motion artifact) were excluded from the study.12 Patients without a normal fellow eye for comparison were excluded. For the purpose of this study, the macula region was defined as a circular area within 3 mm of the foveola. The study adhered to the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of Wills Eye Hospital. Informed consent was obtained from all subjects.
Patient data were reviewed for patient demographics, tumor features at presentation, radiation parameters, postoperative clinical features at the time of SS-OCTA, quantitative OCT and SS-OCTA findings, and clinical outcomes at last follow-up. Patient demographic data collected included age, race, sex, medical history (diabetes, hypertension, hyperlipidemia), and involved eye. Clinical tumor features at presentation included largest basal diameter, thickness, location (superior, inferior, nasal, temporal, macula, juxtapapillary), and distance to the foveola and optic disc. OCT findings in the macula included presence of subfoveal fluid, CME, epiretinal membrane, or retinal atrophy. Radiation parameters included radiation dose (Gy) and dose rate (Gy/hour) to the tumor apex, tumor base, foveola, and optic disc. Adjuvant transpupillary thermotherapy (TTT) for tumor consolidation, sector scatter retinal laser photocoagulation (surrounding the tumor scar for 3 mm on all margins sparing the macular region),13,14 prophylactic intravitreal bevacizumab injections (1.25 mg/0.05 mL at 4-month intervals for a total of 2 years), and therapeutic intravitreal bevacizumab injections for CME (1.25 mg/0.05 mL at 1-month intervals)2 were documented, if performed. CERR was defined by the presence of retinal telangiectasia, microaneurysm, exudation, hemorrhage, nerve fiber layer infarction, vascular sclerosis, or neovascularization on ophthalmoscopic examination. The time from plaque to SS-OCTA, and best-corrected visual acuity at the time of SS-OCTA was recorded. Clinical outcomes at last follow-up, based on the presence or absence of CERR, included best-corrected visual acuity, radiation maculopathy (CME by OCT at any time following plaque radiotherapy) or papillopathy (optic disc edema by OCT at any time following plaque radiotherapy), optic atrophy, retinal or optic disc neovascularization (at any time following plaque radiotherapy), tumor recurrence, enucleation, metastasis, and death.
Image Acquisition Protocol
The food and drug administration-approved Zeiss PlexElite 9000 (PlexElite), software version 18.104.22.16842 (Carl Zeiss Meditec, Dublin, CA), was used to obtain all images in this study, with OCT scans obtained using a swept-source tunable laser centered between 1040 and 1060 nm with an A-scan acquisition rate of 100,000 per second at a 3-mm scanning depth providing transverse resolution of 20 μm and axial resolution in tissue of 6.3 μm (1.95 μm digital). The PlexElite acquires four sequential B-scans at each fixed location over the entirety of the chosen scan size and uses a proprietary Zeiss algorithm named Optical Micro-Angiography to the power of Complex (OMAGc) to analyze both the phase and amplitude data to derive the flow information and create OCTA enface images. The PlexElite utilizes the Zeiss FastTrac real-time eye tracking system to ensure registration of sequential scans and minimize motion and blink artifacts. Resolution of the enface images is 500 × 500 A-scans for the 6 × 6 mm, and 834 × 500 A-scans for the 15 × 9 mm OCTA, respectively. For this study, 15 × 9 mm and 6 × 6 mm scans were performed centered on the foveola. Image segmentation was performed using the built-in software (PlexElite, version 22.214.171.12442). The superficial capillary plexus enface image was automatically segmented with an inner boundary at the internal limiting membrane (ILM) and an outer boundary set at an approximation of the inner plexiform layer (IPL) utilizing the equation ZIPL = ZILM + 70% (TILM-OPL); the deep capillary plexus enface image had the inner boundary at IPL as determined previously and an outer boundary set at an estimation of the outer plexiform layer (OPL) utilizing the equation ZOPL = ZRPEfit – 110 μm. Full retina segmentation had an inner boundary at the ILM to 70 μm above the RPEfit segmentation (ZRPE = ZRPEfit – 70 μm).
For OCT, either the Heidelberg Spectralis, HRA+OCT, version 126.96.36.199 (Heidelberg Engineering, Heidelberg, Germany) or Optovue RTVue XR 100 Avanti, version 2017.0.0.16 (Optovue Inc, Fremont, CA) SD-OCT was used to measure central macular thickness (CMT) with the default macular thickness mapping scans to obtain standard B-scan images of the retina.
