A, Location in the palpebral conjunctiva. B, Location in the bulbar conjunctiva. C, Shown are examples of BRAF wild-type (WT) and mutated (MUT) tumors. Cutoff lines (at 6336 for BRAF V600E and 4103 for BRAF V600K) were determined using the mutation-positive control (MUT CTRL) and mutation-negative control (WT CTRL) DNA samples. The top axes of the graphs show examples of BRAF-mutated conjunctival melanoma samples (28 and 41) and BRAF wild-type conjunctival melanoma samples (19 and 42) compared with MUT CTRL and WT CTRL DNA samples. FAM indicates 1.1 µL of mutation primer or probe mix.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Larsen A, Dahmcke CM, Dahl C, et al. A Retrospective Review of Conjunctival Melanoma Presentation, Treatment, and Outcome and an Investigation of Features Associated With BRAF Mutations. JAMA Ophthalmol. 2015;133(11):1295–1303. doi:10.1001/jamaophthalmol.2015.3200
Large studies investigating clinical presentation and treatment in primary conjunctival melanoma (CM) are rare. Clinicopathological characteristics of BRAF-mutated CM have not been studied thoroughly.
To determine the associations of clinicopathological tumor features and treatment with local recurrence, metastasis, and mortality and to determine the association of BRAF mutations with these features.
Design, Setting, and Participants
Population-based cohort study at the Eye Pathology Institute, Copenhagen, Denmark. Participants included 139 patients with primary CM in Denmark from January 1, 1960, to December 31, 2012. For BRAF analysis, all patients with available formalin-fixed, paraffin-embedded tumor samples from January 1, 1994, to December 31, 2012, were included.
Main Outcomes and Measures
BRAF mutations, local recurrence, regional and distant metastasis, melanoma-related mortality, and all-cause mortality were examined.
A poor prognosis of tumors involving the extrabulbar conjunctiva and adjacent tissue structures was confirmed in multivariable Cox proportional hazards regression models. Patients undergoing incisional biopsy more frequently developed metastasis (hazard ratio [HR], 2.46; 95% CI, 1.08-5.58; P = .03). Excision without adjuvant treatment was associated with local recurrence (HR, 1.97; 95% CI, 0.11-3.48; P = .02), metastatic disease (HR, 2.51; 95% CI, 1.07-5.91; P = .03), and all-cause mortality (HR, 1.80; 95% CI, 1.05-3.08; P = .03). BRAF mutations were identified in 19 of 47 primary CMs (40.4%) and were more frequent in younger patients (P = .005), less frequent in the extrabulbar conjunctiva (P = .05), more frequently classified as T1 tumors (P = .03), and rarely manifested with primary acquired melanosis (P = .001) or with a uniformly pigmented lesion (P = .006). Distant metastases developed in 6 of 19 BRAF-mutated CMs (31.6%) as opposed to 1 of 28 BRAF wild-type CMs (3.6%). No definitive association with distant metastasis was seen in multivariable Cox proportional hazards regression models.
Conclusions and Relevance
Incisional biopsy and excision without adjuvant therapy were associated with a poor outcome in patients with CM. Extrabulbar location was also associated with a poor outcome in multivariable analysis. BRAF-mutated CMs were frequent in younger patients and were rare in tumors involving the extrabulbar conjunctiva. Despite a more favorable location, BRAF-mutated tumors may be associated with more frequent distant metastasis.
Conjunctival melanoma (CM) is an uncommon malignant neoplasm that makes up 5% of ocular melanomas and has an associated mortality of up to 30%.1-3 Several clinical factors have been associated with this high mortality, including tumor size, extrabulbar location (ie, fornical or palpebral), or caruncular location and invasion of adjacent tissues.2,4-7 The management of CM is controversial and has primarily evolved from physician experience and the findings of case series.2,6,8 Previously, many patients having primary conjunctival malignant melanoma were treated with orbital exenteration.9 Local excision and cryotherapy, adjuvant topical mitomycin C, and brachytherapy are widely used primary treatment modalities.4,6,7,10,11 Incisional biopsies and excision without adjuvant therapy are discouraged because these methods appear to correlate with increased tumor recurrence rates, CM metastasis, and CM-related mortality.2,6,8 Recurrent disease is common and has been reported to occur in approximately 50% of patients, and local recurrence has been associated with increased mortality.1,6,7,10,11
In cutaneous melanoma, constitutive activation of the Ras-Raf-MEK-ERK (MAPK) signaling pathway via somatic activating mutations in BRAF (OMIM 164757) is present in 40% to 60% of cases.12BRAF mutations in cutaneous melanoma have been associated with a distinct clinical presentation and a poor clinical outcome.13 As a consequence, targeted treatment with BRAF and MEK inhibitors has improved survival in patients with BRAF-mutated metastatic melanoma.14
Investigations of genetic abnormalities in CM have revealed BRAF mutations in up to 50% of cases.15-21 However, the associations of BRAF mutations in CM with clinical presentation and outcome are unclear. The objective of this study was to determine the associations of clinical presentation and treatment of CM with BRAF mutations, onset of local recurrence, metastasis, and mortality in a clinically well-defined population-based cohort of CM diagnosed in Denmark over a 52-year period.
