Patient with a normal right eye (OD) and oculo(dermal) melanocytosis in the left eye (OS), demonstrating heterochromia with a light brown iris (OD, A) and a dark brown iris (OS). Note the obvious scleral melanocytosis superiorly (C) and inferiorly (OS, D). Funduscopically, there was a normal choroid (OD, E) and melanocytosis with a thin melanoma and overlying orange pigment (OS, F).
Results calculated using log-rank test.
Shields CL, Kaliki S, Livesey M, Walker B, Garoon R, Bucci M, Feinstein E, Pesch A, Gonzalez C, Lally SE, Mashayekhi A, Shields JA. Association of Ocular and Oculodermal Melanocytosis With the Rate of Uveal Melanoma MetastasisAnalysis of 7872 Consecutive Eyes. JAMA Ophthalmol. 2013;131(8):993-1003. doi:10.1001/jamaophthalmol.2013.129
Ocular/oculodermal (oculo[dermal]) melanocytosis is a congenital periocular pigmentary condition that can lead to the development of uveal melanoma, estimated at 1 in 400 affected patients. In this study, patients with melanocytosis who developed uveal melanoma were found to have double the risk for metastasis compared with those without melanocytosis.
To determine the relationship of oculo(dermal) melanocytosis to the prognosis of patients with uveal melanoma.
Design, Setting, and Patients
Retrospective chart review of 7872 patients with uveal melanoma treated at the Ocular Oncology Service, Wills Eye Institute, from August 25, 1970, through August 27, 2008.
Enucleation, plaque radiotherapy, local resection, or thermotherapy.
Main Outcomes and Measures
Metastasis and death.
Of 7872 patients with uveal melanoma, oculo(dermal) melanocytosis was present in 230 (3%). The melanocytosis involved the sclera (92%), iris (17%), choroid (12%), eyelid (8%), and temporal fossa (1%). Eyes with melanoma and oculo(dermal) melanocytosis had a relative risk for metastasis 1.6 times greater compared with those with no melanocytosis (P < .001). Metastasis of uveal melanoma was 2.8 times higher in patients with iris melanocytosis (P < .001), 2.6 times higher with choroidal melanocytosis (P = .02), and 1.9 times higher with scleral melanocytosis (P < .001). By Kaplan-Meier estimates, metastasis in patients with oculo(dermal) melanocytosis vs no melanocytosis was 2% vs 1.8% at 1 year, 27% vs 15% at 5 years, and 48% vs 24% at 10 years (P < .001). By multivariable analysis, the factors predictive of metastasis in patients harboring uveal melanoma associated with oculo(dermal) melanocytosis were increased tumor thickness (P = .001) and the presence of subretinal fluid (P = .05), and the only factor predictive of death was increased tumor thickness (P = .009).
Conclusions and Relevance
Patients with uveal melanoma associated with oculo(dermal) melanocytosis have double the risk for metastasis compared with those with no melanocytosis. All patients with oculo(dermal) melanocytosis should undergo ophthalmic examination and imaging on a twice-yearly basis because this could help with the early detection of melanoma.
In 1939, Ota1 and Ota and Tanino2 described a periocular pigmentary condition termed nevus fusco-caeruleus ophthalmomaxillaris, more recently named ocular or oculodermal melanocytosis. This condition has been labeled with several names including melanosis oculi, nevus of Ota, and oculocutaneous melanosis. This congenital pigmentary abnormality consists of excess melanocytes within the periocular skin, sclera, uvea, orbit, meninges, palate, or tympanic membrane. In most cases, the pigmentation is unilateral. The melanocytic pigmentation can involve the skin or eye alone but can also be seen as a combination of the two, thus the inclusive term oculo(dermal) melanocytosis has been promoted. The main concern with oculo(dermal) melanocytosis is the risk for development of melanoma, predominantly in the uvea.3- 21
Oculo(dermal) melanocytosis is rare in the general population. Cowan and Balistocky6 reviewed 25 000 patients from an eye clinic at the Philadelphia General Hospital between 1956 and 1960, revealing only 4 cases of melanocytosis. They estimated that the incidence was 1 in 6200 in the white and African American combined population.6 In 1982, Gonder et al10 estimated that oculo(dermal) melanocytosis affected 0.04% of the white population and approximately 1.4% of those with uveal melanoma. In 1998, Singh et al15 calculated that the lifetime estimate for a white patient with oculo(dermal) melanocytosis to develop uveal melanoma was 1 in 400.
