January 2005
Volume 46, Issue 1
Free
Clinical and Epidemiologic Research  |   January 2005
Conjunctival Melanoma in The Netherlands: A Nationwide Study
Author Affiliations
  • Guy S. Missotten
    From the Department of Ophthalmology, Leiden University Medical Center, Leiden, The Netherlands.
  • Sander Keijser
    From the Department of Ophthalmology, Leiden University Medical Center, Leiden, The Netherlands.
  • Rob J. W. De Keizer
    From the Department of Ophthalmology, Leiden University Medical Center, Leiden, The Netherlands.
  • Didi De Wolff-Rouendaal
    From the Department of Ophthalmology, Leiden University Medical Center, Leiden, The Netherlands.
Investigative Ophthalmology & Visual Science January 2005, Vol.46, 75-82. doi:10.1167/iovs.04-0344
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Guy S. Missotten, Sander Keijser, Rob J. W. De Keizer, Didi De Wolff-Rouendaal; Conjunctival Melanoma in The Netherlands: A Nationwide Study. Invest. Ophthalmol. Vis. Sci. 2005;46(1):75-82. doi: 10.1167/iovs.04-0344.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

purpose. To evaluate risk factors for local recurrence, regional and distant metastases, and mortality associated with conjunctival melanoma.

methods This was a retrospective study of 194 patients with histologically confirmed conjunctival melanoma diagnosed between 1950 and 2002 in the Netherlands. Data were collected from all university centers and many nontertiary hospitals, using the National Pathology and the Leiden Oncologic Registration Systems. Based on the number of incidences, this study included 70% of the conjunctival melanomas in The Netherlands. Clinical and histopathological data for conjunctival tumors were reviewed and compared with data reported in the literature. Risk factors for local, regional, and distant metastases and survival were analyzed using the Kaplan-Meier and Cox regression analyses.

results. Of 194 patients with conjunctival melanoma, 112 had a local recurrence (median, 1; range, 1–9) during follow-up (median, 6.8 years; range, 0.1–51.5). Location was the most important risk factor for development of local recurrence, and significantly more occurred with nonepibulbar (log rank, P = 0.044) tumors. Significantly fewer local recurrences occurred with tumors initially treated with excision and adjuvant brachytherapy rather than with excision only (log rank, P = 0.008) or with excision and cryotherapy (log rank, P < 0.038). Forty-one (21%) patients had regional lymph node metastases, mostly to the parotid or preauricular lymph nodes (n = 26; 13%). Risk factors for regional metastases were tumor thickness (log rank, P < 0.001) and tumor diameter (log rank, P = 0.010). Forty-nine (25%) patients (mean, 4.37 years) had development of distant metastases, mainly in the lung, liver, skin, and brain. Tumor-related survival was 86.3% (95% confidence interval [CI], 81.0–91.6) at 5 years, 72% (95% CI, 79.7–64.4) at 10 years, and 67% (95% CI, 58.9–76.1) at 15 years. The main mortality risk factors were nonepibulbar location (log rank, P < 0.0001) and tumor thickness (log rank, P = 0.0004).

conclusions. Nonepibulbar tumors more often recur locally and are associated with a shorter survival independent of other risk factors. Tumor thickness is also an important predictor of regional and distant metastases, as well as survival. A prospective study is needed to compare the effect of excision with radiotherapy and excision with cryotherapy on the number of local recurrences, exenteration rate, and survival.

