The Pediatric Choroidal and Ciliary Body Melanoma Study

The Pediatric Choroidal and Ciliary Body Melanoma Study A Survey by the European Ophthalmic Oncology Group Rana’a T. Al-Jamal, MD,1 Nathalie Cassoux, MD,2 Laurence Desjardins, MD,2 Bertil Damato, MD,3 Lazaros Konstantinidis, MD,3 Sarah E. Coupland, MD,3,4 Heinrich Heimann, MD,3 Aleksandra Petrovic, MD,5 Leonidas Zografos, MD,5 Ann Schalenbourg, MD,5 Juan P. Velazquez-Martin, MD,6 Hatem Krema, MD,6 Anna Bogdali, MD,7 Anna Markiewicz, MD,7 Bozena Romanowska-Dixon, MD,7 Claudia H.D. Metz, MD,8 Eva Biewald, MD,8 Norbert Bornfeld, MD,8 Hayyam Kiratli, MD,9 Inge H.G. Bronkhorst, MD,10 Martine J. Jager, MD,10 Marina Marinkovic, MD,10 Maria Fili, MD,11 Stefan Seregard, MD,11 Shahar Frenkel, MD,12 Jacob Pe’er, MD,12 Sachin M. Salvi, MD,13 Ian G. Rennie, MD,13 Iwona Rospond-Kubiak, MD,14 Jaroslaw Kociecki, MD,14 Jens Folke Kiilgaard, MD,15 Steffen Heegaard, MD,15 Victoria M.L. Cohen, MD,16 Mandeep S. Sagoo, MD,16 Anush Amiryan, MD,17 Svetlana Saakyan, MD,17 Nils Eide, MD,18 Jørgen Krohn, MD,19 Edoardo Midena, MD,20 Raffaele Parrozzani, MD,21 Jean-Daniel Grange, MD,22 Emine Kilic, MD,23 Maria Antonietta Blasi, MD,24 Maria Antonia Saornil, MD,25 Tero T. Kivelä, MD1 Purpose: To collect comprehensive data on choroidal and ciliary body melanoma (CCBM) in children and to validate hypotheses regarding pediatric CCBM: children younger than 18 years, males, and those without ciliary body involvement (CBI) have more favorable survival prognosis than young adults 18 to 24 years of age, females, and those with CBI. Design: Retrospective, multicenter observational study. Participants: Two hundred ninety-nine patients from 24 ocular oncology centers, of whom 114 were children (median age, 15.1 years; range, 2.7e17.9 years) and 185 were young adults. Methods: Data were entered through a secure website and were reviewed centrally. Survival was analyzed using Kaplan-Meier analysis and Cox proportional hazards regression. Main Outcome Measures: Proportion of females, tumor-node-metastasis (TNM) stage, cell type, and melanoma-related mortality. Results: Cumulative frequency of having CCBM diagnosed increased steadily by 0.8% per year of age between 5 and 10 years of age and, after a 6-year transition period, by 8.8% per year from age 17 years onward. Of children and young adults, 57% and 63% were female, respectively, which exceeded the expected 51% among young adults. Cell type, known for 35% of tumors, and TNM stage (I in 22% and 21%, II in 49% and 52%, III in 30% and 28%, respectively) were comparable for children and young adults. Melanoma-related survival was 97% and 90% at 5 years and 92% and 80% at 10 years for children compared with young adults, respectively (P ¼ 0.013). Males tended to have a more favorable survival than females among children (100% vs. 85% at 10 years; P ¼ 0.058). Increasing TNM stage was associated with poorer survival (stages I, II, and III: 100% vs. 86% vs. 76%, respectively; P ¼ 0.0011). By multivariate analysis, being a young adult (adjusted hazard rate [HR], 2.57), a higher TNM stage (HR, 2.88 and 8.38 for stages II and III, respectively), and female gender (HR, 2.38) independently predicted less favorable survival. Ciliary body involvement and cell type were not associated with survival. Conclusions: This study confirms that children with CCBM have a more favorable survival than young adults 18 to 25 years of age, adjusting for TNM stage and gender. The association between gender and survival varies between age groups. Ophthalmology 2016;123:898-907 ª 2016 by the American Academy of Ophthalmology. Supplemental material is available at www.aaojournal.org. Uveal melanoma (UM) has an annual incidence of 2 to 8 per million in North America and Europe, varying by age, ethnicity, and latitude.1e3 It is generally a disease of middleaged and older adults, with a low incidence before 45 years 898  2016 by the American Academy of Ophthalmology Published by Elsevier Inc. of age; the median age at diagnosis has increased to 62 years because of increasing life expectancy.3,4 Nevertheless, UM can occur at any age, even as a congenital tumor.5,6 Oculo(dermal) melanocytosis, neurofibromatosis type 1, familial http://dx.doi.org/10.1016/j.ophtha.2015.12.024 ISSN 0161-6420/16 Al-Jamal et al  Choroidal and Ciliary Body Melanoma Study atypical multiple mole and melanoma syndrome, and germline mutations in the BRCA1-associated protein 1 (BAP1) gene have been alleged to play a role in its development, especially in younger patients.7e10 The randomized Collaborative Ocular Melanoma Study did not provide data on UM in patients younger than 21 years, who were ineligible for the study.11 However, in a single-center series of 8033 patients and in several smaller series, patients younger than 21 years have constituted 0.8% to 1.1% of the studied cohorts.7,8,12e15 Given an estimated annual world incidence of 6700 to 7100 cases of UM, this translates to approximately 65 young patients per year.3 The reported features of UM in the latter as compared with adults include a higher incidence of iris melanoma and better survival prognosis, attributed to smaller tumor size and, perhaps, a more active immune system in younger patients.7,12e15 Histopathologic or molecular pathologic studies of UM in children have not demonstrated any differences from their adult counterparts.16,17 Two of the authors (T.T.K. and R.A.J.) recently undertook a meta-analysis of 88 patients younger than 25 years of age with choroidal and ciliary body melanoma (CCBM) that suggested that female gender and higher American Joint Committee on Cancer tumor-node-metastasis (TNM) stage both adversely influence survival.18 The meta-analysis also suggested that patients younger than 18 years may have an excellent life prognosis, especially if they are male, compared with those 18 to 24 years of age, especially if they are female and if the ciliary body is involved.18 To test these hypotheses and