To further explore the relationship with uveal melanoma, Gonder et al11 reported 17 patients with oculo(dermal) melanocytosis among 1250 white patients with uveal melanoma and noted that the melanoma was similar in size, cell type, and metastatic risk compared with those without melanocytosis. In this report, we pursued a comprehensive experience with 7872 patients with uveal melanoma and analyzed the impact of oculo(dermal) melanocytosis on patient risk for melanoma-related metastasis and death.
A retrospective review of patients with clinical diagnosis of uveal melanoma managed at the Ocular Oncology Service, Wills Eye Institute, between August 25, 1970, and August 27, 2008, was conducted. Institutional review board approval was obtained for this study. The study and data collection were compliant with the principles of the Declaration of Helsinki. All patients included in this study had detailed information regarding the presence and extent of ocular melanocytosis. All patients were examined by one of the senior authors (C.L.S. and J.A.S.).
The demographic data included patient age at diagnosis (years), sex, and race/ethnicity (white, African American, Hispanic, Asian, Native American, Middle Eastern, and Asian Indian). A history of skin, orbit, or brain melanoma was noted. The following clinical data were obtained: intraocular pressure, ocular or oculodermal melanocytosis (present or absent; herein referred to as oculo[dermal] melanocytosis to include both ocular and oculodermal types), location of melanocytosis (temporal fossa, eyelid, conjunctiva, sclera, iris, choroid, or palate), laterality of melanocytosis with uveal melanoma (ipsilateral or contralateral), tumor laterality (unilateral or bilateral), general location of tumor epicenter (iris, ciliary body, or choroid), quadrant location of tumor epicenter (superior, nasal, inferior, temporal, or macula), clock-hour location of the tumor epicenter (1:00 to 12:00), anteroposterior location of tumor epicenter (macula, macula equator, equatorora serrata, ciliary body, or iris), distance of posterior tumor margin to optic disc margin and foveola (millimeters), largest tumor basal dimension and thickness (millimeters), tumor configuration (dome, mushroom, tapioca, or plateau), color (pigmented, nonpigmented, or mixed), Bruch membrane rupture, subretinal fluid, intraocular hemorrhage, and extraocular extension. Tumor basal diameter was measured using indirect ophthalmoscopy and ocular ultrasonography. Tumor thickness was measured by ocular ultrasonography. All findings were documented appropriately with anterior and/or posterior segment photography and color-coded fundus drawing, fluorescein angiography, and A-scan and B-scan ultrasonography.
Melanoma management was provided with standard methods (enucleation, plaque radiotherapy, local resection, or thermotherapy) following informed consent. Monitoring for metastatic disease was performed by a medical oncologist with twice-yearly physical examination and liver function tests, once-yearly liver imaging (magnetic resonance imaging, computed tomography, or ultrasonography), and chest imaging (chest radiograph or computed tomography). Outcomes of metastasis and death were recorded. Notification of melanoma-related metastasis and melanoma-related death was obtained, if possible, in letter or telephone format from medical oncologists, family physicians, family, or the patient. Specific death information included date and cause of death.
The demographics and tumor characteristics of patients with uveal melanoma were summarized according to the presence or absence of oculo(dermal) melanocytosis and proportions were analyzed using Fisher exact or χ2 tests. Data collected on continuous or ordinal scale, including age (years), tumor base (millimeters), and tumor thickness (millimeters), were expressed as mean, median, minimum, and maximum, and they were evaluated with independent-sample t test for comparison according to the presence or absence of oculo(dermal) melanocytosis. A comparison of tumor distance to foveola (millimeters) and optic nerve (millimeters), as well as clock-hour location of tumor epicenter, was performed using Wilcoxon rank-sum test.
