Clinical and Cytogenetic Analyses in Uveal Melanoma

Abstract

purpose. Uveal melanoma is one of the most frequently occurring primary intraocular malignancies in the Western world. Cytogenetically these tumors are characterized by typical chromosomal losses and gains, such as loss of 1p, 3, and 6q and gain of 6p and 8q. Whereas most studies focus on known aberrations, in this one, cytogenetic changes were characterized and correlated with clinical and histopathologic parameters.

methods. Karyotypes of 74 primary uveal melanomas were analyzed with respect to the presence or absence of chromosomal gains and losses. In the analysis, classic clinical and histopathologic parameters were analyzed together with the chromosomal aberrations.

results. At a median follow-up of 43 months, 34 patients had died or had metastatic disease. Clonal chromosomal abnormalities were present in 59 tumors. The most frequent chromosomal abnormalities involved chromosome 8 (53%); loss of chromosome 3, p-arm (41%) and q-arm (42%); partial loss of chromosome 1, p-arm (24%); and abnormalities in chromosome 6 that resulted in gain of 6p (18%) and/or loss of 6q (28%). Less-frequent aberrations were abnormalities in chromosome 16, in particular loss of chromosome 16 q-arm (16%). In the univariate analysis, loss of chromosome 3, largest tumor diameter, gain in 8q, and mixed/epithelioid cell type in the tumor compared with tumors without these chromosomal changes or with a spindle cell type was associated with decreased disease-free survival. When corrected for confounding variables, significance of gain of 8q and cell type was decreased, whereas the significance of loss of chromosome 3p or 3q and largest tumor diameter remained the same.

conclusions. Monosomy 3 and largest tumor diameter are the most significant in determining survival of patients with uveal melanoma. Abnormalities in the q-arm of chromosome 16 are relatively common in uveal melanoma, but are not associated with survival or other cytogenetic or histopathologic parameters.

Uveal melanoma (UM) is the most common primary intraocular tumor in the Western world, affecting approximately 7 in 1 million people each year. Tumorigenesis and progression of cancer is in general preceded by the occurrence of genetic changes in normal cells. ¹ In this respect, UMs are quite homogenous, with a few tumor-specific cytogenetic aberrations. Some of these aberrations correlate with the metastatic potential of the tumor, resulting in metastatic disease followed by death. Recurrent aberrations in UM concern loss of 1p, monosomy of chromosome 3, loss of 6q and 8p, and gain of 6p and 8q. 

Loss of chromosome 1, p-arm, was observed in metastases, ² and concurrent loss of 1p and 3 is associated with decreased survival. ^{3 4} Furthermore, monosomy 3 is considered to be an early event in UM, and several studies have shown that it is a strong predictor of survival. ^{5 6 7} Loss of chromosome 3 is frequently associated with amplification of 8q, often seen as isochromosome 8, q-arm. ^{8 9}  

Recently, Hoglund et al. ¹⁰ elucidated a common genetic pathway for both uveal and cutaneous melanoma. Monosomy 3 probably occurs as an early event, and loss of 1p, 8p, and gain of 8q as secondary events. 

Regions of chromosomal loss are thought to harbor tumor-suppressor genes and regions of gain, oncogenes. Previous cytogenetic analyses have focused in general on the known aberrations. In this study, we performed cytogenetic analysis on short-term cell cultures of fresh tissue from 74 primary UMs to characterize all chromosomal changes and correlate these changes with clinical and histopathologic parameters. Significant prognostic parameters for UM, at high-risk for metastases, were identified. 

Materials and Methods

Patients and Tumor Samples

From March 1992 to April 2003, we collected tumor tissue of patients who underwent enucleation for ciliary body or choroidal melanoma. Informed consent was obtained before enucleation, and the study was performed according to the tenets of the Declaration of Helsinki. Fresh tumor tissue was obtained within 1 hour after enucleation and processed as described before. ³ Conventional histopathologic examination was performed on all tumors and the origin of the tumor was confirmed. Cytogenetic studies were also performed on stimulated peripheral blood samples of each patient to exclude the presence of congenital chromosome abnormalities. Follow-up data from time of diagnosis until the end of the study in December 2005 were obtained by reviewing the charts patient’s and contacting their general physicians. 

Cytogenetic Analysis

Chromosome preparations were made according to standard procedures and stained with acridine orange or atebrine to obtain R or Q banding. Cytogenetic abnormalities were described in accordance with the ISCN (International System for Human Cytogenetic Nomenclature, 1995). ¹¹  

Data Classification

Based on the cytogenetic analysis, tumors were classified for gain and or loss for all chromosomal regions, p-arm or q-arm. When different subclones were identified, only the cytogenetic findings of the largest clone were classified. Chromosomal regions with loss in >10% of all tumors and gain in >15% of all tumors were included for analysis. Tumors were identified as small (≤12 mm) and large (>12 mm). 

