Figure. Probability of retention of visual acuity of better than 20/200 using a life-table analysis. Subfoveal melanomas with radiation retinopathy were divided into the following 2 groups: those treated with anti–vascular endothelial growth factor (anti-VEGF) therapy and those not treated or receiving other treatment. The actual percentages of retention for the points in the graph are given in Table 5. There is a trend toward vision retention with anti-VEGF therapy.
Newman H, Chin KJ, Finger PT. Subfoveal Choroidal MelanomaPretreatment Characteristics and Response to Plaque Radiation Therapy. Arch Ophthalmol. 2011;129(7):892-898. doi:10.1001/archophthalmol.2011.161
Author Affiliations: The New York Eye Cancer Center (Drs Newman, Chin, and Finger), Department of Ophthalmology, Ocular Tumor Service, New York University School of Medicine (Drs Newman and Finger), and Department of Ophthalmology, Ocular Tumor Service, The New York Eye and Ear Infirmary (Drs Newman and Finger), New York City, New York.
Objective To evaluate the clinical presentation, tumor characteristics, and response to palladium 103 plaque radiation therapy for subfoveal choroidal melanomas.
Methods Retrospective case series of 50 patients diagnosed as having subfoveal melanoma and treated with plaque brachytherapy. Patients underwent evaluation for tumor characteristics, visual acuity, radiation damage, local tumor control, and metastatic disease.
Results Patients were followed up for a median of 54 (SD, 49.3) months. Forty-nine tumors (98%) were dome shaped. Subretinal fluid (overlying or a dependent exudative retinal detachment) was evident in 34 of 45 patients (76%). Treatment involved an apical radiation dose of 82.8 Gy (delivered across 5-7 days), resulting in a mean dose of 157.7 Gy to the fovea. Pretreatment median visual acuity was 20/50, which declined to 20/180 at last follow-up. Visual acuity was better than 20/200 in 33 patients (66%) at baseline and 25 (50%) at last follow-up; 13 patients (26%) lost 6 or more lines of vision. Twenty-eight patients (56%) developed radiation retinopathy; 16 (32%) required secondary intervention for radiation retinopathy, including intravitreal antivascular endothelial growth factor therapy, laser treatment, cryotherapy, or pars plana vitrectomy. The local tumor control rate of subfoveal tumors was 92%. Four patients (8%) required secondary enucleation. Metastasis developed in 2 patients (4%).
Conclusions Subfoveal choroidal melanomas in this series are almost exclusively dome shaped and likely to have an associated exudative retinal detachment. They are amenable to plaque radiation therapy. However, this tumor location is associated with a high incidence of radiation maculopathy and a low incidence of radiation cataract.
Uveal melanomas are often defined according to their intraocular location (ie, iris, ciliary body, or choroid). Each tumor location is associated with unique morphologic and prognostic characteristics. Owing to their distinctive location, we sought to investigate the features of subfoveal melanomas.
Most uveal melanomas are centered posterior to the equator. In a recent study of 400 patients, 61% had a tumor apex posterior to the equator, 32% were anterior, and 7% were equatorial.1 Similarly, the Collaborative Ocular Melanoma Study (COMS) report 29,2 describing baseline echographic characteristics of 2320 patients with medium and large tumors, found that 1268 patients (55%) had a tumor apex located posterior to the equator (27% equator-posterior, 22% posterior-equator, and only 5% posterior); foveal involvement was not specified.
Subfoveal choroidal melanoma treatment results in high-dose foveal irradiation because it is located within the targeted zone. In 2000, Finger3 found that proximity of the plaque to normal ocular structures is an important risk factor for adverse effects of radiation therapy. Similarly, COMS report 164 found that proximity between the tumor and the foveal avascular zone was strongly associated with less favorable outcomes for vision (visual acuity [VA] of 20/200 or worse or loss of ≥6 lines of VA) at 3 years after treatment.
Most recently, Finger et al5 showed that radiation maculopathy occurs more frequently in posterior tumors compared with anterior tumors (41% vs 7%), with posterior tumor location identified as a risk factor in multivariate analysis (hazard ratio, 6.66). The dose of radiation to the fovea was the most significant risk factor for radiation maculopathy in multivariate analysis. These findings suggest that the subset of subfoveal tumors might have poorer radiation-related outcomes for vision than extrafoveal choroidal melanomas. In this case series, we report on our experience with subfoveal choroidal melanomas with respect to tumor characteristics, VA, radiation retinopathy, and local tumor control. These findings should help patients better understand the risks and potential benefits of plaque radiation therapy for subfoveal choroidal melanoma.