Kaplan-Meier analysis was performed to estimate the cumulative probability of metastasis and death according to the presence or absence of oculo(dermal) melanocytosis at 1, 3, 5, 7, 10, 15, and 20 years. A comparison of survival distribution according to the presence or absence of oculo(dermal) melanocytosis was analyzed by log-rank test. An estimate of metastasis and death according to the presence or absence of oculo(dermal) melanocytosis was analyzed by hazard ratio (HR) with 95% CIs.
Factors predictive of metastasis and death due to uveal melanoma according to the presence or absence of oculo(dermal) melanocytosis were evaluated using the Cox proportional hazard model with HRs and 95% CIs. Significant factors were identified by univariable analysis at 5% level of significance, and these were subsequently considered for multivariable analysis using Cox proportional hazard model by forward stepwise method. The factors significant at the 0.05 level on multivariable analysis were reported.
Of 7872 patients with uveal melanoma included in this study, 230 (3%) displayed oculo(dermal) melanocytosis. The mean age at presentation of uveal melanoma in patients with oculo(dermal) melanocytosis was 56 years compared with 58 years in those with no oculo(dermal) melanocytosis (Table 1). The demographic features are listed in Table 1; there was no statistical difference in age, sex, or involved eye comparing those with melanocytosis present vs absent. The only statistical difference involved patient race/ethnicity, and it was found that nonwhite individuals represented 7% of those with uveal melanoma and melanocytosis compared with 2% of those without melanocytosis (P < .001, χ2 test).
The tissue sites of oculo(dermal) melanocytosis included the sclera (92%), iris (17%), and choroid (12%) (Table 2; Figure 1). A comparison of tumor features in eyes with melanoma and melanocytosis present vs melanocytosis absent is listed in Table 3. The mean tumor basal diameter of 11 mm and mean tumor thickness of 6 mm were similar in both groups. Significant differences (melanoma with melanocytosis present vs absent) were in the location of the tumor in the macula (7% vs 4%; P = .007; χ2 test), temporal quadrant (23% vs 28%; P = .007; χ2 test), and diffuse (multiple quadrants) (7% vs 3%; P = .007; χ2 test). Other significant tumor differences included brown tumor color (pigmented) (69% vs 54%; P < .001; χ2square test) and extraocular extension (6% vs 3%; P = .006; Fisher exact test) (Table 3).
Kaplan-Meier estimates of metastasis in patients with uveal melanoma with oculo(dermal) melanocytosis present vs absent revealed 2% vs 1.8% at 1 year, 27% vs 15% at 5 years, 48% vs 24% at 10 years, and 48% vs 36% at 20 years (P < .001, log-rank test) (Figure 2). Kaplan-Meier estimates of death revealed 0.7% vs 0.8% at 1 year, 13% vs 8% at 5 years, 19% vs 12% at 10 years, and 19% vs 18% at 20 years for those with oculo(dermal) melanocytosis present vs absent (P = .16, Log-rank test) (Figure 3). Metastasis was 1.6 times higher in patients with uveal melanoma associated with oculo(dermal) melanocytosis compared with those with no melanocytosis (P < .001, Cox proportional hazard model) (Table 4). Using the Cox proportional hazard model, the relative risk for metastasis of uveal melanoma differed depending on the main site of pigmentation, with iris melanocytosis at 2.8 (P < .001), choroidal melanocytosis at 2.6 (P = .02), and scleral melanocytosis at 1.9 (P < .001). There was no significant difference in the relative risk for metastasis if the main site of melanocytosis was the eyelid (P = .16) or temporal fossa (P = .08).
