Early Subclinical Macular Edema in Eyes with Uveal Melanoma: Association with Future Cystoid Macular Edema Arman Mashayekhi, MD, Etienne Schönbach, MD, Carol L. Shields, MD, Jerry A. Shields, MD Purpose: To determine the frequency of early subclinical macular edema in eyes with uveal melanoma and its association with future cystoid macular edema (CME). Design: Retrospective cohort study. Participants: A total of 306 patients with uveal melanoma; 260 patients had follow-up of 1 or more years after plaque radiotherapy (follow-up cohort). Methods: Review of medical records and spectral-domain optical coherence tomography (OCT) images. Main Outcome Measures: Frequency of early subclinical macular edema (increased central macular thickness of >10 mm without cystoid changes before or at 4 months after plaque radiotherapy); rate of future CME. Results: At baseline, 164 patients (54%) had subclinical macular edema in the involved eye. On multivariate analysis, factors associated with subclinical macular edema at baseline were increasing tumor diameter (P ¼ 0.001), increasing tumor thickness (P ¼ 0.010), and subretinal fluid (P ¼ 0.001). Of 260 patients in the followup cohort, 105 (40%) developed CME during a median follow-up of 31 months (mean, 34; range, 12e70 months). Eyes with subclinical macular edema at baseline (and at 4 months after plaque radiotherapy) had a significantly higher rate of future CME (n ¼ 66; 50%) compared with eyes without subclinical macular edema at baseline (n ¼ 39; 30%) (P ¼ 0.005; hazard ratio, 1.77; 95% confidence interval, 1.19e2.64). On multivariate analysis, the factors associated with future development of CME included female gender (P ¼ 0.004), increasing tumor thickness (P < 0.001), decreasing tumor distance to foveola (P ¼ 0.002), hemorrhage over tumor (P ¼ 0.017), and increased CMT of >10% at baseline in the involved eyes compared with the opposite eyes (P ¼ 0.012). Conclusions: Subclinical macular edema is common in eyes with uveal melanoma before and at 4 months after plaque radiotherapy and is associated with initial larger tumor size. Eyes with early subclinical macular edema are at significantly higher risk for future CME. These findings suggest that tumor-related factors, most likely mediated through proinflammatory cytokines, may play an important role in development of post-radiation CME. Ophthalmology 2015;122:1023-1029 ª 2015 by the American Academy of Ophthalmology. Supplemental material is available at www.aaojournal.org.
Cystoid macular edema (CME) is a common cause of visual impairment after plaque radiotherapy of uveal melanoma.1,2 In a study on 135 patients with uveal melanoma followed for a median of 24 months after I-125 plaque radiation, Horgan et al1 found evidence of macular edema by timedomain optical coherence tomography (OCT) in 70% of treated eyes. The mean time to onset of macular edema in their study was 12 months, 5 months shorter than the mean time to onset of clinical radiation maculopathy. The aim of the study by Horgan et al1 was to identify early changes in macular morphology after plaque radiotherapy of uveal melanoma. However, because of the limited resolution of time-domain OCT, the researchers were not able to accurately and reliably measure early and subtle changes in macular thickness before the development of OCT-evident CME. Compared with time-domain OCT with an axial resolution of 10 mm or more, spectral-domain 2015 by the American Academy of Ophthalmology Published by Elsevier Inc.
OCT has a higher resolution down to 3 to 5 mm.3 Since March of 2007, all patients with uveal melanoma managed in our center have been evaluated with spectraldomain OCT before treatment and at each visit after plaque radiotherapy. By taking advantage of the higher resolution of spectral-domain OCT, we performed this study to measure early increase in central macular thickness (CMT) in eyes with uveal melanoma before and after plaque radiotherapy and to determine whether such early changes in CMT are predictive of future development of CME.