to collect more comprehensive data on CCBM in patients younger than 25 years of age than what are available from the literature, we established the collaborative Pediatric Choroidal and Ciliary Body Melanoma Study of the European Ophthalmic Oncology Group (http:// www.oog.eu.com). Herein, we present our data obtained from children younger than 18 years compared with young adults 18 to 24 years of age. To the best of our knowledge, ours is the largest series to characterize CCBM in these age groups. Methods Aims of the Study The study aimed to test 3 hypotheses derived from a meta-analysis and a large single-center study of CCBM in young patients18,19: (1) children have a more favorable life prognosis than young adults and, when both groups are combined, (2) males have higher survival rates than females and (3) ciliary body involvement (CBI) is a parameter for poor prognosis.18 Study Design For the purpose of our study, cases were defined as patients younger than 18 years at the time of diagnosis of a CCBM, corresponding to the joint definition of children by the European Medicines Agency and the European Union of Pediatrics,20 and controls were defined as adults younger than 25 years at diagnosis. This age limit was chosen because it was predicted to yield a second group of comparable size on the basis of the meta-analysis18 and because previous series of UM in adolescents adhered to this or a lower age limit.7,13,15,19,21e24 We formulated our hypotheses based on the meta-analysis of 88 patients younger than 25 years (none with an iris melanoma)18 extracted from 6 series and a single-center, referral-based cohort study including 86 patients younger than 20 years19 (25% had an iris melanoma), as follows. First, children have better 10-year survival versus young adults (meta-analysis, 100% vs. 85%; cohort study, 91% for younger than 20 years with CCBM). To calculate the sample needed, we presumed a 10-year survival of 97% for children (allowing for a small number of deaths, based on the cohort study) and 85% for young adults. We further presumed 44% to be children and 44% to be censored from the analysis (both percentages taken from the meta-analysis). Second, for children and young adults combined, males have better 10-year survival than females (meta-analysis, 100% vs. 85%). Third, for children and young adults combined, those without CBI have better 10-year survival than those with such involvement (meta-analysis, 96% vs. 70%). A total sample of 289 patients was needed (power, 0.80; 1sided a, 0.05) for the first comparison, which was a sample in excess to that required for the other 2 comparisons. Eligible for our retrospective cohort study were all patients in whom a CCBM was diagnosed at an age younger than 25 years and for whom at least the following data were available: birth date, date of diagnosis, gender, treatment type, presence or absence of local or systemic tumor recurrence, last survival status, date of last known status, and cause of death (UM, second cancer, or nonmalignant cause) determined by reviewing patient charts, registry data, histologic samples, and death certificates. Patients with iris melanomas were ineligible. All treatment methods were eligible. This investigation was approved by the institutional review boards of the participating centers as required and adhered to the tenets of the Declaration of Helsinki. Data Collection Data on consecutive eligible patients were collected from members of the European Ophthalmic Oncology Group. The data additionally acquired included presence of congenital oculo(dermal) melanocytosis or neurofibromatosis; visual acuity and intraocular pressure at diagnosis and at last visit; tumor thickness; largest basal diameter of tumor; CBI; extraocular extension; tumor distance from the center of the fovea and the margin of the optic disc; tumor cell type; tumor cytogenetic features; dates of any local tumor recurrence; secondary enucleation, and metastasis; and second primary malignancies. We staged the tumors according to the seventh edition of the TNM system of the American Joint Committee on Cancer.25,26 Twenty-four participating ocular oncology services submitted data anonymously through a secure survey website from 356 patients diagnosed between February 1968 and February 2014, of whom 57 patients were excluded upon central review, leaving 299 (84%) for analysis, comprising 114 children (38%) and 185 young adults (62%) (for details see Appendix, available at www.aaojournal.org). Statistical Analysis All analyses were performed with Stata software version 13.0 (Stata Corp., College Station, TX). We used the Fisher exact test and nonparametric test for trend to compare unordered and singly ordered contingency tables, respectively, and the ManneWhitney U test to compare continuous variables between groups. All tests were 2-tailed, and P < 0.05 was taken as statistically significant unless otherwise specified. Statistics other than those related to our 3 predetermined hypotheses should be regarded as exploratory. The percentage of females was compared using the binomial test against the expected percentage, taken from the World Population Prospects of the United Nations for the participating countries,27 and averaged for the observed years of diagnosis. 