Conjunctival melanoma is a rare tumor 1 2 3 with an incidence of 0.02 to 0.08 per 100,000 in a white population, 4 5 6 accounting for some 1% to 3% of all ocular malignancies in adults. This malignancy normally occurs at approximately the age of 60 years and only rarely before the age of 40 (10%). Conjunctival melanomas may be associated with sun exposure, as is skin melanoma, 7 although they can occur at non–sun-exposed sites. 6 In contrast to uveal melanoma, the incidence of conjunctival melanoma is increasing. 6 8 9  
Local recurrence is reported to be 30%–50% at 5 years, 38%–51% at 10 years, and 65% after 15 years, depending on treatment. 10 11 12 13 14 15 16 17 18 19 20 21 Excision without additional treatment is associated with more local recurrences. 22 Main risk factors for patient mortality are thickness and size (basal diameter) as well as a nonepibulbar location for the primary tumor or one of its recurrences. A contributing factor for prognosis is the origin of the tumor—that is, in 55% primary acquired melanosis (PAM) and in 25% melanoma de novo, with a worse prognosis for melanoma de novo. 23 Conjunctival nevi rarely develop into melanoma. 24 The mortality rate is 12% to 20% at 5 years and up to 30% at 10 years. 10 11 12 13 14 15 16 17 18 19 20 21 Conjunctival melanomas are known to spread through the lymph system, although distant metastases are also found without regional lymph node metastasis. 25  
In this report, data from 194 patients with conjunctival melanoma collected from all university hospitals and many nontertiary centers in the Netherlands were analyzed to investigate differences in treatment and surgical approach, as well as the main risk factors for recurrence, metastasis, and survival. 
Patients and Methods
The Netherlands National Ocular and Orbital Tumor Committee, using the Netherlands Pathology Diagnosis System, helped identify 194 patients who had a diagnosis of malignant conjunctival melanoma between 1950 and 2002. Complete clinical and histopathological data were gathered from patients’ records and follow-up data were obtained from death certificates, general practitioners, local ophthalmologists, and the Oncologic Registration System of Leiden University Medical Center (LUMC). The mean follow-up was 9.2 years (SD, 8.1; median, 6.8). Only data for patients with histologically proven conjunctival melanoma were analyzed. Patients with a primary eyelid (cutaneous) melanoma were excluded, as were patients with PAM without an invasive component. The study was approved by the Institutional Review Board and adhered to the tenets of the Declaration of Helsinki. 
Assessment of Tumor Characteristics
Of 194 conjunctival melanomas, 178 (92%) primary tumor specimens were reviewed for histopathologic confirmation of the diagnosis at the LUMC, whereas 16 other specimens were assessed by an experienced ophthalmic pathologist elsewhere. 
Tumor origin was established on the basis of medical and histopathological records and photographs and was classified as nevus, nevus in combination with PAM, PAM, and de novo. The clinical variables included age at diagnosis, sex, unilocular (caruncula, fornix, tarsus, limbus, or epibulbar) or multilocular (diffuse) location of the primary tumor, date for primary treatment, date for diagnosis of local recurrence, and the presence of regional or hematogenous metastases. Local recurrences at the same or different location in the ocular region, but not PAM without melanoma, were considered local recurrence. 20 The proximity of the melanoma to the limbus was classified as limbal or nonlimbal. The basal tumor diameter (in millimeters) was taken from patient records, pathology reports, or photographs, if available (79%), as described by Tuomaala et al. 6 In 42 cases, information about tumor diameter could not be found. Melanoma thickness (in millimeters) was measured by one ophthalmic pathologist on slides of the tumor, according to Breslow. 26 In 41 cases, thickness could not be measured because tumors had been sliced in a tangential plane. Treatment of initial tumors and recurrences were categorized as excision only, excision and chemotherapy (topical mitomycin C), excision and cryotherapy, external irradiation only, brachytherapy after excision, orbital exenteration, and irradiation and exenteration. Cryotherapy and 90Sr brachytherapy were available as adjuvant therapy from 1970 onward and were frequently used after 1980 (Fig. 1)
Assessment of Outcome
Dates of diagnosis and local recurrence were those for histopathologic confirmation of the primary tumor or local recurrence. The interval between diagnosis by a physician and histopathologic confirmation was always <2 months. The date for metastasis was taken as the date when dissemination was confirmed by biopsy, imaging, or clinical examination. Regional metastasis was histopathologically confirmed in all 41 cases. Of 49 cases with distant metastases, the metastases were confirmed by histopathology in 24 (49%) patients and by imaging (mostly computed tomography) in 25 (51%) patients. Autopsy was performed in three patients. In 12 cases, the location of metastases was not specified. Survival time was defined as the time from histopathological confirmation of the primary tumor to death due to conjunctival melanoma (event), death due to other causes, or time of last presentation. 
Statistical Analysis
Risk factors for local or regional recurrence and for tumor-related survival were assessed by Kaplan-Meier estimates and with Cox proportional hazards. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated. P < 0.05 was considered significant. Statistical analysis was performed on computer (SPSS ver. 11.0; SPSS, Inc., Chicago, IL, and Stata/SE ver. 8.0; StataCorp., College Station, TX). Kaplan-Meier analysis for local recurrence was analyzed for time to the first local recurrence. The log rank test was used to compare variables in Kaplan-Meier analysis. In Cox regression analysis, the variables “origin” and “therapy” were considered nominal categorical variables. 
Results
Patient and Tumor Characteristics
Our study population consisted of 107 (55%) female and 87 (45%) male patients (Table 1) , including 99 (51%) left and 95 (49%) right eyes. Mean age at histopathological diagnosis was 57.4 years (median, 58). Regions involved were the limbus (n = 110, 57%), bulbar conjunctiva (limbus included; n = 133, 69%), caruncula (n = 5, 2.6%), tarsus (n = 11, 5.6%), and fornix (n = 2, 1%). More than one location was involved in 43 (22%) patients. Origin of the melanomas was PAM (n = 111, 57.2%), melanoma de novo (n = 50, 25.7%), nevus in combination with PAM (n = 9, 5%), and nevus only (n = 3, 2%). Origin was inconclusive in 21 (11%) patients. Tumor thickness was 2.07 mm (SD 1.9; n = 153) and the basal diameter was 7.21 mm (SD 4.7; n = 152). 
Primary Treatment
Treatment at the moment of histopathological diagnosis (Table 2)was excision alone (n = 127, 65.5%), excision with topical chemotherapy (mitomycin C; n = 4, 2%), excision and adjuvant cryotherapy (n = 17, 8.7%), irradiation (n = 11, 5.6%), and excision with brachytherapy (90Sr/90Y irradiation; 6 × 1000 cGy; n = 20, 10.3%). Exenteration was the primary therapy for advanced conjunctival melanoma in 14 (7.2%) patients and was used in combination with orbital irradiation in 1 (0.5%) patient. 
Today the standard treatment for primary conjunctival melanoma is excision with adjuvant therapy. However, during the period covered by this study, many different treatment strategies were used (Fig. 1) . Although in some cases only a few patients had some of the treatments, we tried to compare different treatments and their outcomes. 
Local Recurrence Analysis and Secondary Treatments
Within the total follow-up period, 112 patients had a local recurrence (62.6% of the 179 patients who did not undergo exenteration). The mean time between the first treatment for histopathologically proven melanoma and first local recurrence was 2.4 years (SD, 3.4; median, 1.3 years; maximum, 22.6 years). The mean number of local recurrences was 1.6 (range, 1–9). The percentage of patients without a local recurrence was 39.3% (95% CI, 31.5–47.1) after 5 years, 33.2% (95% CI, 25.4–41.0) after 10 years, and 27.8% (95% CI, 19.0–36.6) after 15 years (Fig. 2A)
Treatment of the first local recurrence was excision in 63 (56%) patients (4 in combination with topical chemotherapy), 22 of whom ultimately required orbital exenteration for more recurrences or advanced conjunctival melanoma. Twenty-three (20.5%) patients received excision with cryotherapy, with eight patients ultimately requiring exenteration. Excision with brachytherapy (92Ir irradiation, n = 3, 90Sr/90Y application, n = 8) was performed in 11 cases, with two patients undergoing exenteration. One patient was treated for recurrence with irradiation only. In total, 44 (39.3%) patients had exenteration for advanced disease. 
The probability of recurrence of primary tumors using excision with brachytherapy was lower than when other treatment modalities were used (Fig. 2B) . There were significantly fewer recurrences with excision and brachytherapy of primary tumors (n = 19) than with excision with adjuvant cryotherapy (n = 17; P = 0.038, log rank test not adjusted for multiple comparisons). Excision with cryotherapy, in comparison to excision only, did not result in fewer recurrences (P = 0.759, log rank test, not adjusted for multiple comparisons). 