The study was conducted according to the tenets of the Declaration of Helsinki and adhered to the Health Insurance Portability and Accountability Act of 1996. We obtained approval for the study from The New York Eye Cancer Center institutional review board and ethics committee. We searched the database of The New York Eye Cancer Center for patients diagnosed as having a subfoveal choroidal melanoma. The study constituted a retrospective review of the medical records of patients with a minimum of 6 months of follow-up. Patients with 6 months of follow-up were included (in part) because COMS report 19 revealed that 14 of the 69 cases of recurrences had occurred during the first 6 months of follow-up.6
Our methods of diagnosis, informed consent, treatment, and follow-up have been described elsewhere.7 Aspects of particular importance to this study include that tumor basal dimensions and location were determined by ophthalmoscopy, transillumination, fluorescein angiography, and ultrasonography. Patients treated within the framework of the COMS were excluded from this study.
A subfoveal melanoma was defined as a choroidal melanoma with its posterior margin or tumor apex located beneath the fovea. Tumor shape was described as dome, mushroom, or irregular. We noted the presence of subretinal fluid or exudative retinal detachment. Melanomas were classified according to the TMN staging system for location, height, basal diameter, node involvement, and the presence of metastasis.8
Best-corrected VA was recorded by certified COMS9 examiners (K.J.C. and P.T.F.) before the procedure and at every posttreatment visit, using Early Treatment Diabetic Retinopathy Study charts. Using COMS VA grading, if VA was counting fingers, we graded acuity as the distance in feet that the patient was able to count fingers. If the patient had hand movements or light perception VA, acuity was graded as 1/800, and if the patient had undergone enucleation or had no light perception VA, acuity was graded as 00/00 (Snellen notation).
Radiation retinopathy was assessed by ophthalmoscopy, fundus photography, fluorescein angiography, and optical coherence tomography. Retinopathy was graded according to the Finger classification.10 This staging system was developed to classify radiation retinopathy and to predict ocular morbidity. In summary, stage 1 consisted of cotton-wool spots, retinal hemorrhages, microaneurysms, exudation, or choroidopathy outside the macula (area of retina with a diameter of 5.5 mm, centered 4 mm temporal and 0.8 mm inferior to the center of the optic disc). Stage 2 consisted of similar findings in the macula. Stage 3 consisted of any of these findings plus new-onset macular edema or retinal neovascularization, and stage 4 consisted of any of these findings plus vitreous hemorrhage, retinal ischemia of at least 5 disc areas, or the presence of radiation optic neuropathy. We described patients as having radiation optic neuropathy if they had clinical evidence of optic disc neovascularization, edema, leakage, hemorrhage, ischemia, or atrophy as signs of advancing stages.
Palladium 103 (103Pd) seeds (model 200) were affixed into gold COMS-type ophthalmic plaques with a thin layer of acrylic fixative and seed-guide inserts.11 Dosimetry calculations were comparable with the COMS protocol.11,12 We also took into account recommendations of the American Association of Physicists in Medicine task group13 and the American Brachytherapy Society.14 Thus, seeds were calculated as point sources (with no correction for anisotropy), and no attenuation was attributed to the acrylic fixative or the 0.5-mm-thick gold plaque sidewalls. Backscatter effects from the plaque's posterior wall were discounted. Dosimetry was compatible with the National Cancer Institute Brachytherapy Contract Group determinations over time. A specific dose rate constant of 1.09 cGy/h/mCi for 103Pd was used, and our 103Pd radial dose function was obtained from published data.15,16 Our prescription point was the tumor apex (the farthest point of intraocular tumor extension from the inner sclera).14
Surgical techniques for 103Pd episcleral plaque placement have been described elsewhere.7 In addition, intraoperative ultrasonography was used to confirm plaque placement.17,18 In this series, radiation was delivered to a mean apical radiation dose of 82.8 Gy across 5 to 7 continuous days.