899 Ophthalmology Volume 123, Number 4, April 2016 Survival was calculated from the date of diagnosis to death. We based univariate analysis on the Kaplan-Meier product-limit method and log-rank test or test for trend. The small number of deaths in many subcategories did not allow a separate analysis in children and young adults. Multivariate analysis was based on Cox proportional hazards regression. Independent variables are allowed in the model if P < 0.10, and models are compared with the likelihood ratio test.28 We restricted the number of variables in models to 3, based on a rule to have at least 15 to 20 events per each additional variable (we observed 35 melanoma deaths).29 We verified the assumption of proportional hazards according to Therneau and Grambsch.30 Because only 4 patients died of causes other than UM, we did not perform a separate competing risk analysis. Results Characteristics of Primary Tumors The median age of the 114 children was 15.1 years (range, 2.7e17.9 years) and that of the 185 young adults was 21.9 years. The cumulative frequency of having a CCBM diagnosed increased by a mean of 0.8% per year of age between ages 5 and 10 years and, after a 6-year transition from 11 to 16 years, by a mean of 8.8% per year between 17 and 24 years of age (Fig 1). Of children and young adults, 65 (57%) and 116 (63%) were female, respectively (Table 1). This percentage tended to be higher than expected (estimate, 51%) among children (P ¼ 0.053, 1-tailed binomial test) and was significantly higher than expected among young adults (P ¼ 0.0001). Of 268 participants (90%) with a known preexisting condition, 2 children (1.9%; 12 and 14 years of age) and 7 young adults (4.3%) had congenital oculo(dermal) melanocytosis, whereas 2 children (1.9%; 11 and 12 years of age) and 1 young adult (0.6%) had neurofibromatosis (Table 1). The median visual acuity of the tumor eye was 20/80 for children, worse than the median of 20/40 for young adults (P ¼ 0.005; Table 1). Median tumor thickness (6.1 vs. 6.0 mm) and largest basal diameter (12.3 vs. 12.4 mm), respectively, were comparable (Fig 2A, available at www.aaojournal.org), as was CBI (28% vs. 33%; Table 1). Three children (3%; 12, 17, and 17 years of age; diameter 5 mm in the only known case) and 5 young adults (3%; diameter <5 mm in all 4 known cases) had extraocular extension. The median distance to the foveola was shorter in children than in young adults (1.4 vs. 3.0 mm; P ¼ 0.040), whereas the median distance to the disc was comparable (2.5 vs. 3.5 mm, respectively; P ¼ 0.23, Table 1; Fig 2B, available at www.aaojournal.org). The tumor could be staged in 281 participants (94%), and the TNM size categories and stages showed comparable distributions among children and young adults (Table 1): T1 in 27% versus 25%, T2 in 23% versus 34%, T3 in 29% versus 29%, and T4 in 21% versus 12%, respectively (P ¼ 0.27, nonparametric test for trend); and stage I in 22% versus 21%, stage II in 49% versus 52%, and stage III in 30% versus 28%, respectively (P ¼ 0.85). Primary Treatment Three children (3%) were treated with laser (Table 2). Radiotherapy alone was delivered to 72 children (63%) versus 125 young adults (68%), and 42 children (37%) versus 57 young adults (31%) underwent surgical treatment with or without adjuvant radiotherapy. Histopathologic Characteristics Cell type was known for 39 children (34%) and 67 young adults (36%), corresponding to 82 of 99 patients (83%) treated surgically and to 24 of 200 patients (12%) managed conservatively who 900 Figure 1. Cumulative frequency distribution plot of age at diagnosis for 114 children younger than 18 years and 185 young adults 18 to 24 years of age with choroidal and ciliary body melanoma. Note an initial period of a slower (dashed red line) and later faster (dashed green line) steady increase with a gradual transition occurring between 11 and 17 years of age. underwent a biopsy. The thickness (P ¼ 0.99, ManneWhitney U test), largest basal diameter (P ¼ 0.37), TNM category (P ¼ 0.86, nonparametric test for trend), and TNM stage (P ¼ 0.25) of these tumors were comparable for children and young adults. The percentages of spindle-cell, mixed cell, and epithelioid cell melanomas were similar for both groups (P ¼ 0.93; Table 2). Cytogenetic Characteristics Cytogenetic results were available for 15 children (13%) and 25 young adults (14%; for details, see the Appendix, available at www.aaojournal.org). Monosomy 3 was found in 8 children (53%; 11e17 years of age) and 6 young adults (24%; P ¼ 0.089, Fisher exact test), of whom one half in both groups had an additional 8q gain (Table 2). Five children and 5 young adults had been screened for somatic BAP1 mutations. One patient in both groups showed positive results. Local Tumor Control The median follow-up time was 5.9 years (range, 1.4e31.1 years; interquartile range, 3.1e10.6 years) for 35 patients who died of melanoma, 5.5 years (range, 1 daye41.3 years; interquartile range, 2.2e11.9 years) for survivors, and 6.6 years (range, 1 daye41.3 years) for children versus 5.1 years (range, 3 dayse37 years) for young adults (P ¼ 0.12; Table 2). Local tumor recurrence was diagnosed in 7 children (6%; 10e17 years of age) and 9 young adults (5%; Table 2). KaplanMeier estimates of the cumulative proportion with local tumor recurrence were comparable (P ¼ 0.79, log-rank test; Fig 3, available at www.aaojournal.org). The median visual acuity at the last visit was counting fingers (Table 2; Appendix, available at www.aaojournal.org). Univariate Survival Outcome By the end of the follow-up, 8 children (7%) versus 27 young adults (15%) had died of metastatic UM, and 1 patient (1%) versus 4 patients (2%), respectively, were alive with metastases a median of 4.8 months (range, 1 weeke10.7 years) after diagnosis of dissemination. In children, these primary tumors that metastasized were diagnosed at a median age of 14 years (range, 11e17 years). One young adult died of a second cancer (adenocarcinoma of the Al-Jamal et al  Choroidal and Ciliary Body Melanoma Study Table 1. Patient and Tumor Characteristics for 114 Children Younger than 18 Years and 185 Young Adults 18 to 24 Years of Age with Choroidal and Ciliary Body Melanoma Characteristic Median age, y (range) Gender, no. (%) Male Female Predisposing factors, no. (%) Melanocytosis Neurofibromatosis Initial visual acuity, no. (%) 20/20 20/25e20/40 20/50e20/200 CFeLP NLP Median IOP, mmHg (range) IOP, no. (%) Normal High Median thickness, mm (range) Median LBD, mm (range) Median distance to fovea, mm (range) Median distance to disc, mm (range) Ciliary body involvement, no. (%) Yes No Extraocular