In contrast to other tumor locations, epibulbar involvement showed significantly fewer local recurrences (P = 0.044, log rank; Fig. 2C ). Analysis showed that local recurrence was not associated with age at time of diagnosis (categories: <50, 50–68, and >68 years; P = 0.774, log rank test for trend), tumor origin (categories: PAM or non-PAM; P = 0.61, log rank), sex (P = 0.77, log rank), unilocular/multilocular location (P = 0.82, log rank), tumor thickness (<1, 1–2, and >2 mm; P = 0.063, log rank test for trend), or basal tumor diameter (0–6, 6–10, and >10 mm; P = 0.571). Cox regression analysis showed that epibulbar location and treatment modality of the primary tumor were the best predictors of local recurrence (Table 3)
Regional Metastases Analysis and Treatment
Regional metastases developed in 41 patients (n = 9 regional, n = 32 regional and distant). Of these 41 patients, 32 were first thought to have only regional dissemination. The mean time between primary tumor diagnosis and regional metastasis was 4.37 years (SD, 4.2; median, 3.7). Mean survival time after metastasis was 8.1 years (SD, 8.7; median, 2.7) with regional dissemination only (six patients still living, three died of nonrelated causes) and 1.72 years (SD, 1.5; median, 1.3) with both regional and distant metastases. Regional metastases were found in parotid or preauricular (n = 26), cervical (n = 13), submandibular (n = 9), and axillary (n = 2) lymph nodes. Both patients with positive axillary lymph nodes also had parotid positive lymph nodes. 
A univariate Cox regression analysis (Table 4)indicated that tumor thickness (HR 2.8, P < 0.001) and basal diameter (HR 1.6, P = 0.010) were the best predictors of regional metastases. In a multivariate Cox model, only tumor thickness reached statistical significance. 
Distant Recurrence Analysis
Distant metastases developed in 49 patients, of whom 17 showed no regional (skipping) metastases. In 37 patients, the distant metastasis was localized. Metastasis occurred in the lungs (n = 18 cases), liver (n = 15), skin (n = 13), brain (n = 11), spinal cord (n = 5), ileum (n = 3), mesentery (n = 3), skeletal bones (n = 3), thyroid gland (n = 2), and jaw bone (n = 2). 
Mean survival was 1.72 years (SD, 1.5; median, 1.3) for both regional and distant and 0.42 years (SD, 0.34; median, 0.44) for hematogenous metastases. Analysis (Table 5)showed that nonepibulbar lesions were significantly more likely to give rise to a distant metastasis (HR 4.0, P < 0.001). A univariate Cox analysis showed that increasing tumor diameter was associated with an increasing risk of distant recurrences (HR 2.1, P < 0.001). 
Survival Analysis
During the total follow-up period, 47 (24.3%) patients died of metastasis of conjunctival melanoma, 44 (22.7%) of non–melanoma-related causes, and 103 (53%) were still living. Fifty-eight (29.9%) patients had metastasis. The melanoma-specific survival rate was 86.3% (95% CI: 81.0–91.6) at 5 years, 71.2% (95% CI: 79.7–64.4) at 10 years, and 67.2% (95% CI: 58.9–76.1) at 15 years (Fig. 3A)
Survival analysis did not show a significant difference in survival between therapies (overall log rank, P = 0.363). Only patients who underwent a primary exenteration had a significantly lower survival than those with other treatment modalities (P = 0.036, log rank test not adjusted for multiple comparisons). This may be explained by the fact that this group contained significantly larger tumors and at an unfavorable location. 
Neither tumor origin (P = 0.42) nor tumor basal diameter (P = 0.066, log rank test for trend) showed a significant difference in survival rate. However, survival was significantly better for epibulbar tumors (P < 0.0001), unilocular location (P = 0.02), and thin tumors (P = 0.0004, overall log rank for trend). Age at diagnosis (<50, 50–68, and >68 years) correlated significantly with survival (P = 0.023, log rank test for trend), but older patients had significantly larger (P < 0.023) and thicker (P < 0.021) tumors. Cox regression analysis showed the same risk factors as for the development of distant metastases (Table 6) , most important, a nonepibulbar location (HR 3.6, P < 0.001). 
Discussion
Our study suggests that the incidence of conjunctival melanoma in The Netherlands is similar to the previously reported incidence of 0.05/100,000 inhabitants. 4 5 6 8 Since 1970, conjunctival melanoma has been diagnosed in 172 patients. With a mean population of 14,575,000 in The Netherlands (Central Bureau of Statistics, Voorburg/Heerlen, The Netherlands), over the past 30 years, an estimated 240 patients would have received a diagnosis of this type of melanoma. Thus, we probably included >70% of possible patients with conjunctival melanoma in The Netherlands after 1970. 
Treatment of the local primary tumor has changed in the last decades from irradiation or excision as sole therapy to excision in combination with cryotherapy or brachytherapy. This has enabled us to compare different treatments. For local tumor control, excision in combination with brachytherapy appears to result in fewer local recurrences than when using other treatments. Other studies also found a trend toward an increased local recurrence rate after excision without additional treatment compared with excision and ruthenium plaque 27 28 29 30 or additional external radiotherapy. 18 Reduction in local recurrences may also reduce the need for secondary exenteration. 
The other main risk factor for local recurrences was nonepibulbar location. None of the other variables, such as age at diagnosis, unilocular or multilocular location, origin, thickness, or diameter, were significantly correlated with the local recurrence rate. We confirm the findings of Tuomaala et al. 6 that location is the most important risk factor for local recurrence. The 5-year local recurrence rate was 60.7% (95% CI: 53.0–68.5). Others reports have shown lower 5-year rates of between 26% and 52%. 6 11 15 19 20 This difference may be explained by the percentage of tumors treated with adjuvant therapy. However, 10 years after diagnosis, local recurrence seemed to reach comparable levels. 19 20 Therefore, the long-term effect of adjuvant therapy on local recurrences still has to be evaluated. One of our patients had nine local recurrences despite receiving adjuvant brachytherapy and cryotherapy. 
Regional metastases occurred in 41 (21%) patients, mostly in the preauricular or parotid (n = 26) region and cervical (n = 13) regions. Four patients with only regional metastasis survived for >15 years after local treatment, as also reported for parotid gland melanomas. 31 32 This long survival, in the absence of distant metastases, supports an active search for positive lymph nodes in the preauricular, cervical, and submandibular regions. This result may also advocate consideration of sentinel lymph node biopsy 33 in trying to treat regional metastases at a subclinical stage. Lymphoscintigraphy and sentinel lymph node biopsy is a safe procedure, although its sensitivity is still not known, 33 34 35 and care should be taken to avoid facial paresis. The indication for sentinel lymph node biopsy is still uncertain, but the Cox analysis for regional metastases (Table 4)suggests that tumors thicker than 2 mm with a diameter greater than 10 mm warrant this approach. In our study, primary epibulbar tumors tended to give rise to fewer regional metastases (Table 4) , in agreement with Tuomaala et al. 36 who reported significantly fewer regional lymph nodes from limbal tumors. 
In forty-seven (24%) patients distant metastases developed, mainly in lung, liver, skin, and brain, 40 of these patients had distant metastases in these four locations. Esmaeli et al. 25 also reported liver and lung as main sites of metastasis. Seventeen patients had distant metastasis, without prior or concurrent regional nodal involvement, indicating that not all patients with melanoma will benefit from sentinel lymph node biopsy. In contrast to the study of Esmaeli et al., we did not find that the primary tumors that skipped the lymph nodes had a higher rate of local recurrence. 
The survival rate was 86.3% at 5 years, 71.2% at 10 years, and 67.2% at 15 years (Fig. 3A) , similar to other studies (5-year survival 87.6%–74%). 4 21 25 The total exenteration rate was 30% over a mean follow-up period of 9.2 years, comparable to the 37% reported by Paridaens et al., 37 and the estimated 15-year 32% exenteration rate of reported by Shields and Shields. 38 Main risk factors for distant metastases, as well as for mortality, were nonepibulbar location and thick (>2 mm), multilocular tumors. 
Although we found a significant difference in local recurrence analysis, there was no significant difference in survival with different treatment strategies. Only primary exenteration for advanced disease was associated with a significantly lower survival, which can be explained by the fact that these tumors were more extensive and had a more frequently unfavorable location. We think that more knowledge of this rare tumor may help us to detect suspect lesions earlier and thus ensure timely referral of patients to ophthalmic oncology centers. 
Conclusions
Predictors of poor survival were a nonepibulbar location, tumors thicker than 2 mm, and multifocal location. Survival may be improved by early diagnosis of regional metastases. The long survival rate of some patients with regional metastases also suggests that a sentinel lymph node biopsy may increase survival in high-risk patients (e.g., especially in patients with multiple local recurrences, a Breslow tumor thickness of >2 mm and a tumor diameter greater than 10 mm). Adjuvant brachytherapy may reduce the number of local recurrences. A prospective study is needed to compare the effect of excision in combination with radiotherapy and excision with cryotherapy on the number of local recurrences, exenteration rate, and survival. 
 