Data were summarized as mean, median, standard deviation, minimum, and maximum for the following variables: tumor dimensions, radiation dosage, and the tumor's distance to the optic nerve. Number and percentage were tabulated for categorical data. Visual acuity data at baseline, at follow-up visits, and at last visit were summarized as mean, median, minimum, and maximum. Statistical analysis was performed using SPSS software (SPSS for Windows, version 13.0; SPSS, Inc, Chicago, Illinois).
After review of the medical records, 50 eligible patients remained for statistical analysis. Three patients with a follow-up period shorter than 6 months were excluded; none of these 3 were noted to have an early recurrence or enucleation. Patients were followed up for a median of 54 (mean [SD], 62.2 [49.3]) months.
As listed in Table 1, 49 tumors were dome shaped (98%), whereas 1 was irregular (2%). Thirty-four of 45 patients (76%) had subretinal fluid, defined as an inferior or overlying exudative retinal detachment. Five patients could not be assessed for exudative retinal detachment at the time of treatment. The mean distance from the tumor to the optic nerve was 1.9 (SD, 1.2; range, 0-3.6) mm. The mean tumor dimensions were a width of 10.7 mm, length of 10.9 mm, and height of 3.5 mm. According to the TMN classification system, there were 26 T1 tumors (52%), 17 T2 tumors (34%), 5 T3 tumors (10%), and 2 T4 tumors (4%).
Radiation treatment characteristics are listed in Table 2. The mean doses of radiation were 82.8 Gy to the tumor apex, 157.7 Gy to the fovea, 67.0 Gy to the optic nerve, and 5.6 Gy to the lens. All plaque surgical procedures were performed by an experienced ophthalmic oncologist (P.T.F.), who has recently reported 18-year results on 400 patients treated with 103Pd plaque therapy.1 The plaque procedure required temporary extraocular muscle disinsertion in 40 patients (80%).
Ten patients (22% of 46 patients with phakic eyes at baseline) developed a secondary cataract. Twenty-eight patients (56%) developed radiation retinopathy with a mean time to onset of 40.6 months (Table 3). According to the Finger classification, of the 28 patients who had retinopathy, none had stage 1, 4 (14%) had stage 2, 15 (54%) had stage 3, and 9 (32%) had stage 4 retinopathy.10 Eight patients had diabetes mellitus (16%); half have developed radiation retinopathy.
Sixteen patients (32%) received treatment for radiation retinopathy. Treatment included intravitreal anti–vascular endothelial growth factor (anti-VEGF) injections of bevacizumab (Avastin) or ranibizumab (Lucentis) in 11 patients (69%), alone or in combination with pars plana vitrectomy or laser therapy. Six of 16 patients (38%) required laser photocoagulation (primarily scatter laser for neovascularization or focal laser for extramacular disease before the advent of anti-VEGF therapy). Twelve patients did not receive any treatment because they had only early stages of retinopathy with no visual decline or they were followed up before the advent of anti-VEGF treatment.
Pretreatment median VA was 20/50 (Table 4). Pretreatment VA was better than 20/200 in 33 patients (66%), and 22 patients (44%) had 20/40 or better. The median final VA was 20/180 at 62.2 months. Twenty-five patients (50%) had a final VA better than 20/200; 18 (36%), 20/40 acuity or better. Thirteen patients (26%) lost 6 lines or more of VA during the follow-up interval (95% confidence interval, 13.8%-38.2%). This group often started with compromised vision; however, of the 33 patients with a baseline VA better than 20/200, 21 (64%) had better or unchanged VA, and 12 (36%) had worse VA at the final follow-up visit.
A life-table analysis was performed to examine whether VA outcomes were affected by the use of anti-VEGF therapy for radiation retinopathy.19- 21 The Figure and Table 5 demonstrate the probability of retention of VA of better than 20/200. Subfoveal melanomas with radiation retinopathy were divided into the following 2 groups: patients receiving treatment with anti-VEGF therapy (n = 11) and those receiving no treatment or another treatment (n = 17). Nine of 11 patients receiving anti-VEGF therapy and 13 of 17 not receiving anti-VEGF therapy had preoperative VA of better than 20/200 and were included in the analysis.