Methods We reviewed the medical records of patients with the diagnosis of uveal melanoma who were treated with I-125 plaque radiotherapy at the Ocular Oncology Service, Wills Eye Hospital between April 2007 and January 2012. Patients with the following conditions http://dx.doi.org/10.1016/j.ophtha.2014.12.034 ISSN 0161-6420/15
1023
1024
Table 2. Subclinical Macular Edema in Eyes with Uveal Melanoma: Eye and Tumor Features at Initial Visit
All N (%) N ¼ 306
No Subclinical Macular Edema N (%) N ¼ 142
Subclinical Macular Edema N (%) N ¼ 164
All N (%) N ¼ 260
No Cystoid Macular Edema N (%) N ¼ 155
Cystoid Macular Edema N (%) N ¼ 105
273 27 5 1
(89) (9) (2) (0)
133 6 3 0
(94) (4) (2) (0)
140 21 2 1
(85) (13) (1) (1)
239 15 5 1
(92) (6) (2) (0)
142 9 3 1
(92) (6) (2) (1)
97 6 2 0
(92) (6) (2) (0)
128 90 46 8 34 0 7 50
(42) (29) (15) (3) (11) (0) (2) (16)
61 35 27 1 18 0 5 17
(43) (25) (19) (1) (13) (0) (4) (12)
67 55 19 7 16 0 2 33
(41) (34) (12) (4) (10) (0) (1) (20)
110 80 37 3 30 0 5 38
(42) (31) (14) (1) (12) (0) (2) (15)
57 53 24 2 19 0 3 20
(37) (34) (15) (1) (12) (0) (2) (13)
53 27 13 1 11 0 2 18
(50) (26) (12) (1) (10) (0) (2) (17)
96 60 65 74 10 1
(32) (20) (21) (24) (3) (1 Year of Follow-up (n [ 260)
Patients Examined at Baseline (n [ 306)
Feature
Table 2. (Continued.) Patients with >1 Year of Follow-up (n [ 260)
Patients Examined at Baseline (n [ 306)
Feature All N (%) N ¼ 306 8 (3) (87) (12) (1) (62) (12) (3) (40) (9) (3)
125 14 3 78 22 5 61 9 4
(88) (10) (2) (55) (16) (4) (43) (6) (3)
4 (2) 141 22 1 113 16 5 62 18 4
(86) (13) (1) (69) (10) (3) (38) (11) (2)
All N (%) N ¼ 260 2 (1) 224 33 3 167 34 8 106 21 8
No Cystoid Macular Edema N (%) N ¼ 155 1 (1)
(86) (13) (1) (64) (13) (3) (41) (8) (3)
133 19 3 92 18 5 60 6 1
(86) (12) (2) (59) (12) (3) (39) (4) (1)
Cystoid Macular Edema N (%) N ¼ 105 1 (1) 91 14 0 75 16 3 46 15 7
(87) (13) (0) (71) (15) (3) (44) (14) (7)
Table 3. Subclinical Macular Edema in Eyes with Uveal Melanoma: Central Macular Thickness Measurements at Initial Visit All Patients Examined at Baseline (n [ 306)
Feature All N (%) N ¼ 306 OCT machine Optovue (Optovue Inc., Fremont, CA) Spectralis (Heidelberg Engineering GmbH, Heidelberg, Germany) CMT in involved eye (mm) Median (mean, range) CMT in opposite eye (mm) Median (mean, range)
No Subclinical Macular Edema N (%) N ¼ 142
Subclinical Macular Edema N (%) N ¼ 164
Patients with ‡1 Year of Follow-up (n [ 260) All N (%) N ¼ 260
No Cystoid Macular Edema N (%) N ¼ 155
Cystoid Macular Edema N (%) N ¼ 105
233 (76)
106 (75)
127 (77)
195 (73)
108 (68)
87 (79)
73 (24)
36 (25)
37 (23)
65 (25)
47 (30)
18 (17)
278 (285, 201e702)
265 (268, 201e355)
289 (299, 214e702)
278 (280, 201e382)
276 (279, 201e382)
282 (282, 219e356)
263 (266, 164e382)
266 (270, 217e382)
260 (262, 164e353)
265 (267, 189e382)
269 (268, 194e346)
260 (265, 189e382)
CMT ¼ central macular thickness; OCT ¼ optical coherence tomography.
Subclinical Macular Edema and Uveal Melanoma
*Location of tumor epicenter. y Some tumors involved >1 uveal tissue.