SS-OCT-A Quantitative Assessment
The size of the foveal avascular zone (FAZ) was assessed on all the 6 × 6 mm OCTA images using developmental algorithms available via the Zeiss Advanced Retinal Imaging network (Zeiss ARI, Macular Density version 0.7 build RP: 151451 PE: 188.8.131.52840). The perfusion density of the superficial and deep retinal layers was measured separately in each 15 × 9 mm OCTA image. Vascular perfusion density was also measured for each individual plexus, in a peripapillary ring that measured 1 mm in width immediately surrounding the optic disc, and separately in the 2 × 2 mm papillomacular bundle region immediately adjacent to the temporal margin of the optic disc (Fig. 1). Each whole-eye OCTA image set was processed via the PlexElite enface images, which were exported as a bitmap, 1024 × 614 pixel image, with a total pixel count of 628,736. The superficial capillary plexus and deep capillary plexus of the retina enface images were exported and opened in Image J software (National Institutes of Health, Bethesda, MD) for image processing. The images were edited using the Huang auto-threshold.15 Using this technique, any flow above the decorrelation threshold, including the vascular structures, appears bright, whereas areas without perfusion appear dark. The auto-threshold was applied to the whole-scan image before cropping to the 1 mm peripapillary ring and 2 × 2 mm papillomacular bundle area. The total number of pixels selected (·) was then expressed as a percentage (%) of the total number of pixels [x] in the whole image [·/(x) pixels per 15 × 9 mm scan).
Statistical analysis was performed using SPSS Statistics Software Version 22 (IBM, Armonk, NY). Continuous variables were expressed as mean (median, range). We performed 2 different types of analysis: comparison of irradiated eyes (CERR + no CERR) versus contralateral non-irradiated (control) eyes; comparison of irradiated eyes with CERR versus irradiated eyes without CERR. Comparison between groups was performed using student t test and 1-way analysis of variance for continuous variables and Chi-square test (or Fisher exact test when indicated) for categorical variables. Pearson correlation coefficient was used to assess any correlation between continuous variables.
There were 210 eyes of 105 patients imaged with wide-field 15 × 9 mm SS-OCTA, of which 105 eyes were treated with I-125 plaque radiotherapy for CM. At mean time of 49 months (35, 4–297 months) after plaque radiotherapy, there were 52 eyes (50%) with CERR and 53 eyes (50%) without CERR (Fig. 2).
Patient demographics are listed in Table 1. The mean patient age at the time of plaque radiotherapy application was 55 years (58, 18–87), with a predominance for Caucasian race (103, 98%), and a similar proportion of male (50, 48%) and female (55, 52%) patients. No patients had previous retinopathy or received intravitreal anti-vascular endothelial growth factor (VEGF) or corticosteroid injection before plaque radiotherapy. A comparison (CERR vs no CERR) revealed the CERR group with younger age at the time of plaque radiotherapy (51 vs 60 years, P
= 0.001) and no difference in rates of systemic vascular diseases including diabetes mellitus (10% vs 9%, P = 0.99), hypertension (35% vs 40%, P = 0.69), and hyperlipidemia (27% vs 45%, P = 0.07).
Clinical features at presentation are listed in Table 1. The mean largest basal diameter was 10 mm (10, 4–19 mm) and thickness was 4.1 mm (3.5, 1.2–10.8 mm). Most tumors were extramacular, extrapapillary in location (70, 67%), with 21 macular tumors (21, 20%) and 14 juxtapapillary tumors (14, 13%). Mean distance to foveola and optic nerve was 3.5 mm (2.8, 0.0–15.0 mm) and 4 mm (3.0, 0.0–18.0 mm), respectively. A comparison (CERR vs no CERR) revealed no differences in tumor diameter, thickness, quadrant, distance to foveola and optic disc, and macular OCT features.