Figure 1.
 
Treatment modality as a function of time. The study was made on data contributed from all university centers and many general ophthalmologists.
Figure 1.
 
Treatment modality as a function of time. The study was made on data contributed from all university centers and many general ophthalmologists.
Table 1.
 
Patient and Tumor Characteristics of 194 Patients with Conjunctival Melanoma
Table 1.
 
Patient and Tumor Characteristics of 194 Patients with Conjunctival Melanoma
Variables Cases (n) %
Sex
 Male 87 45
 Female 107 55
Location
 Fornix 2 1
 Caruncular 5 2.5
 Tarsal 11 5.5
 Limbal 110 57
 Epibulbar 133 69
 Diffuse 43 22
Origin of melanoma
 Nevus 3 2
 Nevus and PAM 9 4
 PAM 111 57
 De novo 50 26
 Inconclusive 21 11
Cases (n) Mean SD
Measurements
 Basal diameter (mm) 153 7.21 4.7
 Thickness (mm) 152 2.07 1.9
Table 2.
 
Treatment of the Primary Tumor at Moment of Its Histological Diagnosis, and the Number of Patients with a Local Recurrence
Table 2.
 
Treatment of the Primary Tumor at Moment of Its Histological Diagnosis, and the Number of Patients with a Local Recurrence
Treatment Modality Cases Patients with a Local Recurrence
(n) n %
Excision 127 85 67
Excision and chemotherapy (MMC) 4 4 100
Excision and cryotherapy 17 11 65
External irradiation 11 7 64
Excision and brachytherapy 19 5 26
Excision and brachytherapy and cryotherapy 1 0 0
Exenteration and irradiation 1 0 0
Exenteration 14 0 0
Figure 2.
 
Kaplan Meier recurrence analysis in 179 patients (patients with primary exenteration are not included in the recurrence analysis). (A) Overall recurrence rate after the primary treated tumor (time analyzed from primary diagnosis to first local recurrence). (B) Recurrence rate as a function of treatment modality of the primary lesion (overall log rank P = 0.043). Excision in combination with irradiation had significantly fewer recurrences (log rank P = 0.038). There was no significant difference between excision with cryotherapy when compared with excision only (log rank P = 0.759). (C) Recurrence rate in patients with only an epibulbar component (132 patients) or not (47 patients; log rank P < 0.044).
Figure 2.
 
Kaplan Meier recurrence analysis in 179 patients (patients with primary exenteration are not included in the recurrence analysis). (A) Overall recurrence rate after the primary treated tumor (time analyzed from primary diagnosis to first local recurrence). (B) Recurrence rate as a function of treatment modality of the primary lesion (overall log rank P = 0.043). Excision in combination with irradiation had significantly fewer recurrences (log rank P = 0.038). There was no significant difference between excision with cryotherapy when compared with excision only (log rank P = 0.759). (C) Recurrence rate in patients with only an epibulbar component (132 patients) or not (47 patients; log rank P < 0.044).
Table 3.
 
Cox Regression Analysis of the Appearance of First Local Recurrence
Table 3.
 
Cox Regression Analysis of the Appearance of First Local Recurrence
Variable Coefficient (SE) Wald2 P Hazard Ratio (95% CI)
Univariate analysis
 Age at diagnosis* 0.005 (0.005) 0.897 0.344 1.005 (0.994–1.016)
 Nonepibulbar location 0.412 (0.206) 3.996 0.046 1.510 (1.008–2.263)
 Uni/multilocular 0.087 (0.170) 0.263 0.608 1.091 (0.782–1.522)
 Origin, † (referenced to melanoma), ¶ 1.768 0.622
  Nevus −0.515 (0.732) 0.495 0.482 0.598 (0.142–2.509)
  Nevus + PAM −0.624 (0.532) 1.374 0.241 0.536 (0.189–1.520)
  PAM −0.120 (0.219) 0.303 0.582 0.887 (0.578–1.361)
 Thickness, ‡ 0.240 (0.130) 3.433 0.064 1.271 (0.986–1.639)
 Basal diameter, § 0.065 (0.115) 0.320 0.571 1.067 (0.852–1.338)
 Therapy, ∥ (referenced to excision + brachytherapy), ¶ 8.065 0.045
  Excision 1.251 (0.460) 7.404 0.007 3.494 (1.419–8.606)
  Excision + cryotherapy 1.129 (0.540) 4.373 0.037 3.093 (1.073–8.911)
  Irradiation 0.863 (0.588) 2.151 0.142 2.370 (0.748–7.506)
Multivariate analysis
 Model (−2 log likelihood 1006.905)
 Non-epibulbar location 0.533 (0.211) 6.376 0.012 1.704 (1.127–2.577)
 Therapy, ∥ (referenced to excision + brachytherapy), ¶ 9.667 0.022
  Excision 1.331 (0.461) 8.315 0.004 3.783 (1.531–9.347)
  Excision + cryotherapy 1.236 (0.542) 5.196 0.023 3.442 (1.189–9.963)
  Irradiation 0.768 (0.590) 1.699 0.192 2.156 (0.679–6.848)
Table 4.
 
Cox Regression Analysis for the Appearance of Regional Metastases in 194 Patients
Table 4.
 