266 36 4 191 38 10 123 27 8
4 (3)
Subclinical Macular Edema N (%) N ¼ 164
Mashayekhi et al
Data not available Tumor color Pigmented Nonpigmented Data not available Subretinal fluid Drusen Hemorrhage over tumor Orange pigment Bruch’s membrane rupture Retinal invasion
No Subclinical Macular Edema N (%) N ¼ 142
1025
Ophthalmology Volume 122, Number 5, May 2015 were excluded from the study: (1) CME at baseline (5 patients); (2) epiretinal membrane involving the central macula; (3) subretinal fluid involving the central macula; and (4) media opacity precluding acquisition of good-quality OCT images. Of the total of 306 remaining patients, 260 had follow-up of 1 or more years after plaque radiotherapy (follow-up cohort). Patients were followed every 4 months during the first 2 years after plaque radiation and every 6 months afterward. In addition to complete eye examination, patients underwent spectral-domain OCT at baseline and each follow-up visit. Institutional review board approval was obtained from Wills Eye Hospital. For the purpose of this study, subclinical macular edema was defined as increased CMT of >10 mm compared with the opposite eye without OCT-evident cystoid changes. The specific aims of this study were to (1) determine the frequency of subclinical macular edema in eyes with uveal melanoma (involved eyes) at baseline (before plaque radiotherapy) compared with the opposite eyes; (2) determine the factors associated with subclinical macular edema at baseline in eyes with uveal melanoma; (3) determine whether eyes with uveal melanoma and subclinical macular edema at baseline are at higher risk for future development of CME; (4) determine the frequency of subclinical macular edema in eyes with uveal melanoma at 4 months after plaque radiotherapy compared with the opposite eyes; (5) determine whether eyes with uveal melanoma and subclinical macular edema at 4 months after plaque radiotherapy are at higher risk for future development of CME; and (6) determine the factors associated with development of OCT-evident CME after plaque radiotherapy of uveal melanoma. The following data from initial evaluation were extracted from the medical records of the patients: age (years), race, gender, medical history (diabetes mellitus, systemic hypertension), prior cataract surgery, symptoms, duration of symptoms (months), affected eye (right, left), initial Snellen visual acuity, anteroposterior location of tumor epicenter, quadratic location of tumor epicenter, uveal tissue involved by tumor (iris, ciliary body, choroid), tumor diameter (millimeter), tumor thickness (millimeter), tumor distance to optic disc and foveola (millimeter), Bruch’s membrane rupture, tumor shape, retinal invasion by tumor, hemorrhage over tumor, vitreous hemorrhage, orange pigment, drusen, subretinal fluid, extent of subretinal fluid, diabetic retinopathy, diabetic macular edema, and intravenous fluorescein angiography findings. Treatment parameters studied include plaque size (millimeters) and radiation dose and dose rate to the tumor apex, tumor base, fovea, optic disc, and lens. Fine-needle aspiration biopsy of tumor, prophylactic injection of peribulbar steroid before onset of CME, prophylactic injection of intravitreal bevacizumab before onset of CME, panretinal photocoagulation before onset of CME, and use of oral sunitinib malate before onset of CME were recorded. The OCT features recorde include OCT machine used (Spectralis; Heidelberg Engineering GmbH, Heidelberg, Germany) and RTVue100 (Optovue Inc., Fremont, CA), CMT of involved and opposite eyes (microns), difference between CMT of involved and opposite eyes (CMT in the involved eye minus CMT in the opposite eye) (microns), percent difference between CMT of involved and opposite eyes (%), cystoid changes at fovea, extent of CME, configuration of CME, subfoveal fluid, epiretinal membrane, and vitreofoveal traction. The interval between plaque radiotherapy and onset of CME (month) was recorded. Post-radiation complications recorded include radiationinduced maculopathy, radiation-induced papillopathy, and major branch retinal vein occlusion. Information on treatment for postradiation CME was extracted from the charts (intravitreal bevacizumab injection, peribulbar steroid injection, intravitreal
1026
triamcinolone injection, macular photocoagulation, and panretinal photocoagulation). Information from the last visit included followup duration (interval between plaque radiation and final visit) (month), final Snellen visual acuity, final CMT in the involved and opposite eyes, tumor recurrence, enucleation, distant metastasis from uveal melanoma, and death.
Statistical Methods Data were summarized as number and percentages for the categoric variables and as mean, median, and range for those measured on a continuous scale. Patient, eye, and tumor features at baseline associated with subclinical macular edema were analyzed using logistic regression analysis. The factors predictive of development of CME after plaque radiotherapy were analyzed using Cox proportional hazard model. The factors found significant on univariable analysis at 5% level of significance were considered for multivariable analysis using forward stepwise method. The factors significant at 0.05 level on multivariable analysis were reported. The hazard ratios accompanied by their 95% confidence intervals were presented.