extension, no. (%) Yes No TNM category, no. (%) T1a T1b T2a T2b T2d T3a T3b T3c T3d T4a T4b T4d Unknown TNM stage, no. (%) I IIA IIB IIIA IIIB IIIC Unknown Young Adults (18e25 yrs, n [ 185) Children (1e17 yrs, n [ 114) 21.9 (18e24.9) 15.1 (2.7e17.9) 69 (37) 116 (63) 49 (43) 65 (57) 7 (4.3) 1 (0.6) 2 (1.9) 2 (1.9) 46 50 45 31 7 14 125 14 6.0 12.4 3.0 3.5 (26) (28) (25) (17) (4) (8e59) (90) (10) (0.8e16.9) (3.0e24.4) (0e25) (0e22) 16 29 32 25 11 14 83 6 6.1 12.3 1.4 2.5 (14) (26) (28) (22) (10) (7e59) (93) (7) (0.9e20) (2.0e24.3) (0e22) (0e24) 60 (33) 124 (67) 31 (28) 81 (72) 5 (3) 179 (97) 3 (3) 110 (97) Children P Value N/A 0.33* 0.21* 0.13* 0.005y 0.27z 0.48* z 0.69 0.76z 0.040z 0.23z 0.44* 1e10 yrs (n ¼ 15) 7.2 (2.7-10) 11-17 yrs (n ¼ 99) P Value 15.4 (11.1e17.9) N/A 0.58* 5 (33) 1 (67) 44 (44) 55 (56) 0 (0) 0 (0) 2 (2) 2 (2) 3 4 4 2 2 7 9 2 7.2 12.4 2.0 1.25 (20) (27) (27) (13) (13) (3e10) (82) (18) (2.0e14.5) (3.1e24) (0e10.5) (0e10.5) 13 25 28 23 9 15 74 4 6 12.2 1.4 3.0 (13) (26) (29) (23) (9) (11e18) (95) (5) (0.9e20.0) (2.0e24.3) (0e22) (0e24) 5 (33) 10 (67) 26 (27) 71 (73) 0 (0) 15 (100) 3 (3) 110 (97) N/A (20) (4) (27) (5) (0.5) (13) (13) (0.5) (0.5) (3) (8) (0.5) (5) 25 7 21 3 1 22 8 1 0 11 10 1 4 (22) (6) (18) (3) (1) (19) (7) (1) (0) (10) (9) (1) (3) 37 57 33 32 15 1 10 (20) (31) (18) (17) (8) (0.5) (5.5) 25 28 25 21 10 1 4 (22) (25) (22) (18) (9) (1) (3) 0.20z 0.16* 0.54z 0.63z 0.92z 0.20z 0.76* 0.99* 0.99* 37 7 50 9 1 24 24 1 1 6 14 1 10 0.70* 0.69* 0.56y N/A 5 2 0 0 0 4 1 0 0 1 2 0 0 (33) (13) (0) (0) (0) (27) (7) (0) (0) (7) (13) (0) (0) 20 5 21 3 1 18 7 1 0 10 8 1 4 (20) (5) (22) (3) (1) (18) (7) (1) (0) (10) (8) (1) (4) 5 2 4 2 2 0 0 (33) (13) (28) (13) (13) (0) (0) 20 26 21 19 8 1 4 (20) (26) (21) (19) (8) (1) (4) 0.98y 0.46y CF ¼ counting fingers; IOP ¼ intraocular pressure; LBD ¼ largest basal diameter; LP ¼ light perception; N/A ¼ not applicable or meaningful; NLP ¼ no light perception; TNM ¼ tumor-node-metastasis. *Two-tailed Fisher exact test. y Two-tailed nonparametric test for trend. z Two-tailed ManneWhitney U test. colon, Dukes B, pT3), 2 young adults died of nonneoplastic disease, and 1 young adult died of unknown causes. Based on all-cause mortality, survival was 97% versus 89% at 5 years and 91% versus 78% at 10 years for children versus young adults (P ¼ 0.0034; Fig 4A, available at www.aaojournal.org). By Kaplan-Meier analysis, survival based on melanoma-related mortality was 97% (95% confidence interval [CI], 90%e99%) at 5 years and 92% (95% CI, 81%e97%) at 10 years for children versus 90% (95% CI, 84%e94%) and 80% (95% CI, 71%e87%) for young adults (Fig 5A; P ¼ 0.013, log-rank test). For children 1 to 10 years of age versus 11 to 17 years of age at diagnosis, the proportions were 100% at 5 and 10 years versus 97% (95% CI, 901 Ophthalmology Volume 123, Number 4, April 2016 Table 2. Posttreatment Patient and Tumor Characteristics for 114 Children Younger than 18 Years and 185 Young Adults 18 to 24 Years of Age with Choroidal and Ciliary Body Melanoma Characteristic Primary treatment type, no. (%) Laser Transpupillary thermotherapy Photodynamic therapy Radiotherapy Proton beam Iodine plaque Ruthenium plaque Stereotactic Gamma knife Surgery Enucleation Local resection Combined: enucleation þ EBT Combined: resection þ gamma knife Cell type, no. (%) Epithelioid Spindle Mixed Chromosomal abnormality, no. (%) No abnormality Monosomy 3 Monosomy 3 with 8q gain Somatic BAP1 mutation, no. (%) Yes No Median follow-up, yrs (range) Local tumor recurrence, no. (%) Yes No Median local recurrence time, yrs (range) Final visual acuity, no. (%) 20/20 20/25e20/40 20/50e20/200 CFeLP NLP or enucleated Median final IOP (mmHg) Final IOP, no. (%) Normal High Survival status, no. (%) Alive Alive with metastasis Dead from metastasis Dead from other cancer Dead from other causes Dead from unknown cause Children Young Adults Children 1e17 yrs 18e25 yrs (n [ 185) (n [ 114) P Value 1e10 yrs (n ¼ 15) 11e17 yrs (n ¼ 99) P Value N/A 2 (1) 1 (0.5) 0 (0) 0 (0) N/A 0 (0) 0 (0) 0 (0) 0 (0) 67 21 35 1 1 (36) (11) (19) (0.5) (0.5) 39 10 21 0 2 (34) (9) (18) (0) (2) 4 0 1 0 1 (27) (0) (7) (0) (7) 35 10 20 0 1 (36) (10) (20) (0) (1) 38 17 1 1 (20.5) (9) (0.5) (0.5) 26 13 3 0 (23) (11) (3) (0) 7 2 0 0 (46) (13) (0) (0) 19 11 3 0 (19) (11) (3) (0) 0.78* 0.93* 11 (17) 33 (49) 23 (34) 7 (18) 20 (51) 12 (31) 19 (76) 3 (12) 3 (12) 7 (46) 4 (27) 4 (27) 1 (14) 3 (43) 3 (43) 6 (19) 17 (53) 9 (28) 4 (100) 0 (0) 0 (0) 3 (28) 4 (36) 4 (36) 0.069* 0.025* 0.65y 1 (25) 4 (75) 5.1 (0e37) 1 (25) 4 (75) 6.6 (0e41.3) 9 (5) 177 (95) 2.3 (1.2e20) 7 (6) 107 (94) 6.6 (1.4e12) 23 26 28 35 39 14 (15) (17) (19) (23) (26) (2e50) 8 11 17 29 29 13 97 (95) 5 (5) (8) (12) (18) (31) (31) (1e28) 0.12z 0.79y 0.35z 0.087* 0.13z 0.99y 60 (95) 3 (5) 0.51y 0 (0) 1 (100) 7.2 (2.7e10) 1 (7) 14 (93) 1.3 0 1 3 1 4 7.2 (0) (11) (33) (11) (45) (2.7e10) 5 (100) 0 (0) 1 (25) 3 (75) 15.4 (11.2e18) 6 (6) 93 (94) 6.8 (2.8e12) 8 10 14 28 25 15.4 (9) (12) (17) (33) (29) (11.2e18) (81) (2) (15) (0.5) (1) (0.5) 105 1 8 0 0 0 (92) (0.9) (7) (0) (0) (0) 0.13z 0.45* 0.89z 0.99y 55 (95) 3 (5) N/A 150 4 27 1 2 1 0.53z 0.99y N/A 15 0 0 0 0 0 (100) (0) (0) (0) (0) (0) 90 1 8 0 0 0 (91) (1) (8) (0) (0) (0) CF ¼ counting fingers; EBT ¼ external beam radiotherapy; IOP ¼ intraocular pressure; LP ¼ light perception; N/A ¼ not applicable or meaningful; NLP ¼ no light perception. *Two-tailed nonparametric test for trend. y Two-tailed Fisher exact test. z Two-tailed ManneWhitney U test. 88%e99%) at 5 years and 91% (95% CI, 79%e96%) at 10 years, respectively (P ¼ 0.002; Fig 4B, available at www.aaojournal.org). Nine of 65 female children and 1 of 49 male children died of metastases, and males tended to have a more favorable 10-year survival compared with females (100% vs. 85%; P ¼ 0.058; Fig 4C, available at www.aaojournal.org). No such difference was 902 observed in young adults (81% vs. 80%; P ¼ 0.75; Fig 4D, available at www.aaojournal.org) and when combining