Cox Regression Analysis for the Appearance of Regional Metastases in 194 Patients
Variable Coefficient (SE) Wald2 P Hazard Ratio (95% CI)
Univariate analysis
 Age at diagnosis* 0.011 (0.010) 1.234 0.267 1.011 (0.992–1.031)
 Nonepibulbar location 0.596 (0.321) 3.456 0.063 1.815 (0.968–3.404)
 Uni/multilocular 0.231 (0.228) 1.027 0.311 1.259 (0.806–1.967)
 Origin† (referenced to melanoma)¶ 1.260 0.739
  Nevus −12.241 (414.426) 0.001 0.976 0.000 (NA)
  Nevus + PAM 0.312 (0.652) 0.229 0.632 1.366 (0.381–4.903)
  PAM −0.281 (0.370) 0.575 0.448 0.755 (0.366–1.560)
 Thickness‡ 1.019 (0.238) 18.363 <0.001 2.770 (1.738–4.415)
 Basal diameter§ 0.489 (0.189) 6.660 0.010 1.631 (1.125–2.364)
 Therapy∥ (referenced to excision + brachytherapy)¶ 1.483 0.686
  Excision 0.566 (0.732) 0.597 0.440 1.761 (0.419–7.396)
  Excision + cryotherapy −0.092 (1.001) 0.008 0.927 0.912 (0.128–6.485)
  Irradiation 0.741 (0.915) 0.656 0.418 2.098 (0.349–12.602)
Multivariate analysis
 Model 1 (−2 log likelihood 264.401)
 Thickness‡ 0.874 (0.254) 11.851 0.001 2.397 (1.457–3.942)
 Basal diameter§ 0.192 (0.213) 0.811 0.368 1.211 (0.798–1.839)
Table 5.
 
Cox Regression Analysis for the Appearance of Hematogenous Metastases in 194 Patients with Conjunctival Melanoma
Table 5.
 
Cox Regression Analysis for the Appearance of Hematogenous Metastases in 194 Patients with Conjunctival Melanoma
Variable Coefficient (SE) Wald2 P Hazard Ratio (95% CI)
Univariate analysis
 Age at diagnosis* 0.425 (0.180) 5.556 0.018 1.530 (1.074–2.179)
 Nonepibulbar location 1.396 (0.289) 23.269 <0.001 4.039 (2.290–7.122)
 Uni/multilocular 0.034 (0.057) 0.343 0.558 1.034 (0.924–1.157)
 Origin† (referenced to melanoma)¶ 1.234 0.745
  Nevus −12.305 (358.641) 0.001 0.973 0.000 (0.000–8.544)
  Nevus + PAM −0.007 (0.637) 0.000 0.991 0.993 (0.285–3.461)
  PAM −0.347 (0.331) 1.100 0.294 0.707 (0.369–1.352)
 Thickness‡ 0.743 (0.196) 14.306 <0.001 2.102 (1.430–3.089)
 Basal diameter§ 0.322 (0.172) 3.501 0.061 1.379 (0.985–1.932)
 Therapy∥ (referenced to excision + brachytherapy)¶ 3.121 0.373
  Excision 0.711 (0.727) 0.955 0.328 2.036 (0.489–8.471)
  Excision + cryotherapy −0.818 (1.226) 0.466 0.504 0.441 (0.040–4.876)
  Irradiation 0.582 (0.918) 0.401 0.526 1.789 (0.296–10.813)
Multivariate analysis
 Model 1 (−2 log likelihood 333.378)
 Nonepibulbar location 1.269 (0.331) 14.694 <0.001 3.557 (1.859–6.805)
 Thickness‡ 0.533 (0.197) 7.345 0.007 1.704 (1.159–2.505)
Figure 3.
 
Kaplan Meier survival analysis of 194 patients. (A) Overall survival (in years) after primary treated tumor (histopathologically diagnosed conjunctival melanoma). (B) Survival of patients with epibulbar (61 patients) and nonepibulbar (133 patients) conjunctival melanoma. Log rank P < 0.0001. (C) Survival analysis for different tumor thicknesses of the primary lesion (overall log rank test for trend P = 0.0004).
Figure 3.
 
Kaplan Meier survival analysis of 194 patients. (A) Overall survival (in years) after primary treated tumor (histopathologically diagnosed conjunctival melanoma). (B) Survival of patients with epibulbar (61 patients) and nonepibulbar (133 patients) conjunctival melanoma. Log rank P < 0.0001. (C) Survival analysis for different tumor thicknesses of the primary lesion (overall log rank test for trend P = 0.0004).
Table 6.
 
Cox Regression Analysis for Survival in 194 Patients with Conjunctival Melanoma
Table 6.
 