Results All Patients Examined at Baseline (n [ 306 Patients) Table 1 (available at www.aaojournal.org) shows the demographic features, Table 2 shows the eye and tumor features, and Table 3 shows the CMT data of all 306 patients examined at baseline (further classified on the basis of the presence or absence of subclinical macular edema). Overall, 157 patients (51%) were male and 297 patients (97%) were white. The median age of patients was 60 years (mean, 60 years; range, 13e92 years). Diabetes mellitus was present in 43 patients (14%), and systemic hypertension was present in 119 patients (39%). At the initial visit, 173 patients (57%) were asymptomatic, and the most common symptoms were light flashes/floaters (22%) and blurred vision (11%) (Table 1) (available at www.aaojournal.org). As shown in Table 2, median tumor basal dimension and thickness were 12 mm (mean, 12 mm; range, 4e21 mm) and 4.0 mm (mean, 4.8 mm; range, 1.2e12.5 mm), respectively. Subclinical macular edema at baseline (before plaque radiotherapy) in the involved eyes. At baseline, the median CMT was 278 mm (mean, 285; range, 201e702 mm) in the involved eyes (n ¼ 306) and 263 mm (mean, 266; range, 164e382 mm) in the opposite eyes (n ¼ 306). Subclinical macular edema was present in 164 (54%) of the involved eyes at baseline (Table 3). Factors associated with subclinical macular edema at baseline in the involved eyes. Multivariate analysis showed the following factors to be associated with subclinical macular edema at baseline in the involved eyes: (1) tumor diameter (P ¼ 0.001); (2) tumor thickness (P ¼ 0.010); and (3) presence of subretinal fluid (P ¼ 0.001) (Table 4).
Patients with Follow-up of >1 Year after Plaque Radiotherapy (Follow-up Cohort) (n [ 260 Patients) Table 1 shows the demographic features, Table 2 shows the eye and tumor features, Table 3 shows the CMT data, Table 5 (available at www.aaojournal.org) shows the treatment parameters, and Table 6 (available at www.aaojournal.org) shows the complications and final outcome of the 260 patients in the follow-up cohort (further classified on the basis of future
Mashayekhi et al
Subclinical Macular Edema and Uveal Melanoma
Table 4. Subclinical Macular Edema in Eyes with Uveal Melanoma: Patient, Eye, and Tumor Features at Baseline Associated with Subclinical Macular Edema Variable Univariable Analysis Ocular symptoms (Blurred vision vs. no*) (Visual field defect vs. no*) (Flashes/floaters vs. no*) Visual acuity (20/50e20/150 vs. 20/20e20/40*) Lens status (Mild cataract vs. moderate cataract*) (Severe cataract vs. moderate cataract*) Anteroposterior tumor location (Macula to equator vs. macula*) Tumor diameter (mean) Tumor thickness (mean) Subretinal fluid (yes vs. no*) Subretinal fluid extent (1 quadrant vs. no*) (2 quadrants vs. no*) Subretinal fluid distance from foveola (>3 mm vs. no*) Multivariable Analysis Tumor diameter (mean) Tumor thickness (mean) Subretinal fluid (yes vs. no*)
No Subclinical Macular Edema N (%) N [ 142
Subclinical Macular Edema N (%) N [ 164
10 (7) 4 (3) 24 (17)
24 (15) 14 (9) 43 (26)
6 (4)
P Value
Odds Ratio
95% CI
0.004 0.009 0.003
3.21 4.68 2.40
(1.45e7.12) (1.48e14.8) (1.34e4.29)
21 (13)
0.012
3.33
(1.30e8.49)
35 (25) 1 (1)
55 (34) 7 (4)
0.030 0.039
2.23 9.95
(1.08e4.61) (1.13e87.6)
86 (61) 10.2 3.9 78 (55)
108 (66) 12.8 5.5 113 (69)
0.045 15 vs. 15 mm*) Percentage difference in CMT between involved eye and opposite eye at baseline (mean) Percentage difference in CMT between involved eye and opposite eye at baseline (>10% vs. 10% mm*) Multivariable Analysis Gender (female vs. male*) Tumor thickness (mean) Tumor distance to foveola (mean) Hemorrhage over tumor (yes vs. no*) Percentage difference in CMT between involved eye and opposite eye at baseline (>10% vs. 10% mm*)
130 10 3 92
(84) (6) (2) (59)
90 12 5 75
P Value
Risk Ratio
95% CI
1.52
(1.03e2.24)
1.84 1.10y 1.18y 1.05z
(1.03e3.27) (1.04e1.17) (1.09e1.28) (1.004e1.10)
(87) (12) (5) (71)
0.029 0.014 0.006 0.041