The objective of this study was to determine the associations of clinical presentation and treatment with prognosis and BRAF mutations in primary conjunctival melanomas.
Extrabulbar tumor location and invasion of adjacent tissue structures were poor prognostic features.
Incisional biopsy and excision without adjuvant therapy were associated with a poor outcome.
Younger age at diagnosis, bulbar or caruncular tumor location, T1 tumor, lack of clinical melanosis, and mixed or nonpigmented tumor color were features associated with BRAF mutations.
Patients with BRAF-mutated tumors seem to have increased risk of distant metastatic disease.
In total, 139 patients were identified as being diagnosed as having CM from January 1, 1960, to December 31, 2012, by searching the pathology reports at the Eye Pathology Institute, Copenhagen, Denmark (an institution providing nationwide service to Danish hospitals and ophthalmology practices) and the Danish Registry of Pathology. Conjunctival melanoma cases with origin in the conjunctiva or eyelid margin were included. Patients with primary cutaneous melanomas were excluded. Clinical patient records, pathology reports, and death certificates were collected and reviewed, and the Danish Registry of Pathology and the Danish Register of Causes of Death were searched to collect information on clinical presentation, treatment, local recurrence (recorded if recurrence was reported in the same eye at any site and with a histopathological diagnosis of CM), regional and distant metastasis, and cause of death. Tumor thickness (measured from the epithelial surface to the deepest tumor cell) and origin (primary acquired melanosis with atypia [PAM+], PAM+ and nevus, nevus, or de novo) were reviewed using the original hematoxylin-eosin stains. The TNM classification in the seventh edition of the Cancer Staging Manual by the American Joint Committee on Cancer was used.22 Patient data and formalin-fixed, paraffin-embedded (FFPE) tissue were available from 47 consecutive CM samples identified in more recent years (January 1, 1994, to December 31, 2012). All-cause mortality, melanoma-related mortality, development of distant metastasis, development of any (regional or distant) metastasis, and local recurrence were considered primary outcomes.
This study adhered to the tenets of the Declaration of Helsinki, with institutional review board approval and health information privacy agency approval obtained from the Danish Data Protection Agency and the Local Scientific Ethics Committee. The Danish National Ethics Committee did not require informed consent because most patients were deceased at the time of study approval.
DNA was extracted from the 47 FFPE tumor samples using a solution and tissue kit (QIAamp deparaffinization solution and DNA FFPE tissue kit; Qiagen) according to the manufacturer’s instructions. Detection and quantification of BRAF mutations (c.1799T > A, p.V600E and c.1798_1799GT > AA, p.V600K) were performed by droplet digital polymerase chain reaction (ddPCR) analysis using an available system (QX200; Bio-Rad Laboratories) and hydrolysis probe–based assays (PrimePCR ddPCR mutation detection assays; Bio-Rad Laboratories). The PCR mixture contained 11 µL of ddPCR Supermix (Bio-Rad Laboratories) for probes, 1.1 µL of mutation primer or probe mix (FAM dye), 1.1 µL of wild-type primer or probe mix (HEX dye), and 5 µL of DNA (11.6-508.4 ng/µL) in a total volume of 22 μL. Twenty microliters of this mixture and 70 µL of droplet generation oil were transferred to different wells of a droplet generation cartridge. After emulsification of the samples using the droplet generator, samples were transferred to a 96-well PCR plate and amplified for 40 cycles at 94°C for 30 seconds and at 55°C for 60 seconds. Droplets were analyzed on the droplet reader, and a software program (QuantaSoft, version 18.104.22.168; Bio-Rad Laboratories) was used for analyzing DNA concentrations. Cutoff settings were determined using mutation-positive and mutation-negative control DNA samples.
The associations of patient characteristics, clinicopathological features, treatment, and BRAF mutations with primary outcomes were assessed in Cox proportional hazards regression models. Analyses were adjusted for possible confounding by also including sex, age, tumor location, and local invasion as covariates in multivariable Cox proportional hazards regression models.