Using the Cox proportional hazard model for the entire cohort, multivariable analysis for melanoma metastasis revealed significant factors of older age (P < .001; HR, 1.13 per 10 years), the presence of skin melanoma (P = .01; HR, 1.99), anteroposterior tumor epicenter in ciliary body (P = .02; HR, 2.44 compared with the iris), ora serrata to equator (P = .02; HR, 2.42 compared with the iris), tumor base (P < .001; HR, 1.14 per 1-mm increase), tumor thickness (P < .001; HR, 1.08 per 1-mm increase), brown tumor color (P < .001; HR, 1.48), subretinal fluid (P = .005; HR, 1.26), and presence of oculo(dermal) melanocytosis (P = .003; HR, 1.63) (Table 4). Specific factors related to metastasis in eyes with melanocytosis present vs absent are listed in Table 5. In patients with uveal melanoma with oculo(dermal) melanocytosis, factors (Cox proportional hazard model) predictive of metastasis included increasing tumor thickness (P = .001; HR, 1.19 per 1-mm increase) and subretinal fluid (P = .05; HR, 2.66) (Table 5).
Multivariable analysis (Cox proportional hazard model) for melanoma-related death in the entire cohort included increasing age (P < .001; HR, 1.16 per 10-year increase), increasing tumor base (P < .001; HR, 1.15 per 1-mm increase), increasing tumor thickness (P < .001; HR, 1.08 per 1 mm increase), and brown pigmented tumors (P = .003; HR, 1.59) (Table 6). In patients with uveal melanoma with oculo(dermal) melanocytosis, the only factor predictive of death was tumor thickness (P = .009; HR, 1.24 per 1-mm increase) (Table 7).
Oculo(dermal) melanocytosis is an uncommon condition but represents a profound predisposing factor for uveal melanoma. From a clinical perspective, in a general population of 13 150 patients, Gonder et al10 found melanocytosis in 1 in 6915 African American patients (0.014%) and 2 of 5251 white patients (0.038%). From a pathology perspective, Velazquez and Jones12 noted 15 of 1210 eyes enucleated for uveal melanoma had melanocytosis, giving a prevalence of 1.24% in their series. In our series, we noted clinical evidence of melanocytosis in 230 of 7872 patients (3%) with uveal melanoma. Oculo(dermal) melanocytosis can affect the entire uvea and sclera as a diffuse process in all quadrants or can affect only a small portion as a partial or sector pigmentation.4,21 Sector melanocytosis carries the potential for melanoma within the area of excessive pigmentation.21
In some cases, the melanocytosis was obvious with periocular dermal pigmentation or slate gray scleral pigmentation with dark brown heterochromia, contrasting the opposite eye. In other cases, the pigmentation was subtle, requiring detailed inspection of the periocular skin or sclera, with only a mild patch of pigment. Inspection of sclera by external approach and slit lamp biomicroscopy can be revealing of minor melanocytosis. In our oncology practice, all patients with uveal melanoma are carefully inspected at the initial examination for melanocytosis and a comment on the results of examination is recorded on each chart. Owing to diligent recording of this data in patient medical records for more than 40 years, this unique study could be performed.
In this analysis, we explored the relationship of oculo(dermal) melanocytosis on patient survival prognosis. Multivariable analysis for risks for metastatic disease disclosed several anticipated factors22 such as greater melanoma base and thickness, ciliary body or peripheral choroidal tumor location, and older patient age. However, another significant factor for metastatic disease was the presence of oculo(dermal) melanocytosis. Overall, patients with uveal melanoma and melanocytosis had a relative risk for metastasis of 1.6 times greater compared with those without melanocytosis. From a Kaplan-Meier perspective, the 10-year rate of metastasis was 48% for those with melanocytosis compared with 24% for those without melanocytosis. The location of melanocytosis was also important because the relative risk for melanoma-related metastasis with choroidal melanocytosis was 2.6, iris melanocytosis was 2.8, and scleral melanocytosis was 1.9. Eyelid melanocytosis and temporal fossa melanocytosis were nonsignificant factors. When specifically focusing on those eyes with melanoma and melanocytosis, factors for metastasis included greater tumor thickness and presence of subretinal fluid.