4.77 6.60 3.59 1.56
(1.17e19.4) (1.47e29.6) (1.44e8.91) (1.02e2.39)
11 (7) 2 (1)
17 (16) 5 (5)
0.006 0.003
2.32 4.24
(1.27e4.23) (1.63e11.1)
61 (40) 11.2
58 (55) 17.0
0.022 0.006
1.69 1.12x
(1.08e2.64) (1.03e1.22)
65 (42)
66 (63)
0.005
1.77
(1.19e2.64)
49 (32) 3.6
50 (48) 5.8
0.004 0.004
1.75 1.49x
(1.19e2.58) (1.14e1.94)
19 (12)
30 (29)
Cystoid macular edema (CME) is a common cause of visual impairment after plaque radiotherapy of uveal melanoma.1,2 In a study on 135 patients with uveal melanoma followed for a median of 24 months after I-125 plaque radiation, Horgan et al1 found evidence of macular edema by timedomain optical coherence tomography (OCT) in 70% of treated eyes. The mean time to onset of macular edema in their study was 12 months, 5 months shorter than the mean time to onset of clinical radiation maculopathy. The aim of the study by Horgan et al1 was to identify early changes in macular morphology after plaque radiotherapy of uveal melanoma. However, because of the limited resolution of time-domain OCT, the researchers were not able to accurately and reliably measure early and subtle changes in macular thickness before the development of OCT-evident CME. Compared with time-domain OCT with an axial resolution of 10 mm or more, spectral-domain 2015 by the American Academy of Ophthalmology Published by Elsevier Inc.
OCT has a higher resolution down to 3 to 5 mm.3 Since March of 2007, all patients with uveal melanoma managed in our center have been evaluated with spectraldomain OCT before treatment and at each visit after plaque radiotherapy. By taking advantage of the higher resolution of spectral-domain OCT, we performed this study to measure early increase in central macular thickness (CMT) in eyes with uveal melanoma before and after plaque radiotherapy and to determine whether such early changes in CMT are predictive of future development of CME.
Methods We reviewed the medical records of patients with the diagnosis of uveal melanoma who were treated with I-125 plaque radiotherapy at the Ocular Oncology Service, Wills Eye Hospital between April 2007 and January 2012. Patients with the following conditions http://dx.doi.org/10.1016/j.ophtha.2014.12.034 ISSN 0161-6420/15
1023
1024
Table 2. Subclinical Macular Edema in Eyes with Uveal Melanoma: Eye and Tumor Features at Initial Visit
All N (%) N ¼ 306
No Subclinical Macular Edema N (%) N ¼ 142
Subclinical Macular Edema N (%) N ¼ 164
All N (%) N ¼ 260
No Cystoid Macular Edema N (%) N ¼ 155
Cystoid Macular Edema N (%) N ¼ 105
273 27 5 1
(89) (9) (2) (0)
133 6 3 0
(94) (4) (2) (0)
140 21 2 1
(85) (13) (1) (1)
239 15 5 1
(92) (6) (2) (0)
142 9 3 1
(92) (6) (2) (1)
97 6 2 0
(92) (6) (2) (0)
128 90 46 8 34 0 7 50
(42) (29) (15) (3) (11) (0) (2) (16)
61 35 27 1 18 0 5 17
(43) (25) (19) (1) (13) (0) (4) (12)
67 55 19 7 16 0 2 33
(41) (34) (12) (4) (10) (0) (1) (20)
110 80 37 3 30 0 5 38
(42) (31) (14) (1) (12) (0) (2) (15)
57 53 24 2 19 0 3 20
(37) (34) (15) (1) (12) (0) (2) (13)
53 27 13 1 11 0 2 18
(50) (26) (12) (1) (10) (0) (2) (17)
96 60 65 74 10 1
(32) (20) (21) (24) (3) (1 Year of Follow-up (n [ 260)
Patients Examined at Baseline (n [ 306)
Feature
Table 2. (Continued.) Patients with >1 Year of Follow-up (n [ 260)
Patients Examined at Baseline (n [ 306)
Feature All N (%) N ¼ 306 8 (3) (87) (12) (1) (62) (12) (3) (40) (9) (3)
125 14 3 78 22 5 61 9 4
(88) (10) (2) (55) (16) (4) (43) (6) (3)
4 (2) 141 22 1 113 16 5 62 18 4
(86) (13) (1) (69) (10) (3) (38) (11) (2)
All N (%) N ¼ 260 2 (1) 224 33 3 167 34 8 106 21 8
No Cystoid Macular Edema N (%) N ¼ 155 1 (1)
(86) (13) (1) (64) (13) (3) (41) (8) (3)
133 19 3 92 18 5 60 6 1
(86) (12) (2) (59) (12) (3) (39) (4) (1)
Cystoid Macular Edema N (%) N ¼ 105 1 (1) 91 14 0 75 16 3 46 15 7
(87) (13) (0) (71) (15) (3) (44) (14) (7)
Table 3. Subclinical Macular Edema in Eyes with Uveal Melanoma: Central Macular Thickness Measurements at Initial Visit All Patients Examined at Baseline (n [ 306)