Fisher exact test, χ2 test, Mann-Whitney test, and log-rank test were used, where appropriate, to compare BRAF mutation status with patient characteristics, clinicopathological features, and outcome. Statistical calculations were performed using software programs (SPSS, version 20; SPSS Science and SAS, version 9.4; SAS Institute Inc).
Patient data and status were available for 132 of 139 patients with CM (Table 1), including 70 women and 62 men. The median age at diagnosis was 67.2 years (age range, 14-100 years). The median follow-up period was 6.1 years (range, 4 months to 44 years).
Most CMs (66.2% [86 of 130]) were confined to the epibulbar conjunctiva. Local invasion occurred in 13.1% (17 of 130), most frequently to the eyelid skin (82.3% [14 of 17]). Most CMs (58.3% [74 of 127]) were uniformly pigmented lesions, and more than half (54.1% [66 of 122]) were initially seen with PAM.
Data on primary treatment were obtained for 129 patients. Incisional or diagnostic biopsy was performed before other treatment in 27 patients. Most patients were treated with excision of the tumor combined with adjuvant therapy, which consisted of cryotherapy (n = 52), cryotherapy and topical chemotherapy (n = 3), brachytherapy with a ruthenium 106 plaque (n = 3), and irradiation with strontium 90 (n = 12). Thirty-seven patients were treated with excision without any further treatment, and enucleation or exenteration was the choice of treatment in 22 patients. Sentinel lymph node biopsy was performed in 3 patients, all initially seen with eyelid skin involvement. Among 129 cases, the histopathological origin was PAM+ in 62.0% (n = 80), nevus in 25.6% (n = 33), de novo in 10.9% (n = 14), and PAM+ and nevus in 1.6% (n = 2). All CMs invaded the substantia propria, and 57.6% (72 of 125) exceeded 1.5 mm in thickness. In total, 73.0% (27 of 37) of CMs involving the extrabulbar conjunctiva exceeded 1.5 mm in thickness. Local recurrence developed in 61 of 126 patients (46.2%), and regional metastasis was recorded in 16 of 132 patients (12.1%) and most commonly involved the preauricular lymph nodes, lymph nodes within the parotid and submandibular glands, and unspecified lymph nodes in the head and neck region. Distant metastasis developed in 30 of 132 patients (22.7%) and most commonly involved skin (cutaneous or subcutaneous) at several different sites (scalp, neck, shoulder, or leg), parotid or submandibular glands, brain, liver, lungs, or multiple sites. Thirteen of 132 patients (9.8%) developed subsequent or concurrent regional and distant metastasis. Fifty of 132 patients (37.9%) were alive at the end of the follow-up period, 33 of 132 (25.0%) died of melanoma-related causes, 45 of 132 (34.1%) died of other causes, and 4 of 132 (3.0%) died of an unknown cause.
Extrabulbar location of tumors (with involvement of the fornix, palpebral conjunctiva, or eyelid margin) was associated with increased melanoma-related and all-cause mortality hazard ratios (HRs) of 2.72 (95% CI, 1.12-6.59; P = .03) and 2.21 (95% CI, 1.25-3.90; P = .006), respectively. Invasion of adjacent tissue was associated with the development of distant metastasis or any metastasis (HR, 3.38; 95% CI, 1.21-9.45; P = .02 and HR, 3.04; 95% CI, 1.13-8.18; P = .03). Six CM cases involving the caruncle showed no significant associations with local recurrence, metastasis, or mortality. No significant associations of tumor origin with outcome were identified. However, increased risk of all-cause mortality in melanoma developed de novo (HR, 1.94; 95% CI, 0.98-3.84; P = .06) was identified. Compared with PAM+, melanoma from a nevus may have been associated with a lower risk of local recurrence (HR, 0.41; 95% CI, 0.16-1.07; P = .07) and any metastasis (HR, 0.51; 95% CI, 0.25-1.02; P = .06).
Patients undergoing incisional biopsy more frequently developed distant metastatic disease (HR, 2.46; 95% CI, 1.08-5.58; P = .03). Patients treated with excision without adjuvant treatment more frequently developed local recurrence (HR, 1.97; 95% CI, 0.11-3.48; P = .02) and distant metastasis (HR, 2.51; 95% CI, 1.07-5.91; P = .03) and had increased melanoma-related mortality (HR, 2.27; 95% CI, 0.97-5.33; P = .06) and all-cause mortality (HR, 1.80; 95% CI, 1.05-3.08; P = .03).