Statistical Analysis

The primary end point for disease-free survival (DFS) was the time to development of metastatic disease, whereas death due to other causes was censored. The influence of single prognostic factors on DFS was assessed using the log rank test (for categorical variables) or the Cox proportional hazards analysis (for continuous variables), and Kaplan-Meier curves were plotted to illustrate the differences in survival. To examine the possibility that other clinical, histopathologic, or chromosomal variations may affect the prognosis we performed Cox proportional hazards analysis for each confounding variable. An effect was considered significant if the P ≤ 0.05. The odds-ratios with corresponding probabilities were calculated to identify association between the different parameters (SPSS, ver. 11.0; SPSS, Chicago, IL). 

Results

Patients

From March 1992 to April 2003, of the 152 patients available for the study, chromosome analysis was successfully performed in 74. The clinical and histopathologic features of the 74 primary UMs are listed in Supplementary Table S1. The median age of the patients at the time of enucleation was 60 years (range, 21–87 years), 29 women and 45 men. One patient was lost to follow-up after 27 months. At the end of follow-up time 31 patients had died of melanoma-related disease, 3 had diagnosed metastases, 9 had died due to other causes, and 31 were still alive without metastases. The median follow-up time was 42.8 months (range, 6.4–164.4 months). 

Histopathology

All tumors were confirmed histopathologically as UM. Based on cell type, 16 tumors were classified as epithelioid, 24 as mixed, and 34 tumors as spindle. The mean tumor diameter and thickness were 13.2 mm (range, 6–19 mm) and 8.4 mm (range, 2–22 mm), respectively. Four tumors were located in the ciliary body and 70 were located in the choroid. Of the tumors located in the choroid four showed involvement of the ciliary body. 

Cytogenetic Analysis

Seventy-four UMs were analyzed for cytogenetic changes (Supplementary Table S1) and classified for gain and loss in all chromosomal regions (Table 1) . Clonal chromosomal abnormalities were present in 59 tumors. The most frequent chromosomal abnormality involved chromosome 8, trisomy of chromosome 8 or gain in 8q, most often in the form of an i(8q) (53%). Other abnormalities involved loss of chromosome 3, p-arm (41%) and q-arm (42%), partial loss of 1p (24%), and abnormalities in chromosome 6, resulting in gain of 6p (18%) and/or loss of 6q (28%). Other less-frequent aberrations were abnormalities of chromosome 16, in particular loss of chromosome 16, q-arm (16%; Fig. 1 ). Other chromosomal aberrations, such as loss of 6p, 9p, 15p, 15q, 21p, and 22p and gain of 2p, 2q, 7q, 9p, and 11q were present but did not reach 10%. 

Statistical Analysis

Univariate analysis was performed for all clinical, histopathologic, and cytogenetic parameters (Table 2 , Fig. 2 ). Univariate analysis of the single prognostic factors showed significant lower DFS in patients with loss of chromosome 3, largest tumor diameter (LTD), gain of 8q, and a mixed/epithelioid cell type in the tumor compared with tumors without these chromosomal changes or of a spindle cell type. Other potential prognostic factors such as gender, age at time of diagnosis, and tumor location (i.e., involvement of ciliary body) did not reach significance. Also chromosomal changes such as loss of chromosome 1p, gain of 6p, and loss of 6q were not significantly associated with DFS. To examine the possibility that other clinical, histopathologic, or chromosomal variations may affect the prognosis, we performed Cox proportional hazards analysis for each confounding variable (Table 2) . Parameters presented in the columns are the investigated prognostic parameters; in the rows the same parameters resemble the confounders with a possible modifying effect. Significance of loss of 3p and 3q did not alter after correcting for the possible confounders. A similar pattern was observed for LTD and cell type. Odds ratios were calculated to identify association between the different parameters (Table 3) . Associations were shown for loss of chromosome 3 with gain of 8q, loss of 8p, vascular patterns and LTD (>12 mm), and a weak association with mixed-epithelioid cell type. Presence of vascular patterns and LTD (>12 mm) showed also association with gain of 8q. Associations were also present for loss of 1p with loss of 16q and loss of 3p, and weak association with cell type, vascular patterns LTD, 3q loss, and 8q gain. Loss of 6q was weakly associated with gain of 8q. Loss of 16q was weakly associated with gain of 8p. 

Discussion

By means of karyotyping we analyzed chromosomal aberrations in UM. Previous reports have revealed that abnormalities in chromosome 1, 3, 6, and 8 occur in a nonrandom fashion in UM. ^{9 13 16} Some of these tumor-specific aberrations have been associated with the metastatic potential of the tumor. In this study, loss of 1p and chromosome 3 and aberrations in chromosomes 6, 8, and 16 were most often encountered. Furthermore, tumors with abnormalities of chromosome 3, gain of 8q, epithelioid–mixed cell type, and a larger tumor diameter were strongly associated with a poor prognosis. 