both groups (90% vs. 82%; P ¼ 0.16; Fig 5B). Survival was associated with oculo(dermal) melanocytosis; 3 of 9 affected patients died (P ¼ 0.0016; Fig 5C). Too few patients had neurofibromatosis to allow analysis. Combining both age groups, the 10-year survival Al-Jamal et al  Choroidal and Ciliary Body Melanoma Study Figure 5. Kaplan-Meier estimates of melanoma-related mortality among 114 children younger than 18 years and 185 young adults 18 to 24 years of age with choroidal and ciliary body melanoma according to (A) age group, (B) gender, (C) presence of congenital oculo(dermal) melanocytosis, (D) ciliary body involvement, (E) extraocular extension, and (F) tumor-node-metastasis stage. Ticks show censored observations, and numbers below graphs represent patients at risk; (AeE) log-rank test and (F) test for trend. Three young adults died of other causes and 1 died of an unknown cause; all were censored. One further melanoma death occurred at 31 years of follow-up (age, 18e24 years; male; no ocular melanocytosis; ciliary body involved; no extraocular extension; stage III). of young patients did not differ by CBI (87% for not involved vs. 80% for involved; P ¼ 0.17; Fig 5D), whereas extraocular extension predicted a significantly worse survival; 4 of 8 patients died (87% for no extension vs. 39% for extension; P ¼ 0.0002; Fig 5E). TNM stage showed a trend with survival among children (stage I vs. stage II vs. stage III; 100% vs. 96% vs. 82%, respectively, at 10 years; P ¼ 0.091, log-rank test for trend; Fig 4E, available at www.aaojournal.org) and was associated strongly with survival 903 Ophthalmology Volume 123, Number 4, April 2016 both in young adults (100% vs. 80% vs. 73%, respectively; P ¼ 0.0043; Fig 4F, available at www.aaojournal.org) and when both groups were combined (100% vs. 86% vs. 76%, respectively; P ¼ 0.0011; Fig 5F). Cell type was not associated with survival (78% vs. 80% vs. 72% for spindle-cell, mixed cell, and epithelioid cell melanomas, respectively; P ¼ 0.93, log-rank test for trend; Fig 4G, available at www.aaojournal.org). Of the 8 children with monosomy 3, 1 died of metastasis 4.2 years after diagnosis, 1 is alive with metastasis at 2.2 years (both had an additional 8q gain), and 6 survive without metastases after a median followup of 2.2 years (range, 1 daye4.3 years; 2 with 8q gain), whereas of the 6 young adults, 1 (with 8q gain) died of an unknown cause and 5 survive after a median follow-up of 1.2 years (range, 3 dayse21 years; 2 with 8q gain). None of the 26 patients with disomy 3 so far have demonstrated metastases (median follow-up, 3.8 years; range, 0.5e13 years; Fig 4H, available at www.aaojournal.org). Table 4. Final Preferred Cox Proportional Hazards Multivariate Regression Model of Survival for 114 Children Younger than 18 Years and 185 Young Adults 18 to 24 Years of Age with Choroidal and Ciliary Body Melanoma Characteristic Chi-Square P Value TNM stage* Stage I (reference) Stage II Stage III Age groupy Genderz 1.96 7.84 4.71 4.26 0.16 0.005 0.030 0.041 Hazard Rate (95% Confidence Interval) 1.00 2.88 8.38 2.57 2.38 (0.65e12.7) (1.89e37.1) (1.09e6.01) (1.04e5.48) TNM ¼ tumor-node-metastasis. Trivariate analysis. Final preferred model: 2 log likelihood ¼ 263.58. *Categories: stage I ¼ 0, stage II ¼ 1, stage III ¼ 2. y Categories: age 0e17 years ¼ 0, age 18e24 years ¼ 1. z Categories: male ¼ 0, female ¼ 1. Multivariate Survival Outcome Univariate Cox regression confirmed the associations between survival and age group (hazard rate [HR], 2.64), congenital melanocytosis, extraocular extension, and TNM stage (Table 3, available at www.aaojournal.org). Of bivariate models that combined TNM stage with age group (HR, 2.66), melanocytosis (HR, 16.6), or gender (HR, 2.46; an independent predictor when combined with TNM stage), the one including melanocytosis fit best to our data (likelihood ratio test, 268.23 vs. 232.60, chisquare test with 1 degree of freedom; P < 0.0001). However, because melanocytosis was associated with only 3 deaths, this model is easy to fit, but may not be reproducible. We propose a trivariate model that combines TNM stage with age group (HR, 2.57) and gender (HR, 2.38; Table 4). The latter 2 variables are independent predictors of survival, and the model is preferred to the bivariate ones that include either of them (likelihood ratio test, 268.23 vs. 263.58, chi-square test with 1 degree of freedom; P ¼ 0.031). Alternative trivariate models with congenital melanocytosis and age or gender are shown in Table 3 (available at www.aaojournal.org). Discussion In our collaborative study, 52% of children younger than 18 years with CCBM came from the oldest (abortive) 5-year cohort of 15 to 17 years of age; 41% were 10 to 14 years of age, 11% were 5 to 9 years of age, and only 2% were younger than 5 years. A recent referral-based single-center series of 122 young patients (25% with iris melanomas) reported essentially identical percentages of 50%, 43%, 11%, and 3%.19 Our novel observation is that the cumulative frequency of having CCBM diagnosed increased steadily but slowly until 11 years of age, at which point a transition to a more than 10-times faster increase after 17 years of age took place, and 90% of CCBM in children were diagnosed during this transition period. A similar transition is known to occur again between 40 and 45 years of age.31 Taken together, observations suggest the existence of 3 periods of development of CCBM that may reflect age-dependent differences in initiation and progression of CCBM. We had strong evidence of a higher percentage of females compared with males among young adults younger 904 than 25 years. We believe this to be true also among children, in line with the prior meta-analysis18 and most series,13,15,18,21,23 but even our study was not large enough to confirm this (P ¼ 0.053). The reasons for the gender difference in young patients are unclear. Preexisting conditions predisposing to the development