Radiation parameters are listed in Table 1. The mean tumor apex dose was 71 Gy (70, 70–122 Gy) and base dose was 161 Gy (137, 75–431). The mean foveola dose was 67 Gy (45, 8–485 Gy) and optic disc dose was 38 Gy (29, 9–122 Gy). A comparison (CERR vs no CERR) revealed no difference in radiation dose, dose rate, or adjuvant treatment except that eyes with CERR received more intravitreal bevacizumab injections (11 vs 5 injections, P
Quantitative OCT and SS-OCTA features in irradiated (CERR + no CERR) versus nonirradiated (control) eyes are listed in Table 2. The total group of irradiated eyes (CERR + no CERR) was imaged with SS-OCTA at a mean time of 49 months (35, 4–297 months) after plaque radiotherapy. A comparison [irradiated (CERR + no CERR) vs control] revealed irradiated eyes with worse visual acuity (0.81 vs 0.1 logMAR, Snellen equivalent 20/150 vs 20/25, P = 0.002), greater CMT (313 vs 285 μm, P = 0.01), CME (36% vs 0%, P = 0.01), and (1.4 vs 0.23 mm2, P = 0.01), and reduced capillary vascular density in the superficial and deep plexus of the total wide field (44% vs 47%, P < 0.001, and 47% vs 48%, P < 0.001, respectively), peripapillary region (63% vs 77%, P < 0.001, and 63% vs 72%, P = 0.01, respectively), and superficial plexus of the papillomacular bundle (63% vs 68%, P < 0.001). A comparison (CERR vs control) revealed the CERR group with worse visual acuity (1.0 vs 0.1 logMAR, Snellen equivalent 20/200 vs 20/25, P = 0.004), greater CMT (342 vs 285 μm, P = 0.003), CME (52% vs 0%, P
< 0.001), FAZ (1.7 vs 0.23 mm2, P = 0.03), and reduced capillary vascular density in the superficial and deep plexus of the total wide field (43% vs 47%, P < 0.001, and 46% vs 48%, P = 0.001, respectively), peripapillary region (66% vs 77%, P < 0.001, and 66% vs 72%, P = 0.001, respectively), and papillomacular bundle (60% vs 68%, P < 0.001, and 61% vs 64%, P
= 0.03, respectively). A comparison (no CERR vs control) revealed the no CERR group with worse visual acuity (0.6 vs 0.1 logMAR, Snellen equivalent 20/80 vs 20/25, P < 0.001), greater CME (21% vs 0%, P < 0.001), and reduced capillary vascular density in the superficial plexus (46% vs 47%, P < 0.008) of the total wide-field. A comparison (CERR vs no CERR) revealed the CERR group was imaged with SS-OCTA at a longer time after plaque radiotherapy (59 vs 40 months, P = 0.04), with worse visual acuity (1.0 vs 0.6 logMAR, Snellen equivalent 20/100 vs 20/80, P = 0.002), greater CME (52% vs 21%, P = 0.002), reduced capillary vascular density in the superficial and deep plexus of the total wide field (43% vs 46%, P < 0.006, and 46% vs 48%, P < 0.02, respectively) and peripapillary region (66% vs 74%, P < 0.001, and 66% vs 72%, P < 0.01, respectively), and the superficial plexus in the papillomacular bundle (60% vs 65%, P
= 0.03). The area of the FAZ correlated with visual acuity in the total group of irradiated eyes (CERR + no CERR) (r = 0.33, P < 0.05), and in eyes with no CERR (r = 0.44, P
= 0.001) but not in eyes with CERR (r = 0.24, P = 0.11).
Clinical outcomes at last follow-up are listed in Table 3. At mean post treatment follow-up of 50 months (35, 4–297 months), mean visual acuity was 20/120 (20/60, 20/20–HM). A comparison (CERR vs no CERR) revealed the CERR group with longer follow up (60 vs 40 months, P
= 0.04), worse visual acuity (0.98 vs 0.59, Snellen equivalent 20/190 vs 20/78, P
= 0.02), and higher incidence of radiation maculopathy (60% vs 23%, P
Radiation retinopathy is clinically characterized by the presence of retinal microaneurysm, telangiectasia, exudation, hemorrhage, nerve fiber layer infarction, vascular sclerosis, and neovascularization.16 Radiation maculopathy can be vision-threatening17 with no long-term effective therapy.18 Recently, prophylactic intravitreal anti-VEGF injections have been used in an attempt to prevent post-radiation CME and associated vision loss after plaque radiotherapy of uveal melanoma. Shah et al2 in a 2-year outcome study of 292 patients with uveal melanoma treated with plaque radiotherapy and 6 prophylactic intravitreal bevacizumab injections (1.25 mg/0.05 mL) at 4-month intervals, found (bevacizumab group vs control group) lower rates of OCT-evident macular edema (26% vs 40%, P = 0.004), moderate vision loss (33% vs 57%, P < 0.001), and poor final visual acuity (15% vs 28%, P = 0.004) in eyes receiving bevacizumab injections. In that retrospective study, clinically evident radiation maculopathy occurred in 40 patients (14%), clinically evident radiation papillopathy in 35 patients (12%), and best-corrected Snellen visual acuity of less than 5/200 was seen in 35 patients (12%). Similar benefits have been reported by other groups who have used prophylactic intravitreal injections of anti-VEGF agents following proton beam radiation19 or plaque brachytherapy of different radioactive isotopes.20