In UM, numerous parameters have been used to predict survival, with the conventional parameters being tumor size, tumor location, cell type, and vascular patterns. ¹² None of these factors is entirely solid, and there has been considerable variation in interpretation among observers. In contrast to a previous report, ¹³ we did not find chromosomes 11 and 21 to occur very often (Table 1) , and therefore these aberrations were not included in the analysis. In addition, we identified loss of 16q. Loss of chromosome 16, in particular 16q, also mentioned in earlier reports ^{10 13} occurred in >10% of the UMs. Even though it was not significantly associated with DFS, it still may be involved in tumor progression. A remarkable association was shown for loss of 16q with loss of 1p. Delineation of a region on 16q may depict a region of interest with possible candidate genes. Other tumors, such as breast cancer and neuroectodermal tumors have also shown deletion on 16q. ^{14 15} In these tumors, candidate genes have not yet been identified. Because UM cells are derived from neuroectodermal tissue this might be of potential interest. In many reports outcome was correlated with tumor location. ^{7 16} Because we had limited sample size in the group tumors located in the ciliary body, we were not able to make reliable assumptions on association of outcome with tumor location. LTD in our study was histopathologically measured. This parameter may be used noninvasively in a clinical setting (measurement on ultrasound) and may be the most reliable noninvasive prognostic parameter. However, there is a variation between clinical and histopathologic measurements. The tumor size measured on ultrasound is in general larger than the histopathologic measurement. In contrast, the detection of specific chromosomal aberrations by routine fluorescence in situ hybridization (FISH), comparative genomic hybridization (CGH), and karyotyping provides a more objective measurement of potential tumor behavior. Identification of monosomy 3 in a tumor sample is widely accepted as the most reliable prognostic parameter. ^{5 6 7} Monosomy of chromosome 3 is considered an early event, occurring before alterations of chromosome 8, 1, and 6. ^{5 6 7} Moreover, it may cause isochromosome formation of especially isochromosomes 6, p-arm, and 8, q-arm. ^{8 9} Table 3may also support this hypothesis, because the odds ratios for loss of chromosome 3, p-arm or q-arm, and gain of 8q or loss of 8p were higher than the combination of losss of 3p or 3q and gain of 8p. However, in our series, we cannot conclude the same for isochromosome 6, p-arm. In addition, gain of 8q was significantly associated with survival in the univariate analysis (Table 2) , but when corrected for confounding variables, such as vascular pattern, cell type, LTD, and 3p or 3q loss, significance was absent, implying that gain of 8q occurs together with at least one of those other variables. On the contrary, when this same procedure was followed for 3p or 3q loss we observed that the significance remained. In Table 3the odds ratios were shown for different chromosomal parameters. If we put the odds ratios in the following order, 8q gain, and consequently 8p loss, follows monosomy 3, and loss of 1p and 16q occur thereafter. This is consistent with the findings observed by Hoglund et al. ¹⁰ Moreover, tumor diameter is associated with most of the chromosomal aberrations, implying that larger tumors have more aberrations. Our study involves patient samples from relatively large tumors that were treated by enucleation. Considering monosomy 3 as an early event, ¹⁷ it is likely that it would be observed in even the smallest amount of tissue despite the heterogeneity of UM. Though, there are no studies to date that confirm the uniform distribution of cytogenetic abnormalities in UM, and it is at least theoretically possible that small amounts of tissue (e.g., used for karyotyping, FISH, and CGH) do not contain the cytogenetic markers of interest. 

 Supported by the Revolving Fund of the University Hospital Rotterdam, The Dutch Foundation Fighting against Blindness, The Rotterdam Blindness Foundation, The Hague Ophthalmic Foundation, Grant EUR 2002-2695 from the Dutch Cancer Society; and the Henkes Stichting, The Netherlands.

 Submitted for publication January 31, 2006; revised March 31 and April 22, 2006; accepted June 19, 2006.

 Disclosure: E. Kilic, None; W. van Gils, None; E. Lodder, None; H.B. Beverloo, None; M.E. van Til, None; C.M. Mooy, None; D. Paridaens, None; A. de Klein, None; G.P.M. Luyten, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Recurrent Changes in Karyotype of Primary Uveal Melanoma

Loss and gain >10% of all tumors (n = 74)
1p loss	18 (24)
3p loss	30 (41)
3q loss	31 (42)
6p gain	13 (18)
6q loss	21 (28)
8p gain	13 (18)
8p loss	18 (24)
8q gain	39 (53)
16q loss	12 (16)
Loss and gain <10% of all tumors
2p gain	4 (5)
2q gain	4 (5)
6p loss	7 (9)
7q gain	4 (5)
9p gain	4 (5)
9p loss	7 (9)
11q gain	7 (9)
15p loss	7 (9)
15q loss	7 (9)
21p loss	7 (9)
22p loss	7 (9)

 Data are the number of tumors (% of total group).

Figure 1.

 

Karyotype of tumor EOM 63. This tumor showed chromosomal changes caused by UM: −3, I(6)(p), I(8)q (multiple copies), and del(16)(q21).

Table 2.

 

Prognostic Significance of Clinical, Histopathological, and Chromosomal Aberrations in Uveal Melanoma*†

 

Figure 2.

 

Kaplan-Meier survival curves for clinical, histopathologic, and chromosomal aberrations.

Table 3.

 

Relation between Different Histopathological, Clinical and Chromosomal Aberations

 

Supplementary Materials

Supplementary Table S1 - (PDF) 

The authors thank Anne Hagemeijer, Rosalyn Slater, and Ellen van Drunen for performing most of the cytogenetic analyses during the early years of the study.