of CCBM were infrequent. Nevertheless, congenital oculo(dermal) melanocytosis that is estimated to affect 0.04% of white persons32 was present in 1.9% of children and 4.3% of young adults, percentages that were 47 and 107 times higher, respectively, than the general population frequency. These percentages were of the same order of magnitude as the 3% reported in the single-center series of 122 patients younger than 20 years of age.19 Neurofibromatosis type 1, also reported to be associated with UM,33e35 was present in 1.9% of children in our series, but was rare among young adults. Recently, a germline BAP1 mutation was found in 1 of 3 young adults younger than 25 years with a CCBM, but in neither of 2 children younger than 18 years.36 An affected 16-year old patient is on record, however.37 The primary CCBM was equally advanced in children and young adults. A smaller tumor thickness in children compared with young adults in the previous meta-analysis thus was likely the result of chance.18 As was the case with anatomic extent, the groups were comparable with regard to cell type. Cytogenetic analysis was available for a small but equal percentage of children and young adults. Monosomy 3 tended to be more frequent among children, but none of them were younger than 10 years, consistent with a hypothesis of 2 distinct periods of development of pediatric CCBM, the first one ending when the transition period in cumulative incidence begins. Our main hypothesis was that children have a favorable survival prognosis compared with young adults. The 10-year survival percentages were 92% and 80%, respectively, with a 12 percentage points difference, and KaplanMeier analysis led us to reject the null hypothesis of similar survival. Local recurrence was equally frequent and unlikely to lead to the survival difference between the Al-Jamal et al  Choroidal and Ciliary Body Melanoma Study groups, which also was maintained in multivariate analyses. The single-center series of 122 patients 20 years of age or younger also found a proportion of 91% surviving at 10 years.18 None of our 15 patients diagnosed before the age of 10 years so far have demonstrated metastases, again supporting our hypothesis of distinct periods of development of pediatric CCBM. Regarding our second hypothesis that males have a better survival prognosis than females, we could not discard the null hypothesis of similar survival after univariate analysis. However, gender became an independent predictor of survival when we adjusted for TNM stage, age group, or both. Notably among children, 8 females but only 1 male demonstrated metastases, and females tended to have a worse survival than males (85% vs. 100% at 10 years, the percentages we had postulated for both groups combined). However, the worse survival of girls began after 10 years of age, coinciding with the transition period in the cumulative frequency plot. Age and gender thus interact with regard to survival. Our third hypothesis that CBI would translate to worse life prognosis of young patients with CCBM clearly was not substantiated. Additionally, we confirmed that TNM stage is a predictor of survival among young patients with CCBM.18 This finding recently was supported by a single-center study of 43 patients 20 years of age or younger that reported metastasis only in TNM size categories T4 (very large) and T3 (large).14 We observed metastatic death in this age group also from 6 T2 (medium-sized) CCBMs (in 1 child and 5 young adults) and 1 T1 (small) CCBM (in 1 child), but the latter, along with another T1 melanoma in a young adult, metastasized more than 10 years after diagnosis. The stage-specific 10-year survival proportions of 100%, 96%, and 82% for children and 100%, 80%, and 72% for young adults suggest a better-than-average survival in both groups compared with that of 5403 patients with a CCBM across all ages, reported to be 88%, 75%, and 30%, respectively.26 We also were able to confirm that congenital melanocytosis is associated with increased mortality from CCBM. In a recent study of 7872 patients, the risk of metastasis was 1.9 to 2.8 times higher for those 230 patients who had melanocytosis, depending on its extent.38 In our study, the risk was 5.6 times higher in young patients and even higher by multivariate analysis. Our preliminary data on monosomy 3 support the observation that monosomy 3 with 8q gain predicts the highest risk for metastasis and disomy 3 predicts the lowest risk for metastasis39e42 also in young patients with CCBM. In contrast, cell type was unassociated with survival, although twice as many nonespindle-cell (mixed or epithelioid) than spindle-cell melanomas showed monosomy 3. In summary, children younger than 18 years with CCBM have a more favorable life prognosis than young adults 18 to 24 years of age, and TNM stage and gender are additional independent predictors of survival in young patients. We propose a hypothesis that the biology of CCBM may differ between children 10 years of age or younger and those older than 10 years at the time of diagnosis, based on our evidence of a lower incidence of CCBM and the apparent rarity of both monosomy 3 and metastasis in the younger of these age groups. Acknowledgments. The authors thank all personnel in the participating centers who collected patient data without participating in data analysis and writing the manuscript. References 1. Singh AD, Topham A. Incidence of uveal melanoma in the United States: 1973e1997. Ophthalmology 2003;110:956–61. 2. Virgili G, Gatta G, Ciccolallo L, et al. Incidence of uveal melanoma in Europe. Ophthalmology 2007;114:2309–15. 3. Kivelä T. Prevalence and epidemiology of ocular melanoma. In: Murray T, Boldt HC, eds. Ocular Melanoma: Advances in Diagnostic and Therapeutic Strategies. London: Future Science; 2014:21–38. 4. Singh AD, Bergman L, Seregard S. Uveal melanoma: epidemiologic aspects. Ophthalmol Clin North Am 2005;18: 75–84. 