Radiation exposure is thought to cause preferential loss of capillary vascular endothelial cells in the retinal arterial circulation as a result of free radical formation,21,22 resulting in increased vascular permeability, microaneurysm formation, and capillary non-perfusion and ischemia. Changes in the choroidal vasculature have also been reported in studies using fluorescein angiography (FA) and indocyanine green angiography.23–25 By histology, early changes occur in the small, deep retinal vessels, followed by involvement of larger vessels later in the course of the disease.21,22 Although FA has been the traditional method used to assess for retinal ischemia and neovascularization, more recently OCTA, a noninvasive OCT-based imaging system that generates 3D angiograms by detection of motion contrast of circulating blood cells without the need for contrast injection, has been increasingly used to image and quantify the retinal microvasculature to a microscopic level.26,27 Given that early radiation retinopathy seems to affect the small vessels of the retina, OCTA is a useful imaging modality for the evaluation of retinal microvascular changes in radiation retinopathy, with capabilities including high resolution, the ability to segment individual capillary plexuses, and quantitative measurements.
In 2016, Shields et al3 reported findings in 65 consecutive patients with CM treated with I-125 plaque brachytherapy using SD-OCTA and 3 × 3 mm scans centered on the fovea. Comparison with the fellow nonirradiated eye demonstrated significant enlargement of the FAZ at the superficial (P < 0.0001) and deep (P < 0.0001) plexus and decreased parafoveal density of both the superficial and deep capillary plexus, even in eyes without clinically-evident radiation maculopathy. Similar results were reported by Sellam et al6 in a small series of 17 patients with small posterior CM treated with proton beam radiation. Another study by Matet et al5 on 93 patients with CM treated with proton beam radiotherapy, demonstrated SD-OCTA structural and microvascular changes to have significant impact on visual acuity, with worse visual acuity associated with a larger FAZ (P < 0.0001), and reduced vascular density at the superficial (P = 0.001) and deep (P < 0.0001) capillary plexus. Evaluation of the peripapillary retinal circulation after plaque radiotherapy for CM by Skalet et al10 in a series of 10 eyes, revealed reduced overall peripapillary capillary density in irradiated eyes compared with the fellow non-irradiated eye (52.9% vs 73.3%, P = 0.004), that inversely correlated with radiation dose to the optic nerve (r = −0.528, P = 0.043) and visual acuity (r = −0.564, p = 0.028).
To the best of our knowledge, this is the first study to demonstrate retinal microvasculature abnormalities after plaque radiotherapy of CM using wide-field SS-OCTA. Our findings are consistent with previous SD-OCTA studies,3–7,10 with quantitative microvascular changes including enlargement of the FAZ and reduced capillary vascular density in the superficial and or deep plexus in irradiated eyes with CM. Our study, however, has furthered this information to the widefield spectrum. We have shown in this report that retinal microvascular abnormalities are detected following plaque radiotherapy for CM in the superficial and/or deep vascular plexuses of the total 15 × 9 mm wide-field, and in 2 structurally and functionally important areas of the retina, the peripapillary region, and the papillomacular bundle, which has not been examined in previous studies.
This study confirms that SS-OCTA microvascular abnormalities are not limited to eyes with CERR and can also be seen in eyes without CERR but with distinct SS-OCTA differences between the 2 groups. Eyes with CERR showed statistically significant enlargement of the FAZ (compared with nonirradiated eyes) (1.7 vs 0.23 mm2, P = 0.03), whereas in eyes with no CERR, enlargement of the FAZ (compared with nonirradiated eyes) was not statistically significant (1.2 vs 0.23 mm2, P = 0.16) (Table 2). This follows the trend seen in the previous study of 65 eyes by Shields et al,3 in which the FAZ was significantly enlarged compared with non-irradiated control eyes, in eyes with and to a lesser extent, in eyes without clinically evident radiation maculopathy (P < 0.001 and P = 0.03, respectively). In the present study, the area of the FAZ correlated with visual acuity in the total group of irradiated eyes (CERR + no CERR) (r = 0.33, P < 0.05), and in eyes with no CERR (r = 0.44, P
= 0.001), but not in eyes with CERR (r = 0.24, P = 0.11), suggesting that other factors such as macular edema, subretinal fluid, and outer retinal atrophy, might also contribute to reduced visual acuity in eyes with CERR.