Feature All N (%) N ¼ 306 OCT machine Optovue (Optovue Inc., Fremont, CA) Spectralis (Heidelberg Engineering GmbH, Heidelberg, Germany) CMT in involved eye (mm) Median (mean, range) CMT in opposite eye (mm) Median (mean, range)
No Subclinical Macular Edema N (%) N ¼ 142
Subclinical Macular Edema N (%) N ¼ 164
Patients with ‡1 Year of Follow-up (n [ 260) All N (%) N ¼ 260
No Cystoid Macular Edema N (%) N ¼ 155
Cystoid Macular Edema N (%) N ¼ 105
233 (76)
106 (75)
127 (77)
195 (73)
108 (68)
87 (79)
73 (24)
36 (25)
37 (23)
65 (25)
47 (30)
18 (17)
278 (285, 201e702)
265 (268, 201e355)
289 (299, 214e702)
278 (280, 201e382)
276 (279, 201e382)
282 (282, 219e356)
263 (266, 164e382)
266 (270, 217e382)
260 (262, 164e353)
265 (267, 189e382)
269 (268, 194e346)
260 (265, 189e382)
CMT ¼ central macular thickness; OCT ¼ optical coherence tomography.
Subclinical Macular Edema and Uveal Melanoma
*Location of tumor epicenter. y Some tumors involved >1 uveal tissue.
266 36 4 191 38 10 123 27 8
4 (3)
Subclinical Macular Edema N (%) N ¼ 164
Mashayekhi et al
Data not available Tumor color Pigmented Nonpigmented Data not available Subretinal fluid Drusen Hemorrhage over tumor Orange pigment Bruch’s membrane rupture Retinal invasion
No Subclinical Macular Edema N (%) N ¼ 142
1025
Ophthalmology Volume 122, Number 5, May 2015 were excluded from the study: (1) CME at baseline (5 patients); (2) epiretinal membrane involving the central macula; (3) subretinal fluid involving the central macula; and (4) media opacity precluding acquisition of good-quality OCT images. Of the total of 306 remaining patients, 260 had follow-up of 1 or more years after plaque radiotherapy (follow-up cohort). Patients were followed every 4 months during the first 2 years after plaque radiation and every 6 months afterward. In addition to complete eye examination, patients underwent spectral-domain OCT at baseline and each follow-up visit. Institutional review board approval was obtained from Wills Eye Hospital. For the purpose of this study, subclinical macular edema was defined as increased CMT of >10 mm compared with the opposite eye without OCT-evident cystoid changes. The specific aims of this study were to (1) determine the frequency of subclinical macular edema in eyes with uveal melanoma (involved eyes) at baseline (before plaque radiotherapy) compared with the opposite eyes; (2) determine the factors associated with subclinical macular edema at baseline in eyes with uveal melanoma; (3) determine whether eyes with uveal melanoma and subclinical macular edema at baseline are at higher risk for future development of CME; (4) determine the frequency of subclinical macular edema in eyes with uveal melanoma at 4 months after plaque radiotherapy compared with the opposite eyes; (5) determine whether eyes with uveal melanoma and subclinical macular edema at 4 months after plaque radiotherapy are at higher risk for future development of CME; and (6) determine the factors associated with development of OCT-evident CME after plaque radiotherapy of uveal melanoma. The following data from initial evaluation were extracted from the medical records of the patients: age (years), race, gender, medical history (diabetes mellitus, systemic hypertension), prior cataract surgery, symptoms, duration of symptoms (months), affected eye (right, left), initial Snellen visual acuity, anteroposterior location of tumor epicenter, quadratic location of tumor epicenter, uveal tissue involved by tumor (iris, ciliary body, choroid), tumor diameter (millimeter), tumor thickness (millimeter), tumor distance to optic disc and foveola (millimeter), Bruch’s membrane rupture, tumor shape, retinal invasion by tumor, hemorrhage over tumor, vitreous hemorrhage, orange pigment, drusen, subretinal fluid, extent of subretinal fluid, diabetic retinopathy, diabetic macular edema, and intravenous fluorescein angiography findings. Treatment parameters studied include plaque size (millimeters) and radiation dose and dose rate to the tumor apex, tumor