5. Pukrushpan P, Tulvatana W, Pittayapongpat R. Congenital uveal malignant melanoma. J AAPOS 2014;18:199–201. 6. Broadway D, Lang S, Harper J, et al. Congenital malignant melanoma of the eye. Cancer 1991;67:2642–52. 7. Shields CL, Kaliki S, Furuta M, et al. Clinical spectrum and prognosis of uveal melanoma based on age at presentation in 8,033 cases. Retina 2012;32:1363–72. 8. Singh AD, Wang MX, Donoso LA, et al. Genetic aspects of uveal melanoma: a brief review. Semin Oncol 1996;23: 768–72. 9. Singh AD, Shields CL, Shields JA, et al. Uveal melanoma and familial atypical mole and melanoma (FAM-M) syndrome. Ophthalmic Genet 1995;16:53–61. 10. Russo A, Coupland SE, O’Keefe M, Damato BE. Choroidal melanoma in a 7-year-old child treated by trans-scleral local resection. Graefes Arch Clin Exp Ophthalmol 2010;248: 747–9. 11. The Collaborative Ocular Melanoma Study (COMS) randomized trial of pre-enucleation radiation of large choroidal melanoma I: characteristics of patients enrolled and not enrolled. COMS report no. 9. Am J Ophthalmol 1998;125: 767–78. 12. Singh AD, Shields CL, Shields JA, Sato T. Uveal melanoma in young patients. Arch Ophthalmol 2000;118:918–23. 13. Shields CL, Shields JA, Milite J, et al. Uveal melanoma in teenagers and children. A report of 40 cases. Ophthalmology 1991;98:1662–6. 14. Petrovic A, Bergin C, Schalenbourg A, et al. Proton therapy for uveal melanoma in 43 juvenile patients: long-term results. Ophthalmology 2014;121:898–904. 15. Vavvas D, Kim I, Lane AM, et al. Posterior uveal melanoma in young patients treated with proton beam therapy. Retina 2010;30:1267–71. 16. Barr CC, McLean IW, Zimmerman LE. Uveal melanoma in children and adolescents. Arch Ophthalmol 1981;99:2133–6. 17. Apt L. Uveal melanomas in children and adolescents. Int Ophthalmol Clin 1962;2:403–10. 18. Al-Jamal RT, Kivelä T. Uveal melanoma among Finnish children and young adults. J AAPOS 2014;18:61–6. 19. Shields CL, Kaliki S, Arepalli S, et al. Uveal melanoma in children and teenagers. Saudi J Ophthalmol 2013;27: 197–201. 20. Stalford H. Children and the European Union: Rights, Welfare and Accountability. Oxford: Hart Publishing; 2012:21. 905 Ophthalmology Volume 123, Number 4, April 2016 21. Leonard BC, Shields JA, McDonald PR. Malignant melanomas of the uveal tract in children and young adults. Can J Ophthalmol 1975;10:441–9. 22. Pogrzebielski A, Orlowska-Heitzman J, RomanowskaDixon B. Uveal melanoma in young patients. Graefes Arch Clin Exp Ophthalmol 2006;244:1646–9. 23. Verdaguer J Jr. Prepuberal and puberal melanomas in ophthalmology. Am J Ophthalmol 1965;60:1002–11. 24. Kaliki S, Shields CL, Mashayekhi A, et al. Influence of age on prognosis of young patients with uveal melanoma: a matched retrospective cohort study. Eur J Ophthalmol 2013;23:208–16. 25. Malignant melanoma of the uvea. In: Edge S, Byrd D, Compton C, et al., eds. AJCC Cancer Staging Manual. 7th ed. New York: Springer; 2010:547–59. 26. Kujala E, Damato B, Coupland SE, et al. Staging of ciliary body and choroidal melanomas based on anatomic extent. J Clin Oncol 2013;31:2825–31. 27. United Nations Department of Economic and Social Affairs. World Population Prospects, 2012 Revision. Available at: http://esa.un.org/wpp/index.htm. Accessed June 15, 2013. 28. Hosmer DW Jr, Lemeshow S. Applied Survival Analysis. Regression Modeling of Time to Event Data. New York: Wiley & Sons; 1999:207–43. 29. Parmar MKB, Machin D. Survival Analysis. A Practical Approach. Chichester: John Wiley & Sons; 1995:129. 30. Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model. New York: Springer; 2000:127–52. 31. Kujala E, Mäkitie T, Kivelä T. Very long-term prognosis of patients with malignant uveal melanoma. Invest Ophthalmol Vis Sci 2003;44:4651–9. 32. Gonder JR, Ezell PC, Shields JA, Augsburger JJ. Ocular melanocytosis. A study to determine the prevalence rate of ocular melanocytosis. Ophthalmology 1982;89:950–2. 33. Honavar SG, Singh AD, Shields CL, et al. Iris melanoma in a patient with neurofibromatosis. Surv Ophthalmol 2000;45: 231–6. 34. Croxatto JO, Charles DE, Malbran ES. Neurofibromatosis associated with nevus of Ota and choroidal melanoma. Am J Ophthalmol 1981;92:578–80. 35. Friedman SM, Margo CE. Choroidal melanoma and neurofibromatosis type 1. Arch Ophthalmol 1998;116:694–5. 36. Cebulla CM, Binkley EM, Pilarski R, et al. Analysis of BAP1 germline gene mutation in young uveal melanoma patients. Ophthalmic Genet 2015;1–6. 37. Höiom V, Edsgard D, Helgadottir H, et al. Hereditary uveal melanoma: a report of a germline mutation in BAP1. Genes Chromosomes Cancer 2013;52:378–84. 38. Shields CL, Kaliki S, Livesey M, et al. Association of ocular and oculodermal melanocytosis with the rate of uveal melanoma metastasis: analysis of 7872 consecutive eyes. JAMA Ophthalmol 2013;131:993–1003. 39. Damato B, Dopierala JA, Coupland SE. Genotypic profiling of 452 choroidal melanomas with multiplex ligationdependent probe amplification. Clin Cancer Res 2010;16: 6083–92. 40. Cassoux N, Rodrigues MJ, Plancher C, et al. Genome-wide profiling is a clinically relevant and affordable prognostic test in posterior uveal melanoma. Br J Ophthalmol 2014;98: 769–74. 41. van Beek JG, Koopmans AE, Vaarwater J, et al. The prognostic value of extraocular extension in relation to monosomy 3 and gain of chromosome 8q in uveal melanoma. Invest Ophthalmol Vis Sci 2014;55:1284–91. 42. Versluis M, de Lange MJ, van Pelt SI, et al. Digital PCR validates 8q dosage as prognostic tool in uveal melanoma. PLoS One 2015;10:e0116371. Footnotes and Financial Disclosures Originally received: August 14, 2015. Final revision: December 13, 2015. Accepted: December 14, 2015. Available online: February 4, 2016. 12 Department of Ophthalmology, Hadassah-Hebrew University Medical Center, Jerusalem, Israel. Manuscript no. 2015-1387. 1 Ocular Oncology Service, Department of Ophthalmology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland. 2 Department of Ophthalmology, Institute Curie, Paris, France. 