Cox Regression Analysis for Survival in 194 Patients with Conjunctival Melanoma
Variable Coefficient (SE) Wald2 P Hazard Ratio (95% CI)
Univariate analysis
 Age at diagnosis* 0.018 (0.010) 3.425 0.064 1.018 (0.999–1.037)
 Nonepibulbar location 1.286 (0.294) 19.113 <0.001 3.617 (2.033–6.438)
 Uni/multilocular 0.439 (0.196) 5.210 0.025 1.551 (1.056–2.278)
 Origin† (referenced to melanoma)¶ 1.417 0.702
  Nevus −12.293 (354.707) 0.001 0.973 0.000 (0.000–8.948)
  Nevus + PAM 0.192 (0.644) 0.089 0.765 1.212 (0.343–4.279)
  PAM −0.333 (0.341) 0.954 0.329 0.716 (0.367–1.399)
 Thickness‡ 0.678 (0.200) 11.465 0.001 1.969 (1.330–2.916)
 Basal diameter§ 0.320 (0.176) 3.299 0.069 1.378 (0.975–1.946)
 Therapy∥ (referenced to excision + brachytherapy)¶ 2.788 0.426
  Excision 0.649 (0.728) 0.795 0.372 1.914 (0.460–7.972)
  Excision + cryotherapy −0.819 (1.226) 0.446 0.504 0.441 (0.040–4.873)
  Irradiation 0.544 (0.918) 0.351 0.553 1.723 (0.285–10.417)
Multivariate analysis
Model 1 (−2 log likelihood 319.991)
Nonepibulbar location 1.148 (0.337) 11.595 0.001 3.152 (1.628–6.103)
Thickness‡ 0.491 (0.202) 5.914 0.015 1.634 (1.100–2.426)
The authors thank participating universities and ophthalmologists, especially Leo E. Blank (Radiotherapy, Academic Medical Center, Amsterdam, The Netherlands) and Amy S. Keegan, members of the National Committee on Ocular and Orbital Tumors; Nico Bleichrodt of the Cancer Registration Leiden for help with the survival data; Paul H. C. Eilers (Department of Biomedical Statistics, Leiden University Medical Center) for statistical advice; and, especially, the reviewers and editorial board member for many suggestions and statistical advice that improved the manuscript substantially. 
FolbergR, McLeanIW, ZimmermanLE. Malignant melanoma of the conjunctiva. Hum Pathol. 1985;16:136–143. [CrossRef] [PubMed]
ShieldsJA, ShieldsCL. Tumors of the conjunctiva and cornea.SmolinG TholftRA eds. The Cornea: Scientific Foundations and Clinical Practice. 1993; 3rd ed. 579–595.Little, Brown & Co. Boston.
ShieldsJA, ShieldsCL. Tumors of the conjunctiva.StephensonCM eds. Ophthalmic Plastic, Reconstructive and Orbital Surgery. 1997;253–271.Butterworth-Heinemann Stoneham, MA.
LommatzschP, LommatzschR, KirschI, FuhrmannP. Therapeutic outcome of patients suffering from malignant melanomas of the conjunctiva. Br J Ophthalmol. 1990;74:615–619. [CrossRef] [PubMed]
SeregardS. Conjunctival melanoma: major review. Surv Ophthalmol. 1998;42:321–350. [CrossRef] [PubMed]
TuomaalaS, EskelinS, TarkkanenA, KivelaT. Population-based assessment of clinical characteristics predicting outcome of conjunctival melanoma in whites. Invest Ophthalmol Vis Sci. 2002;43:3399–3408. [PubMed]
de VriesE, SchoutenLJ, VisserO, EggermontAM, CoeberghJW. Rising trends in the incidence of and mortality from cutaneous melanoma in The Netherlands: a Northwest to Southeast gradient?. Eur J Cancer. 2003;39:1439–1446. [CrossRef] [PubMed]
YuG-P, HuD-N, McCormickS, FingerP. Conjunctival melanoma: is it increasing in the United States?. Am J Ophthalmol. 2003;135:800–806. [CrossRef] [PubMed]
BergmanL, SeregardS, NilssonB, et al. Incidence of uveal melanoma in Sweden from 1960–1998. Invest Ophthalmol Vis Sci. 2002;43:2579–2583. [PubMed]
de Wolff-RouendaalD, OosterhuisJA. Conjunctival melanoma in the Netherlands: a follow-up study. Doc Ophthalmol. 1983;56:49–54. [CrossRef] [PubMed]
FuchsU, KiveläT, LiestoK, TarkkanenA. Prognosis of conjunctival melanomas in relation to histopathological features. Br J Cancer. 1989;59:261–267. [CrossRef] [PubMed]
de Wolff-RouendaalD. Conjunctival melanoma in the Netherlands: a clinicopathological and follow-up study. Thesis. 1990;University of Leiden Leiden, The Netherlands.
SeregardS, KockE. Conjunctival malignant melanoma in Sweden 1969–1991. Acta Ophthalmol. 1992;70:289–296.
De PotterP, ShieldsC, ShieldsJ, MendukeH. Clinical predictive factors for development of recurrence and metastasis in conjunctival melanoma: a review of 68 cases. Br J Ophthalmol. 1993;77:624–630. [CrossRef] [PubMed]
ParidaensAD, MinassianDC, McCartneyAC, HungerfordJL. Prognostic factors in primary malignant melanoma of the conjunctiva: a clinicopathological study of 256 cases. Br J Ophthalmol. 1994;78:252–259. [CrossRef] [PubMed]
NorregaardJC, GernerN, JensenOA, PrauseJU. Malignant melanoma of the conjunctiva: occurrence and survival following surgery and radiotherapy in a Danish population. Graefes Arch Clin Exp Ophthalmol. 1996;234:569–572. [CrossRef] [PubMed]
Bobic-RadovanovicA, LatkovicZ, MarinkovicJ, RadovanovicZ. Predictors of survival in malignant melanoma of the conjunctiva: a clinico-pathological and follow-up study. Eur J Ophthalmol. 1998;8:4–7. [PubMed]
DesjardinsL, PoncetP, LevyC, et al. Facteurs pronostiques du melanoma malin de la conjonctive: etude anatomo-clinique sur 56 patients. J Fr Ophtalmol. 1999;22:315–321. [PubMed]
ShieldsCL. Conjunctival melanoma: risk factors for recurrence, exenteration, metastasis, and death in 150 consecutive patients. Trans Am Ophthalmol Soc. 2000;98:471–492. [PubMed]
ShieldsC, ShieldsJ, GündüzK, et al. Conjunctival melanoma: risk factors for recurrence, exenteration, metastasis, and death in 150 consecutive patients. Arch Ophthalmol. 2000;118:497–1507.
WerschnikC, LommatzschPK. Long-term follow-up of patients with conjunctival melanoma. Am J Clin Oncol. 2002;25:248–255. [CrossRef] [PubMed]
LommatzschPK, WerschnikC. Das maligne Melanom der Bindehaut. Klinische Übersicht mit Empfehlungen zur Diagnose, Therapie und Nachsorge. Klin Monatsbl Augenheilkd. 2002;219:710–721. [CrossRef] [PubMed]
de Wolff-RouendaalD. Melanozytäre tumoren der bindehaut.LommatzschP eds. Ophthalmologische onkologie. 1990;81–95.Enke Stuttgart, Germany.
ShieldsCL, FasiuddenA, MashayekhiA, ShieldsJA. Conjunctival nevi: clinical features and natural course in 410 consecutive patients. Arch Ophthalmol. 2004;122:167–175. [CrossRef] [PubMed]
EsmaeliB, WangX, YoussefA, GershenwaldJE. Patterns of regional and distant metastasis in patients with conjunctival melanoma: experience at a cancer center over four decades. Ophthalmology. 2001;108:2101–2105. [CrossRef] [PubMed]
BreslowA. Thickness, cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg. 1970;172:902–908. [CrossRef] [PubMed]
LommatzschP, VollmarR. Ergebnisse der Beta-Therapie bei epibulbären tumoren. Klin Monatsbl Augenheilkd. 1964;144:856–871. [PubMed]
LommatzschP. Beta-ray treatment of malignant epithelial tumors of the conjunctiva. Am J Ophthalmol. 1976;81:198–206. [CrossRef] [PubMed]
LommatzschP. Beta-ray treatment of conjunctival melanomas. Trans Ophthalmol Soc UK. 1977;97:378–380. [PubMed]
AnastassiouG, HeiligenhausA, BechrakisN, et al. Prognostic value of clinical and histopathological parameters in conjunctival melanoma retrospective study. Br J Ophthalmol. 2002;86:163–167. [CrossRef] [PubMed]
PraysonR, SebekB. Parotid gland malignant melanomas. Arch Pathol Lab Med. 2000;124:1780–1784. [PubMed]
WilsonMW, FlemingJC, FlemingRM, HaikBG. Sentinel node biopsy for orbital and ocular adnexal tumors. Ophthalmic Plast Reconstr Surg. 2001;17:338–345. [CrossRef]
EsmaeliB, EicherS, PoppJ, et al. Sentinel lymph node biopsy for conjunctival melanoma. Ophthal Plast Reconstr Surg. 2001;17:436–442. [CrossRef] [PubMed]
EsmaeliB. Sentinel node biopsy in conjunctival lesions.HonavarSG eds. XIIIth International Congress Ocular Oncology. 2004;Hyderabad, India.Abstract 35
EsmaeliB, ReiflerD, PrietoVG, et al. Conjunctival melanoma with a positive sentinel lymph node. Arch Ophthalmol. 2003;121:1779–1783. [CrossRef] [PubMed]
TuomaalaS, KivelaT. Metastatic pattern and survival in disseminated conjunctival melanoma: implication for sentinel lymph node biopsy. Ophthalmology. 2004;111:816–821. [CrossRef] [PubMed]
ParidaensAD, MinassianDC, McCartneyAC, HungerfordJL. Orbital exenteration in 95 cases of primary conjunctival malignant melanoma. Br J Ophthalmol. 1994;78:520–528. [CrossRef] [PubMed]
ShieldsC, ShieldsJ. Tumors of the conjunctiva and cornea. Surv Ophthalmol. 2004;49:3–24. [CrossRef] [PubMed]
Figure 1.
 