With regard to capillary vascular density, this study showed the superficial and deep vascular plexus in eyes with CERR (compared with nonirradiated eyes) to be reduced in all regions (total widefield, papillomacular bundle, and peripapillary), whereas eyes without CERR (compared with nonirradiated eyes) only showed reduced capillary vascular density of the superficial plexus of the total widefield. Furthermore, comparison of eyes with and without CERR showed significant differences in vascular density in the superficial and deep vascular plexuses of the total widefield and peripapillary region, and the superficial plexus of the papillomacular bundle. These findings suggest that SS-OCTA retinal microvascular density changes might reflect the progression from no CERR to CERR over time, but longitudinal studies are needed to further explore this.
Although the microvascular changes noted in our study are most likely a direct result of radiation damage to retinal vasculature, the presence of subretinal fluid, macular edema, and tumor location in the macula or peripapillary region could also play some role in these changes.8,9 Furthermore, interventions including adjuvant transpupillary thermotherapy for tumor consolidation, sector scatter retinal laser photocoagulation, and prophylactic or therapeutic intravitreal bevacizumab injections might also be contributary to SS-OCTA changes seen following plaque radiotherapy. Tumor inflammatory growth factors and cytokines associated with tumor regression have also been hypothesized to have an effect on the retinal microvasculature post radiation.28,29
Strengths of this study include the relatively large number of patients with a mean follow-up of approximately 4 years treated at a single institution with fundus photography and SS-OCTA images in all cases. Limitations of this study include its retrospective nature, heterogeneity of the study cohort, variable time between plaque radiotherapy and SS-OCTA imaging, and potential bias introduced by the exclusion of patients who did not have SS-OCTA of sufficient quality for analysis or a normal fellow eye for comparison. As SS-OCTA technology is based upon velocity of flow, more restricted flow may be misrepresented and interpreted as ischemia.27 Furthermore, analysis of the SS-OCTA enface images in ImageJ, while allowing for selection and measurement of discrete areas of the retina (foveal, parafoveal, peripapillary ring, and papillomacular bundle), limits measurement to processed and exported enface images rather than the raw OCTA data. Lastly, the size of the FAZ was calculated using developmental algorithms available via the Zeiss Advanced Retinal Imaging network, which resulted in FAZ size readings including 0 mm2, indicating that such algorithms may not always provide the most reliable measurements.
Wide-field SS-OCTA following plaque radiotherapy for CM demonstrates retinal microvascular abnormalities that are not limited to the fovea, but can also occur in the peripapillary region and papillomacular bundle on wide-field images. Furthermore, microvascular abnormalities are not limited to eyes with CERR but can also be seen in eyes without CERR, with distinct SS-OCTA differences between the 2 groups. SS-OCTA is a useful imaging modality for early detection and monitoring of radiation retinopathy in the post-equatorial wide-field portion of the eye.