These findings underscore the importance of careful examination for all patients with oculo(dermal) melanocytosis because thin tumors could blend into the background uveal pigmentation and not be clinically manifest. In all cases of choroidal melanocytosis, we perform fundus photography to confirm the absence of tumor, fundus autofluorescence for the detection of overlying retinal pigment epithelial–related lipofuscin abnormalities that could signal an underlying tumor or subretinal fluid, optical coherence tomography with enhanced depth imaging to inspect the retina for subretinal fluid and the choroid for thickening, and ultrasonography to confirm the lack of mass. For iris melanocytosis, we advise imaging the anterior segment with slitlamp photography and anterior segment optical coherence tomography. The monitoring of ciliary body melanocytosis is more challenging because this tissue is hidden behind the iris. In such cases, we use scleral depressed indirect ophthalmoscopic examination and image with standard ultrasonography and screen with either ultrasound biomicroscopy or ciliary body anterior segment optical coherence tomography, both of which can document ciliary body anatomy with contact or noncontact methods, respectively.
Despite the importance of melanocytosis as a factor for metastasis in this analysis, it did not reach significance in the outcome of melanoma-related death. Multivariable factors for death included older age, greater tumor base and thickness, and brown tumor color. However, when evaluating only those eyes with uveal melanoma and oculo(dermal) melanocytosis for death, the single risk factor was greater tumor thickness. The importance of tumor thickness22 with this condition cannot be overemphasized.
We acknowledge that there are limitations to this analysis including its retrospective nature; however, all patients were examined by one of the senior authors (C.L.S. and J.A.S.) over the 40 years of data collection, thus uniformity of examination was maintained. Despite the rigorous data collection, information on patient metastasis and death relied on letter or telephone communication by physicians, families, or patients. Much of this data was gathered at the patient annual examination. However, notification of death relied on physicians and families and was occasionally incomplete. Although we have correlated melanocytosis with higher metastatic rate, there may remain unknown factors that directly predispose these patients such as the underlying cellular or genetic constitution of the melanocytosis. Horgan et al23 documented chromosome 3 monosomy in uveal melanoma and its absence in melanocytosis in a single case. Future study of the genetic composition of congenital melanocytosis will hopefully be revealing to its relationship with melanoma.
In summary, oculo(dermal) melanocytosis is an important predisposing factor for uveal melanoma.3,4 Uveal melanoma carries prominent risk for metastasis in 25% by 10 years, with each millimeter increase in thickness promoting increasing metastatic rate.22 Recognition of congenital melanocytosis early in life and understanding of the possible magnitude of monitoring these patients are emphasized. We suggest that all patients with oculo(dermal) melanocytosis undergo ophthalmic examination and imaging on a twice-yearly basis. Symptomatic patients should be examined promptly, especially if photopsia, floaters, visual acuity loss, or visual field abnormalities are noted because these can be a manifestation of uveal melanoma.22 This strategy could assist in the early detection of melanoma in this at-risk population. Recognition of thin melanoma in eyes with melanocytosis is critical because greater thickness and the presence of subretinal fluid impart greater risk for ultimate metastatic disease.
Corresponding Author: Carol L. Shields, MD, Ocular Oncology Service, Ste 1440, Wills Eye Institute, 840 Walnut St, Philadelphia, PA 19107 (firstname.lastname@example.org).
Submitted for Publication: September 3, 2012; final revision received December 22, 2012; accepted December 28, 2012.
Published Online: May 16, 2013. doi:10.1001/jamaophthalmol.2013.129.
Author Contributions: Dr C. L. Shields had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by the Eye Tumor Research Foundation, Philadelphia (Drs C. L. Shields and J. A. Shields); Mellon Charitable Giving from the Martha W. Rogers Charitable Trust (Dr C. L. Shields); Lift for a Cure, Morrisdale (Drs C. L. Shields and J. A. Shields); and the Lucille Wiedman Fund for Pediatric Eye Cancer, Philadelphia (Drs J. A. Shields and C. L. Shields).
Disclaimer: The funders had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review or approval of the manuscript.
Presentation: This study will be presented at the Eugene Chalfin Lecture; June 13, 2013; Brooklyn, New York.
Additional Contributions: Rishita Nutheti, PhD, provided statistical analysis.