Treatment modality as a function of time. The study was made on data contributed from all university centers and many general ophthalmologists.
Figure 1.
 
Treatment modality as a function of time. The study was made on data contributed from all university centers and many general ophthalmologists.
Figure 2.
 
Kaplan Meier recurrence analysis in 179 patients (patients with primary exenteration are not included in the recurrence analysis). (A) Overall recurrence rate after the primary treated tumor (time analyzed from primary diagnosis to first local recurrence). (B) Recurrence rate as a function of treatment modality of the primary lesion (overall log rank P = 0.043). Excision in combination with irradiation had significantly fewer recurrences (log rank P = 0.038). There was no significant difference between excision with cryotherapy when compared with excision only (log rank P = 0.759). (C) Recurrence rate in patients with only an epibulbar component (132 patients) or not (47 patients; log rank P < 0.044).
Figure 2.
 
Kaplan Meier recurrence analysis in 179 patients (patients with primary exenteration are not included in the recurrence analysis). (A) Overall recurrence rate after the primary treated tumor (time analyzed from primary diagnosis to first local recurrence). (B) Recurrence rate as a function of treatment modality of the primary lesion (overall log rank P = 0.043). Excision in combination with irradiation had significantly fewer recurrences (log rank P = 0.038). There was no significant difference between excision with cryotherapy when compared with excision only (log rank P = 0.759). (C) Recurrence rate in patients with only an epibulbar component (132 patients) or not (47 patients; log rank P < 0.044).
Figure 3.
 
Kaplan Meier survival analysis of 194 patients. (A) Overall survival (in years) after primary treated tumor (histopathologically diagnosed conjunctival melanoma). (B) Survival of patients with epibulbar (61 patients) and nonepibulbar (133 patients) conjunctival melanoma. Log rank P < 0.0001. (C) Survival analysis for different tumor thicknesses of the primary lesion (overall log rank test for trend P = 0.0004).
Figure 3.
 
Kaplan Meier survival analysis of 194 patients. (A) Overall survival (in years) after primary treated tumor (histopathologically diagnosed conjunctival melanoma). (B) Survival of patients with epibulbar (61 patients) and nonepibulbar (133 patients) conjunctival melanoma. Log rank P < 0.0001. (C) Survival analysis for different tumor thicknesses of the primary lesion (overall log rank test for trend P = 0.0004).
Table 1.
 
Patient and Tumor Characteristics of 194 Patients with Conjunctival Melanoma
Table 1.
 
Patient and Tumor Characteristics of 194 Patients with Conjunctival Melanoma
Variables Cases (n) %
Sex
 Male 87 45
 Female 107 55
Location
 Fornix 2 1
 Caruncular 5 2.5
 Tarsal 11 5.5
 Limbal 110 57
 Epibulbar 133 69
 Diffuse 43 22
Origin of melanoma
 Nevus 3 2
 Nevus and PAM 9 4
 PAM 111 57
 De novo 50 26
 Inconclusive 21 11
Cases (n) Mean SD
Measurements
 Basal diameter (mm) 153 7.21 4.7
 Thickness (mm) 152 2.07 1.9
Table 2.
 
Treatment of the Primary Tumor at Moment of Its Histological Diagnosis, and the Number of Patients with a Local Recurrence
Table 2.
 
Treatment of the Primary Tumor at Moment of Its Histological Diagnosis, and the Number of Patients with a Local Recurrence
Treatment Modality Cases Patients with a Local Recurrence
(n) n %
Excision 127 85 67
Excision and chemotherapy (MMC) 4 4 100
Excision and cryotherapy 17 11 65
External irradiation 11 7 64
Excision and brachytherapy 19 5 26
Excision and brachytherapy and cryotherapy 1 0 0
Exenteration and irradiation 1 0 0
Exenteration 14 0 0
Table 3.
 
Cox Regression Analysis of the Appearance of First Local Recurrence
Table 3.
 
Cox Regression Analysis of the Appearance of First Local Recurrence
Variable Coefficient (SE) Wald2 P Hazard Ratio (95% CI)
Univariate analysis
 Age at diagnosis* 0.005 (0.005) 0.897 0.344 1.005 (0.994–1.016)
 Nonepibulbar location 0.412 (0.206) 3.996 0.046 1.510 (1.008–2.263)
 Uni/multilocular 0.087 (0.170) 0.263 0.608 1.091 (0.782–1.522)
 Origin, † (referenced to melanoma), ¶ 1.768 0.622
  Nevus −0.515 (0.732) 0.495 0.482 0.598 (0.142–2.509)
  Nevus + PAM −0.624 (0.532) 1.374 0.241 0.536 (0.189–1.520)
  PAM −0.120 (0.219) 0.303 0.582 0.887 (0.578–1.361)
 Thickness, ‡ 0.240 (0.130) 3.433 0.064 1.271 (0.986–1.639)
 Basal diameter, § 0.065 (0.115) 0.320 0.571 1.067 (0.852–1.338)
 Therapy, ∥ (referenced to excision + brachytherapy), ¶ 8.065 0.045
  Excision 1.251 (0.460) 7.404 0.007 3.494 (1.419–8.606)
  Excision + cryotherapy 1.129 (0.540) 4.373 0.037 3.093 (1.073–8.911)
  Irradiation 0.863 (0.588) 2.151 0.142 2.370 (0.748–7.506)
Multivariate analysis
 Model (−2 log likelihood 1006.905)
 Non-epibulbar location 0.533 (0.211) 6.376 0.012 1.704 (1.127–2.577)
 Therapy, ∥ (referenced to excision + brachytherapy), ¶ 9.667 0.022
  Excision 1.331 (0.461) 8.315 0.004 3.783 (1.531–9.347)
  Excision + cryotherapy 1.236 (0.542) 5.196 0.023 3.442 (1.189–9.963)
  Irradiation 0.768 (0.590) 1.699 0.192 2.156 (0.679–6.848)
Table 4.
 