Life-table analysis showed that, at 3 years after plaque brachytherapy, those with anti-VEGF–treated retinopathy had an 85.7% retention rate, whereas those not treated with anti-VEGF had a 66.3% retention rate. This study also suggests that, at 8 years after plaque radiation therapy, all patients not treated with anti-VEGF developed VA of 20/200 or worse, while those treated with anti-VEGF still had a 19% (better than 20/200) VA retention rate. Multivariate analysis was not performed because of the small number in this subset.
In this series, tumor regrowth (local failure) was noted in 4 patients (8%); 2 underwent a second treatment of 103Pd brachytherapy, whereas the other 2 underwent enucleation. Overall, 4 patients (8%) required secondary enucleation. Two patients (4%) developed neovascular glaucoma along with radiation retinopathy (without local tumor regrowth). One was treated with panretinal laser, whereas the other required enucleation after conservative treatment with laser and panretinal and ciliary body cryotherapy failed. Neither received anti-VEGF therapy, one because of patient preference and the other because treatment was received before the advent of anti-VEGF therapy. Metastatic melanoma was documented in 2 patients (4%) who died 27 and 36 months after diagnosis. The overall mortality rate from all causes was 12% (n = 6). Four other patients died of nonmelanoma-related causes.
Herein we examine the initial clinical characteristics and treatment outcomes of subfoveal melanomas treated with plaque radiation therapy. In this series, most tumors (76%) had subretinal fluid and exudative retinal detachment at the initial examination. This rate is within the upper range of that reported in the literature because exudative retinal detachment was detected clinically in up to 75% of eyes with malignant uveal melanoma.22 This is not surprising because posterior location of choroidal melanomas has been described as one of the predictors of exudative retinal detachment.23
This study found that almost all subfoveal tumors (98%) were dome shaped and that their mean height was 3.5 mm. These findings are reasonable because a mushroom shape occurs more frequently in larger tumors.2 For example, the COMS report 29 stated that only 2% of tumors thinner than 5 mm were mushroom shaped vs 66% of tumors thicker than 10 mm. That report has additionally found that most (73%) of the smaller tumors (those thinner than 5 mm) had their apex located posterior to the equator, whereas only 24% of tumors thicker than 10 mm had their apex located posterior to the equator.2 This finding suggests that subfoveal tumors are probably diagnosed earlier than anterior tumors owing to the presence of visual symptoms and a greater likelihood of early detection by periodic ophthalmic examinations.
Subfoveal choroidal melanoma baseline VAs were inferior to those reported for all-location studies of uveal melanomas.4 The COMS report 16 describing medium choroidal melanoma found a baseline median VA of 20/32, with 70% of the eyes having 20/40 or better acuity and 10% having 20/200 or worse.4 In our study, the median baseline acuity was 20/50, with 44% of the eyes having a VA of 20/40 or better and 34% having VA of 20/200 or worse. The poorer baseline acuities can be explained by damage to the foveal photoreceptors, by direct retinal invasion by the tumor, or by the presence of subretinal fluid or exudative retinal detachment, as was found in most of our cases. Exudative retinal detachment is a known poor prognostic factor for VA before and after treatment.24- 29
With regard to outcome, VA was better than 20/200 in 25 eyes (50%) at last follow-up. Although we could find no other reports on the subset of subfoveal choroidal melanoma, all-location reports, such as that of Summanen et al,30 noted a 41% retention rate of at least 20/200 vision at 3-year VA after ruthenium 106 (106Ru) brachytherapy for uveal melanoma. Damato et al31 found a 57% actuarial rate of conservation of 20/200 or better vision at 9 years in choroidal melanomas (also in all locations) treated with 106Ru. However, comparisons with other studies are of limited value because previous studies included eyes with melanoma in all locations (even peripheral melanomas).
One would expect eyes with subfoveal melanoma to have a worse outcome for vision after plaque brachytherapy. Subfoveal tumors meet 3 criteria for poor visual outcome found in the COMS report: short distance between the tumor and the foveal avascular zone, poorer baseline VA, and tumor-associated retinal detachment.4 The COMS used a higher radiation dose and a more energetic radionuclide (iodine 125). Both factors may have increased the dose to the fovea and diminished VA results.5 Also, the COMS took place before the era of intravitreal anti-VEGF suppression of radiation maculopathy and optic neuropathy.19- 21
The advent of anti-VEGF therapy has altered VA outcomes in patients with radiation retinopathy.19- 21 Life-table analysis comparing patients with retinopathy treated with anti-VEGF vs without anti-VEGF demonstrates a trend toward vision retention with anti-VEGF therapy over time. Although the study size was relatively small, it also shows that, in most cases of subfoveal choroidal melanoma, vision loss has been progressive despite anti-VEGF therapy. Further studies with larger sample sizes are required to investigate this topic.