base, fovea, optic disc, and lens. Fine-needle aspiration biopsy of tumor, prophylactic injection of peribulbar steroid before onset of CME, prophylactic injection of intravitreal bevacizumab before onset of CME, panretinal photocoagulation before onset of CME, and use of oral sunitinib malate before onset of CME were recorded. The OCT features recorde include OCT machine used (Spectralis; Heidelberg Engineering GmbH, Heidelberg, Germany) and RTVue100 (Optovue Inc., Fremont, CA), CMT of involved and opposite eyes (microns), difference between CMT of involved and opposite eyes (CMT in the involved eye minus CMT in the opposite eye) (microns), percent difference between CMT of involved and opposite eyes (%), cystoid changes at fovea, extent of CME, configuration of CME, subfoveal fluid, epiretinal membrane, and vitreofoveal traction. The interval between plaque radiotherapy and onset of CME (month) was recorded. Post-radiation complications recorded include radiationinduced maculopathy, radiation-induced papillopathy, and major branch retinal vein occlusion. Information on treatment for postradiation CME was extracted from the charts (intravitreal bevacizumab injection, peribulbar steroid injection, intravitreal
1026
triamcinolone injection, macular photocoagulation, and panretinal photocoagulation). Information from the last visit included followup duration (interval between plaque radiation and final visit) (month), final Snellen visual acuity, final CMT in the involved and opposite eyes, tumor recurrence, enucleation, distant metastasis from uveal melanoma, and death.
Statistical Methods Data were summarized as number and percentages for the categoric variables and as mean, median, and range for those measured on a continuous scale. Patient, eye, and tumor features at baseline associated with subclinical macular edema were analyzed using logistic regression analysis. The factors predictive of development of CME after plaque radiotherapy were analyzed using Cox proportional hazard model. The factors found significant on univariable analysis at 5% level of significance were considered for multivariable analysis using forward stepwise method. The factors significant at 0.05 level on multivariable analysis were reported. The hazard ratios accompanied by their 95% confidence intervals were presented.
Results All Patients Examined at Baseline (n [ 306 Patients) Table 1 (available at www.aaojournal.org) shows the demographic features, Table 2 shows the eye and tumor features, and Table 3 shows the CMT data of all 306 patients examined at baseline (further classified on the basis of the presence or absence of subclinical macular edema). Overall, 157 patients (51%) were male and 297 patients (97%) were white. The median age of patients was 60 years (mean, 60 years; range, 13e92 years). Diabetes mellitus was present in 43 patients (14%), and systemic hypertension was present in 119 patients (39%). At the initial visit, 173 patients (57%) were asymptomatic, and the most common symptoms were light flashes/floaters (22%) and blurred vision (11%) (Table 1) (available at www.aaojournal.org). As shown in Table 2, median tumor basal dimension and thickness were 12 mm (mean, 12 mm; range, 4e21 mm) and 4.0 mm (mean, 4.8 mm; range, 1.2e12.5 mm), respectively. Subclinical macular edema at baseline (before plaque radiotherapy) in the involved eyes. At baseline, the median CMT was 278 mm (mean, 285; range, 201e702 mm) in the involved eyes (n ¼ 306) and 263 mm (mean, 266; range, 164e382 mm) in the opposite eyes (n ¼ 306). Subclinical macular edema was present in 164 (54%) of the involved eyes at baseline (Table 3). Factors associated with subclinical macular edema at baseline in the involved eyes. Multivariate analysis showed the following factors to be associated with subclinical macular edema at baseline in the involved eyes: (1) tumor diameter (P ¼ 0.001); (2) tumor thickness (P ¼ 0.010); and (3) presence of subretinal fluid (P ¼ 0.001) (Table 4).