1. Gündüz K, Shields CL, Shields JA, Cater J, Freire JE, Brady LW. Radiation retinopathy following plaque radiotherapy for posterior uveal melanoma. Arch Ophthalmol
2. Shah SU, Shields CL, Bianciotto CG, et al. Intravitreal bevacizumab at 4-month intervals for prevention of macular edema after plaque radiotherapy of uveal melanoma. Ophthalmology
3. Shields CL, Say EA, Samara WA, Khoo CT, Mashayekhi A, Shields JA. Optical coherence tomography angiography of the macula after plaque radiotherapy of choroidal melanoma: comparison of irradiated versus nonirradiated eyes in 65 patients. Retina
4. Veverka KK, AbouChehade JE, Iezzi R Jr, Pulido JS. Noninvasive grading of radiation retinopathy: the use of optical coherence tomography angiography. Retina
5. Matet A, Daruich A, Zografos L. Radiation maculopathy after proton beam therapy for uveal melanoma: optical coherence tomography angiography alterations influencing visual acuity. Invest Ophthalmol Vis Sci
6. Sellam A, Coscas F, Lumbroso-Le Rouic L, et al. Optical coherence tomography angiography of macular features after proton beam radiotherapy for small choroidal melanoma. Am J Ophthalmol
7. Say EA, Samara WA, Khoo CT, et al. Parafoveal capillary density after plaque radiotherapy for choroidal melanoma: analysis of eyes without radiation maculopathy. Retina
8. Li Y, Say EA, Ferenczy S, Agni M, Shields CL. Altered parafoveal microvasculature in treatment-naive choroidal melanoma eyes detected by optical coherence tomography angiography. Retina
9. Valverde-Megías A, Say EA, Ferenczy SR, Shields CL. Differential macular features on optical coherence tomography angiography in eyes with choroidal nevus and melanoma. Retina
10. Skalet AH, Liu L, Binder C, et al. Quantitative OCT angiography evaluation of peripapillary retinal circulation after plaque brachytherapy. Ophthalmol Retina
11. Jia Y, Tan O, Tokayer J, et al. Split-spectrum amplitude-decorrelation angiography with optical coherence tomography. Opt Express
12. Say EAT, Ferenczy S, Magrath GN, Samara WA, Khoo CTL, Shields CL. Image quality and artifacts on optical coherence tomography angiography: Comparison of pathologic and paired fellow eyes in 65 patients with unilateral choroidal melanoma treated with plaque radiotherapy. Retina
13. Finger PT, Kurli M. Laser photocoagulation for radiation retinopathy after ophthalmic plaque radiation therapy. Br J Ophthalmol
14. Materin MA, Bianciotto CG, Wu C, Shields CL. Sector laser photocoagulation for the prevention of macular edema after plaque radiotherapy for uveal melanoma: a pilot study. Retina
15. Huang LK, Wang MJJ. Image thresholding by minimizing the measures of fuzziness. Pattern Recog
16. Boldt HC, Melia BM, Liu JC, Reynolds SM, Collaborative Ocular Melanoma Study Group. I-125 brachytherapy for choroidal melanoma photographic and angiographic abnormalities: the Collaborative Ocular Melanoma Study: COMS Report No. 30. Ophthalmology
17. Horgan N, Shields CL, Mashayekhi A, Shields JA. Classification and treatment of radiation maculopathy. Curr Opin Ophthalmol
18. Wen JC, McCannel TA. Treatment of radiation retinopathy following plaque brachytherapy for choroidal melanoma. Curr Opin Ophthalmol
19. Kim IK, Lane AM, Jain P, Awh C, Gragoudas ES. Ranibizumab for the prevention of radiation complications in patients treated with proton beam irradiation for choroidal melanoma. Trans Am Ophthalmol Soc
20. Finger PT, Chin KJ, Semenova EA. Intravitreal anti-VEGF therapy for macular radiation retinopathy: a 10-year study. Eur J Ophthalmol
21. Archer DB, Amoaku WM, Gardiner TA. Radiation retinopathy-clinical, histopathological, ultrastructural and experimental correlations. Eye (Lond)
22. Irvine AR, Wood IS. Radiation retinopathy as an experimental model for ischemic proliferative retinopathy and rubeosis iridis. Am J Ophthalmol
23. Amoaku WM, Lafaut B, Sallet G, De Laey JJ. Radiation choroidal vasculopathy: an indocyanine green angiography study. Eye (Lond)
24. Elmassri A. Radiation chorioretinopathy. Br J Ophthalmol
25. Midena E, Segato T, Valenti M, Degli Angeli C, Bertoja E, Piermarocchi S. The effect of external eye irradiation on choroidal circulation. Ophthalmology
26. Spaide RF, Fujimoto JG, Waheed NK, Sadda SR, Staurenghi G. Optical coherence tomography angiography. Prog Retin Eye Res
27. Kashani AH, Chen CL, Gahm JK, et al. Optical coherence tomography angiography: a comprehensive review of current methods and clinical applications. Prog Retin Eye Res
28. Guyer DR, Mukai S, Egan KM, Seddon JM, Walsh SM, Gragoudas ES. Radiation maculopathy after proton beam irradiation for choroidal melanoma. Ophthalmology
29. Peña LA, Fuks Z, Kolesnick RN. Radiation-induced apoptosis of endothelial cells in the murine central nervous system: protection by fibroblast growth factor and sphingomyelinase deficiency. Cancer Res