Subfoveal location was associated with a lower rate of radiation-induced cataract than that reported in a recent study32 of 282 patients with all-locations uveal melanomas (24% at a mean follow-up of 39.8 months). In that study, anterior tumors had a higher incidence of cataract than did posterior tumors. Multivariate analysis revealed that the lens dose, and not the tumor location, was the best predictor for radiation cataract. Furthermore, the percentages of patients who developed a secondary cataract were 11.5% (dose, <4.0 Gy), 27.7% (dose, 4.0-9.9 Gy), 39.4% (dose, 10.0-19.9 Gy), and 75% (dose, >20.0 Gy).32 In this series of subfoveal choroidal melanomas, the mean central lens dose was 5.6 Gy, yielding a similar rate (21.7%) of secondary cataract formation.
Radiation retinopathy developed in 56% of plaque-irradiated subfoveal melanomas at a mean follow-up of 5.2 years (range, 6-223 months). Although this percentage could be higher if we lengthened the inclusion criteria for follow-up (from 6 months), our mean follow-up far exceeds the average time for tumors in all locations to develop radiation retinopathy (at 2 years).1,5 Most radiation retinopathy in this subfoveal study was advanced (stage 3 in 54%; stage 4 in 32%). This can be explained by the high foveal dose (157.5 Gy) and that the staging system was created to upstage risk factors for vision loss (including anatomic tumor location). A recent study5 examined which tumor and treatment factors might influence the incidence of radiation maculopathy. Although posterior tumor location was a significant risk factor, multivariate analysis revealed that the dose to the fovea was the most significant factor associated with the incidence of radiation maculopathy. Furthermore, compared with a dose of less than 35 Gy, the hazard risks of radiation maculopathy were 1.74 for doses of 35 to 70 Gy and 2.43 for doses of 70 Gy or more. We expect the incidence of radiation retinopathy to increase with additional follow-up.
In this series, we used 103Pd seeds in gold ophthalmic plaques. Therefore, our results may not be the same as those centers using alternative radionuclides (eg, 125I or 106Ru) or external beam techniques (eg, proton or cyberknife). This study offers further evidence that pretreatment comparative dosimetry studies using these radiation modalities is likely to allow for selection of the most favorable source for vision retention.
This study presents a good local control rate of 92% with 103Pd ophthalmic plaque radiation therapy. This rate is lower than the 97% previously reported (by us) of all-location uveal melanomas.1 This difference represents a selection that includes more small posterior melanomas that have been noted to more commonly recur.1 However, these results are still comparable to those found with the use of 125I plaque by the COMS report 19 of a local control rate of 89.7% at 5 years.6
In conclusion, this study suggests that subfoveal melanomas are typically seen at first examination as relatively small tumors with comparatively poor initial VAs. Most are dome shaped. They are more likely to have an associated retinal detachment. Patients should expect good local control, with limited although better than expected VA results (50% of patients with better than 20/200 at 5.2 years). However, this is likely related (in part) to the advent of anti-VEGF therapy. Subfoveal melanomas were associated with lower rates of radiation cataract and increased risk for radiation maculopathy. We hope this work is found helpful to those who care for patients with choroidal melanoma and to those patients affected by this disease.
Correspondence: Paul T. Finger, MD, The New York Eye Cancer Center, Ste 5B, 115 E 61st St, New York City, NY 10065 (email@example.com).
Submitted for Publication: August 6, 2010; final revision received and accepted October 21, 2010.
Financial Disclosure: None reported.
Funding/Support: This study was supported by The Eye Cancer Foundation, Inc.
Additional Contributions: Praveen K. Nirmalan, MBBS, MPH, PRASHASA Health Consultants Pvt Ltd, Hyderabad, India, contributed statistical expertise.