Cox Regression Analysis for the Appearance of Regional Metastases in 194 Patients
Table 4.
 
Cox Regression Analysis for the Appearance of Regional Metastases in 194 Patients
Variable Coefficient (SE) Wald2 P Hazard Ratio (95% CI)
Univariate analysis
 Age at diagnosis* 0.011 (0.010) 1.234 0.267 1.011 (0.992–1.031)
 Nonepibulbar location 0.596 (0.321) 3.456 0.063 1.815 (0.968–3.404)
 Uni/multilocular 0.231 (0.228) 1.027 0.311 1.259 (0.806–1.967)
 Origin† (referenced to melanoma)¶ 1.260 0.739
  Nevus −12.241 (414.426) 0.001 0.976 0.000 (NA)
  Nevus + PAM 0.312 (0.652) 0.229 0.632 1.366 (0.381–4.903)
  PAM −0.281 (0.370) 0.575 0.448 0.755 (0.366–1.560)
 Thickness‡ 1.019 (0.238) 18.363 <0.001 2.770 (1.738–4.415)
 Basal diameter§ 0.489 (0.189) 6.660 0.010 1.631 (1.125–2.364)
 Therapy∥ (referenced to excision + brachytherapy)¶ 1.483 0.686
  Excision 0.566 (0.732) 0.597 0.440 1.761 (0.419–7.396)
  Excision + cryotherapy −0.092 (1.001) 0.008 0.927 0.912 (0.128–6.485)
  Irradiation 0.741 (0.915) 0.656 0.418 2.098 (0.349–12.602)
Multivariate analysis
 Model 1 (−2 log likelihood 264.401)
 Thickness‡ 0.874 (0.254) 11.851 0.001 2.397 (1.457–3.942)
 Basal diameter§ 0.192 (0.213) 0.811 0.368 1.211 (0.798–1.839)
Table 5.
 
Cox Regression Analysis for the Appearance of Hematogenous Metastases in 194 Patients with Conjunctival Melanoma
Table 5.
 
Cox Regression Analysis for the Appearance of Hematogenous Metastases in 194 Patients with Conjunctival Melanoma
Variable Coefficient (SE) Wald2 P Hazard Ratio (95% CI)
Univariate analysis
 Age at diagnosis* 0.425 (0.180) 5.556 0.018 1.530 (1.074–2.179)
 Nonepibulbar location 1.396 (0.289) 23.269 <0.001 4.039 (2.290–7.122)
 Uni/multilocular 0.034 (0.057) 0.343 0.558 1.034 (0.924–1.157)
 Origin† (referenced to melanoma)¶ 1.234 0.745
  Nevus −12.305 (358.641) 0.001 0.973 0.000 (0.000–8.544)
  Nevus + PAM −0.007 (0.637) 0.000 0.991 0.993 (0.285–3.461)
  PAM −0.347 (0.331) 1.100 0.294 0.707 (0.369–1.352)
 Thickness‡ 0.743 (0.196) 14.306 <0.001 2.102 (1.430–3.089)
 Basal diameter§ 0.322 (0.172) 3.501 0.061 1.379 (0.985–1.932)
 Therapy∥ (referenced to excision + brachytherapy)¶ 3.121 0.373
  Excision 0.711 (0.727) 0.955 0.328 2.036 (0.489–8.471)
  Excision + cryotherapy −0.818 (1.226) 0.466 0.504 0.441 (0.040–4.876)
  Irradiation 0.582 (0.918) 0.401 0.526 1.789 (0.296–10.813)
Multivariate analysis
 Model 1 (−2 log likelihood 333.378)
 Nonepibulbar location 1.269 (0.331) 14.694 <0.001 3.557 (1.859–6.805)
 Thickness‡ 0.533 (0.197) 7.345 0.007 1.704 (1.159–2.505)
Table 6.
 
Cox Regression Analysis for Survival in 194 Patients with Conjunctival Melanoma
Table 6.
 
Cox Regression Analysis for Survival in 194 Patients with Conjunctival Melanoma
Variable Coefficient (SE) Wald2 P Hazard Ratio (95% CI)
Univariate analysis
 Age at diagnosis* 0.018 (0.010) 3.425 0.064 1.018 (0.999–1.037)
 Nonepibulbar location 1.286 (0.294) 19.113 <0.001 3.617 (2.033–6.438)
 Uni/multilocular 0.439 (0.196) 5.210 0.025 1.551 (1.056–2.278)
 Origin† (referenced to melanoma)¶ 1.417 0.702
  Nevus −12.293 (354.707) 0.001 0.973 0.000 (0.000–8.948)
  Nevus + PAM 0.192 (0.644) 0.089 0.765 1.212 (0.343–4.279)
  PAM −0.333 (0.341) 0.954 0.329 0.716 (0.367–1.399)
 Thickness‡ 0.678 (0.200) 11.465 0.001 1.969 (1.330–2.916)
 Basal diameter§ 0.320 (0.176) 3.299 0.069 1.378 (0.975–1.946)
 Therapy∥ (referenced to excision + brachytherapy)¶ 2.788 0.426
  Excision 0.649 (0.728) 0.795 0.372 1.914 (0.460–7.972)
  Excision + cryotherapy −0.819 (1.226) 0.446 0.504 0.441 (0.040–4.873)
  Irradiation 0.544 (0.918) 0.351 0.553 1.723 (0.285–10.417)
Multivariate analysis
Model 1 (−2 log likelihood 319.991)
Nonepibulbar location 1.148 (0.337) 11.595 0.001 3.152 (1.628–6.103)
Thickness‡ 0.491 (0.202) 5.914 0.015 1.634 (1.100–2.426)
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×