BRAF mutation analysis was performed in a subset of 47 consecutive patients (Table 2). Analysis of tumor DNA using ddPCR led to identification of BRAF mutations in 19 of 47 patients (40.4%). Mutations were V600E type in 15 of 19 (78.9%) and V600K type in 4 of 19 (21.1%) (Figure). BRAF mutations were more common in younger patients (P = .005, Mann-Whitney test), were less common in the extrabulbar conjunctiva (P = .05, Fisher exact test), and rarely were initially seen with PAM (P = .001, χ2 test). The tumor most commonly had a mixed pigmented or nonpigmented appearance (P = .006, χ2 test). BRAF-mutated CMs were more frequently classified as T1 tumors (P = .03, χ2 test). Conjunctival melanomas with a histopathological origin in PAM+ were BRAF mutated in 36.1% (13 of 36), whereas CMs associated with a nevus were BRAF mutated in 66.7% (6 of 9).
Distant metastasis occurred in 7 of 47 patients (14.9%) and more frequently developed in patients with BRAF-mutated CM (n = 6) as opposed to BRAF wild-type CM (n = 1). A Kaplan-Meier model showed decreased distant metastasis–free survival in BRAF-mutated CM (P = .02, log-rank test). The presence of BRAF mutations in primary CM showed no association with local recurrence, regional metastasis, or mortality (Table 3). Multivariable Cox proportional hazards regression showed no association with local recurrence, metastasis, or mortality (Table 1).
We present a retrospective investigation of clinical prognostic features, treatment, and BRAF mutations in patients with CM in Denmark. Our findings represent unique national data acquired for more than 50 years.
The sex ratio of patients with CM was approximately 1:1, and the median age at diagnosis was 67.2 years, which is slightly older than is reported in other Nordic countries.3,7 Most CMs were located in the bulbar conjunctiva, similar to what has been reported previously.3,6,7,10,11,23,24 Conjunctival melanomas involving the extrabulbar conjunctiva and manifesting with involvement of adjacent tissue structures were associated with a poor prognosis. The higher risk of metastasis and melanoma-related death for extrabulbar CM is in agreement with other studies.6,7,10,11,23-27 However, increased mortality for CM cases involving the plica and caruncle could not be confirmed.10,24 Significant associations with outcome were not identified for tumor diameter, which could be because of the few measurements obtained. Increasing tumor thickness was not associated with a poor outcome, which may be due to the adjustment for tumor location in multivariable analysis because 73.0% (27 of 37) of extrabulbar tumors exceeded 1.5 mm.
All tumors invaded the substantia propria. However, depth of invasion was excluded in the present study because it is not included in the current TNM classification.22
Caution against performing incisional biopsies has been stated in several studies.2,6,23,28 However, to our knowledge, the present study is the first to show an association of incisional biopsies with distant metastasis regardless of subsequent treatment. Therefore, the use of incisional biopsies is discouraged.
Despite the retrospective nature of our study, a poor prognosis in patients treated with excision alone was observed, which has also been reported in other case series.6,23 Tumor origin may influence the poor outcome observed in patients treated with excision alone because of the tendency for more frequent local recurrences and metastases in CM originating in PAM+ and the fact that treatment practices changed over the considerably long follow-up period. Therefore, our results also reflect the variation in practice over time in the performance of incisional biopsies and excision without adjuvant therapy.
We had insufficient data to evaluate differences in surgical techniques (lamellar scleroconjunctival excision vs simple conjunctival excision) or prognostic information from sentinel lymph node biopsies. Recent results showed improvement in local recurrence rates when brachytherapy alone or in combination with local chemotherapy was performed.4,8,29 This treatment is the preferred choice at our institution.
Forty-seven CMs were analyzed for BRAF mutations in the present study, representing a large number of patients compared with previous studies. BRAF mutations were identified in 40.4% (19 of 47) of the tumors. The frequency of BRAF mutations in CM has varied between 0% and 50% in previous studies.15-21 Small sample size, the use of different detection methods, and a lack of assessment of V600K-type mutations could explain these differences. The distribution of V600E and V600K-type mutations in the present study corresponded well with that reported in cutaneous melanoma, in which V600K may be present in up to 20% of BRAF-mutated tumors.30 Therefore, future studies should also include investigations of the V600K-type mutation.
Patients having tumors with the BRAF mutation were younger at diagnosis compared with patients having wild-type CM, which has also been reported in cutaneous melanoma.13BRAF mutations occurred more frequently in the bulbar, corneolimbal, and caruncular conjunctiva and were more frequently classified as T1 tumors. An increase in the incidence of CM in recent years has been reported, particularly for cases developing in areas of the conjunctiva exposed to UV radiation.3,7,31 It is likely that this increase is also linked to an increase in BRAF mutations because these mutations appear more frequently at these sites. However, this theory needs to be investigated in larger patient series.