3 Ocular Oncology Service, Royal Liverpool University Hospital, Liverpool, United Kingdom. 4 Department of Cellular Pathology, Royal Liverpool University Hospital, Liverpool, United Kingdom. 5 Department of Ophthalmology, University of Lausanne, Jules-Gonin Eye Hospital, Fondation Asile des Aveugles (FAA), Lausanne, Switzerland. 6 Department of Ocular Oncology, Princess Margaret Hospital, University of Toronto, Toronto, Canada. 7 Department of Ophthalmology and Ocular Oncology, Jagiellonian University, Collegium Medicum, Krakow, Poland. 8 Department of Ophthalmology, Faculty of Medicine, University Duisburg-Essen, Essen, Germany. 9 Ocular Oncology Service, Department of Ophthalmology, Hacettepe University School of Medicine, Ankara, Turkey. 13 Department of Ophthalmology, Royal Hallamshire Hospital, Sheffield, United Kingdom. 14 Department of Ophthalmology, Poznán University of Medical Sciences, Poznán, Poland. 15 Department of Ophthalmology, Copenhagen University Hospital Glostrup, Copenhagen, Denmark. 16 Ocular Oncology Service, St Bartholomew’s and Moorfields Eye Hospital, London, United Kingdom. 17 Department of Ophthalmic Oncology and Radiology, Helmholtz Institute, Moscow, Russia. 18 Department of Ophthalmology, Oslo University Hospital-HF and University of Oslo, Oslo, Norway. 19 Department of Clinical Medicine, Section of Ophthalmology, University of Bergen, Bergen, Norway. 20 G. B. Bietti Foundation, Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS), Ocular Oncology and Toxicology Research Unit, Rome, Italy. 22 10 Department of Ophthalmology, Leiden University Medical Center, Leiden, The Netherlands. 23 11 24 Department of Ophthalmic Oncology, St. Erik’s Eye Hospital, Stockholm, Sweden. 906 Department of Ophthalmology, University of Padova, Padova, Italy. 21 Department of Ophthalmology, Croix-Rousse Hospital, Lyon, France. Department of Ophthalmology, Erasmus University Medical Center, Rotterdam, The Netherlands. Department of Ophthalmology, Catholic University of Rome, Rome, Italy. Al-Jamal et al  Choroidal and Ciliary Body Melanoma Study 25 Department of Ophthalmology, Ocular Oncology Unit, Valladolid University Hospital, Valladolid, Spain. Financial Disclosure(s): The author(s) have made the following disclosure(s): I.G.R.: Consultant Oraya Therapeutics Inc.; Moorfields Eye Hospital NHS Foundation Trust, London, United Kingdom; Lecturer Allergan; British Medical Association. R.T.A.: Grant The Finnish Eye Foundation, Eye and Tissue Bank Foundation, Evald and Hilda Nissi Foundation, Mary and Georg C. Ehrnrooth Foundation (Helsinki, Finland), and Emil Aaltonen Foundation (Tampere, Finland). E.B.: Board member and lecturer e Novartis. N.B.: Board member and lecturer e Novartis. T.T.K.: Grant Helsinki University Central Hospital Research Fund and Sigrid Jusélius Foundation (Helsinki, Finland). Supported by the Finnish Eye Foundation (R.T.A.); the Eye and Tissue Bank Foundation (R.T.A.); the Evald and Hilda Nissi Foundation (R.T.A.); the Mary and Georg C. Ehrnrooth Foundation (R.T.A.); the Emil Aaltonen Foundation, Finland (R.T.A.); Helsinki University Central Hospital Research Fund (grant no.: TYH2013316 [T.T.K.]); the Sigrid Jusélius Foundation, Helsinki, Finland (T.T.K.); and Fight for Sight, Copenhagen, Denmark (J.K.). The funding organizations had no role in the design or conduct of this research. Author Contributions: Conception and design: Al-Jamal, Cassoux, Desjardins, Kivelä Analysis and interpretation: Al-Jamal, Cassoux, Desjardins, Damato, Konstantinidis, Coupland, Heimann, Petrovic, Pe’er, Salvi, Rennie, Rospond-Kubiak, Kociecki, Folke Kiilgaard, Heegaard, Cohen, Kilic, Blasi, Saornil, Kivelä, Sagoo, Amiryan, Saakyan, Eide, Krohn, Midena, Parrozzani, Grange, Schalenbourg Data collection: Al-Jamal, Cassoux, Desjardins, Damato, Konstantinidis, Coupland, Heimann, Petrovic, Zografos, Velazquez-Martin, Krema, Bogdali, Markiewicz, Romanowska-Dixon, Metz, Biewald, Bornfeld, Kiratli, Bronkhorst, Jager, Marinkovic, Fili, Seregard, Frenkel, Pe’er, Salvi, Rennie, Rospond-Kubiak, Kociecki, Folke Kiilgaard, Heegaard, Cohen, Kilic, Blasi, Saornil, Kivelä, Sagoo, Amiryan, Saakyan, Eide, Krohn, Midena, Parrozzani, Grange, Schalenbourg Obtained funding: Al-Jamal, Kiilgaad, Kivelä Overall responsibility: Al-Jamal, Cassoux, Desjardins, Damato, Konstantinidis, Coupland, Heimann, Petrovic, Zografos, Velazquez-Martin, Krema, Bogdali, Markiewicz, Romanowska-Dixon, Metz, Biewald, Bornfeld, Kiratli, Bronkhorst, Jager, Marinkovic, Fili, Seregard, Frenkel, Pe’er, Salvi, Rennie, Rospond-Kubiak, Kociecki, Folke Kiilgaard, Heegaard, Cohen, Kilic, Blasi, Saornil, Kivelä, Sagoo, Amiryan, Saakyan, Eide, Krohn, Midena, Parrozzani, Grange, Schalenbourg Abbreviations and Acronyms: CBI ¼ ciliary body involvement; CCBM ¼ choroidal and ciliary body melanoma; CI ¼ confidence interval; HR ¼ hazard rate; TNM ¼ tumornode-metastasis; UM ¼ uveal melanoma. Correspondence: Rana’a T. Al-Jamal, MD, PhD, Department of Ophthalmology, Helsinki University Hospital, PL 220, Haartmaninkatu 4 C, FI-00029 HUS, Helsinki, Finland. E-mail: ranaa.aljamal@hus.fi. Pictures & Perspectives Central Retinal Artery Occlusion in a Teenager A 17-year-old boy presented with recurrent headache for several weeks and sudden vision loss on the left eye to light perception. On slit lamp examination partial aniridia with transillumination was identified (Fig 1A). Ophthalmoscopy and angiography of the left eye revealed a central retinal artery occlusion (Fig 1B, D). Computed tomography of the cardiovascular and cerebrovascular system discovered ectasia of all major arterial vessels, including carotid and cerebral arteries (Fig 1C). Genetic analysis revealed a missense mutation of the ACTA2 gene, encoding for a smooth muscle isoform of a-actin. Despite anticoagulant treatment, the vision of the patient remained at light perception. MATTHIAS FUEST, MD YASSIN DJALALI-TALAB, MD PETER WALTER, MD Department of Ophthalmology, RWTH Aachen University, Aachen, Germany 907