Patients with Follow-up of >1 Year after Plaque Radiotherapy (Follow-up Cohort) (n [ 260 Patients) Table 1 shows the demographic features, Table 2 shows the eye and tumor features, Table 3 shows the CMT data, Table 5 (available at www.aaojournal.org) shows the treatment parameters, and Table 6 (available at www.aaojournal.org) shows the complications and final outcome of the 260 patients in the follow-up cohort (further classified on the basis of future
Mashayekhi et al
Subclinical Macular Edema and Uveal Melanoma
Table 4. Subclinical Macular Edema in Eyes with Uveal Melanoma: Patient, Eye, and Tumor Features at Baseline Associated with Subclinical Macular Edema Variable Univariable Analysis Ocular symptoms (Blurred vision vs. no*) (Visual field defect vs. no*) (Flashes/floaters vs. no*) Visual acuity (20/50e20/150 vs. 20/20e20/40*) Lens status (Mild cataract vs. moderate cataract*) (Severe cataract vs. moderate cataract*) Anteroposterior tumor location (Macula to equator vs. macula*) Tumor diameter (mean) Tumor thickness (mean) Subretinal fluid (yes vs. no*) Subretinal fluid extent (1 quadrant vs. no*) (2 quadrants vs. no*) Subretinal fluid distance from foveola (>3 mm vs. no*) Multivariable Analysis Tumor diameter (mean) Tumor thickness (mean) Subretinal fluid (yes vs. no*)
No Subclinical Macular Edema N (%) N [ 142
Subclinical Macular Edema N (%) N [ 164
10 (7) 4 (3) 24 (17)
24 (15) 14 (9) 43 (26)
6 (4)
P Value
Odds Ratio
95% CI
0.004 0.009 0.003
3.21 4.68 2.40
(1.45e7.12) (1.48e14.8) (1.34e4.29)
21 (13)
0.012
3.33
(1.30e8.49)
35 (25) 1 (1)
55 (34) 7 (4)
0.030 0.039
2.23 9.95
(1.08e4.61) (1.13e87.6)
86 (61) 10.2 3.9 78 (55)
108 (66) 12.8 5.5 113 (69)
0.045 15 vs. 15 mm*) Percentage difference in CMT between involved eye and opposite eye at baseline (mean) Percentage difference in CMT between involved eye and opposite eye at baseline (>10% vs. 10% mm*) Multivariable Analysis Gender (female vs. male*) Tumor thickness (mean) Tumor distance to foveola (mean) Hemorrhage over tumor (yes vs. no*) Percentage difference in CMT between involved eye and opposite eye at baseline (>10% vs. 10% mm*)
130 10 3 92
(84) (6) (2) (59)
90 12 5 75
P Value
Risk Ratio
95% CI
1.52
(1.03e2.24)
1.84 1.10y 1.18y 1.05z
(1.03e3.27) (1.04e1.17) (1.09e1.28) (1.004e1.10)
(87) (12) (5) (71)
0.029 0.014 0.006 0.041
4.77 6.60 3.59 1.56
(1.17e19.4) (1.47e29.6) (1.44e8.91) (1.02e2.39)
11 (7) 2 (1)
17 (16) 5 (5)
0.006 0.003
2.32 4.24
(1.27e4.23) (1.63e11.1)
61 (40) 11.2
58 (55) 17.0
0.022 0.006
1.69 1.12x
(1.08e2.64) (1.03e1.22)
65 (42)
66 (63)
0.005
1.77
(1.19e2.64)
49 (32) 3.6
50 (48) 5.8
0.004 0.004
1.75 1.49x
(1.19e2.58) (1.14e1.94)
19 (12)
30 (29)