A high frequency of BRAF mutations in tumors without PAM and in tumors with a mixed pigmented or nonpigmented appearance was observed. The appearance of these tumors could be due to different tumor clones. The distribution of BRAF mutations within these clones remains to be investigated. A higher proportion of BRAF mutations in CM with origin in a nevus was observed, although statistical analysis of this finding does not rule out the possibility that this difference was due to chance. An association of BRAF mutations with nevi has been reported previously17,18 and corresponds well with the few BRAF-mutated tumors with clinically evident PAM in the present study.
When associating BRAF mutation status with outcome, an association with the development of distant metastatic disease was identified in univariate analysis. However, a definitive association was no longer identified in multivariable Cox proportional hazards regression models, which could be owing to the small proportion of tumors tested or the possibility that confounding factors rather than the BRAF mutations accounted for the association noted in univariate analysis. Additional studies evaluating the association of BRAF mutations with this outcome would be needed to increase the confidence that BRAF mutations are truly associated with this outcome. Bulbar location in CM is usually associated with a fair prognosis,6,7,10,11,27 which contrasts with our findings of more frequent distant metastases in BRAF-mutated bulbar CM. Previous studies15-21 investigating BRAF mutations in CM have not shown significant associations with prognosis. In a study by Lake et al,19 however, 50% of BRAF mutations were found in CM metastases. In addition, BRAF-mutated cutaneous melanoma has been associated with increased risk of melanoma-related mortality and a high propensity to metastasize.32,33 Considering the similarities between cutaneous melanoma and CM, it seems reasonable to assume that the same associations would apply to CM. However, the results should be interpreted with caution because of the few distant metastases and the limitations of our statistical analyses.
A poor prognosis of extrabulbar CM and CM with local invasion was confirmed. A higher risk of distant metastasis was identified in patients with CM in whom incisional biopsy was performed.
BRAF-mutated CMs were more frequently classified as T1 tumors and were associated with mixed or nonpigmented appearance, absence of PAM, and caruncular or bulbar localization, with bulbar location otherwise having a good prognosis. However, BRAF-mutated CMs more frequently developed distant metastases. Therefore, identification of BRAF mutations after excision and relevant adjuvant therapy is important for studies investigating BRAF and MEK inhibitor therapy and more intensive surveillance in patients with BRAF-mutated CMs. However, the limitations of this study preclude definitive conclusions, and larger studies evaluating BRAF mutations and prognosis are warranted. Because of the morbidity of this condition, there is a need for studies of adjuvant BRAF inhibitor therapy and more intensive surveillance in patients with BRAF-mutated CMs.
Submitted for Publication: April 13, 2015; final revision received July 15, 2015; accepted July 22, 2015.
Corresponding Author: Steffen Heegaard, MD, DMSc, Eye Pathology Institute, Department of Neuroscience and Pharmacology, University of Copenhagen, Frederik V’s vej 11, 1 Sal, 2100 Copenhagen Ø, Denmark (email@example.com).
Published Online: October 1, 2015. doi:10.1001/jamaophthalmol.2015.3200.
Author Contributions: Dr Larsen 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.
Study concept and design: Larsen, Toft, Coupland, Prause, Guldberg, Heegaard.
Acquisition, analysis, or interpretation of data: Larsen, Dahmcke, Dahl.
Drafting of the manuscript: Larsen, Dahmcke, Siersma.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Larsen, Siersma.
Obtained funding: Larsen.
Administrative, technical, or material support: Prause, Guldberg, Heegaard.
Study supervision: Toft, Coupland, Prause, Guldberg, Heegaard.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest, and none were reported.
Funding/Support: The study was supported by grants from Fight for Sight Denmark, Danish Eye Research Foundation, Synoptik-Fonden, Danish Cancer Research Foundation, Kleinsmed Schrøder og Hustrus Fond, Købmand Kjærs Fond, and Christian den X’s Fond.
Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
Previous Presentations: An abstract of this study was presented at the European Ocular Oncology Group Spring Meeting; February 7-9, 2013; Lapland, Finland; at the 16th Conference of the International Society of Ocular Oncology; September 30, 2013, to October 3, 2013; Cleveland, Ohio; at the World Ophthalmology Congress 2014; April 2-6, 2014; Tokyo, Japan; and at the European Ocular Oncology Group Spring Meeting; March 13-15, 2015; Moscow, Russia.
Create a personal account or sign in to: