Uveal melanoma has an incidence of 5.2 per million people and is the most common primary ocular tumor in adults [1]. Local recurrence is rare following identification and treatment, but up to 50% of patients develop metastatic disease, most commonly to the liver and lungs [2]. There are no treatments for metastatic disease with poor 5-year survival outcomes [1]. Thus, the discovery of prognostic indicators and predictors of development of metastatic disease has been an important area of research.
There are a number of clinical, histopathologic, and cytogenetic features that have been associated with an increase in metastatic risk. From a clinical perspective, age, gender, tumor size, tumor location, and presence of extraocular extension have been shown to have significant prognostic value [3]. Histopathologically, there are a number of factors that convey prognostic value including tumor cell morphology (e.g., epithelioid vs spindle) and mitotic activity, as well as various receptor and antigen expression levels [4]. Some of the most significant prognostic indicators for uveal melanoma, however, are its cytogenetic features.
Chromosomal abnormalities, such as chromosome 3 loss, have been shown to be predictive of disease-specific mortality [5]. Monosomy 3 has a reported frequency of approximately 40% in uveal melanomas, making it one of the most common genetic alterations in this disease [6]. Originally established as a significant prognostic indicator in 1996, chromosome 3 loss is one of the most well established predictive factors in uveal melanoma [7]. In the original study, Prescher et al., examined 54 cases of uveal melanoma treated by enucleation. Monosomy 3 was detected via karyotyping and comparative genomic hybridization in 30 of these cases. On 3-year follow up, 50% of these cases had developed metastases, while no patients in the disomy 3 group had developed metastases. Since 1996, additional studies have confirmed the negative prognostic value of monosomy 3 and suggest its value is most useful when interpreted in the context of clinical and pathology features such as tumor size and cell type of the tumor [5, 8]. Current research has focused on partial chromosomal alterations. The frequency of these alterations varies widely in the literature and their clinical significance is still under debate. Some reports have implicated these alterations in metastatic disease, while other results have suggested they have no significant clinical value [9].
Two of the most common molecular pathology techniques used to assess for monosomy 3 are FISH and CMA. FISH uses fluorochromes linked to DNA probes to rapidly identity and visualize targeted DNA sequences under the fluorescence microscope. First used in uveal melanoma in 1997, FISH is now widely used in the study of uveal melanoma cytogenetics, specifically in the detection of monosomy 3 [10].
Another common technique is CMA, which provides an alternative method for assessing chromosome copy number variations. One form of CMA, known as array comparative genomic hybridization (CGH), combines the use of a microarray with the classic CGH technique of in which two fluorochromes are used to differentially label genomic material from two samples, a test and a control. The labelled samples are then competitively hybridized to either reference metaphase spreads (classical CGH) or a microarray created from cloned DNA fragments which represent the normal genome (array CGH). Signal intensity is then compared between test and control fluorochromes to identify variations in copy numbers in the test genome [11]. CGH was first described in the analysis of chromosomal aberrations in uveal melanoma in 1996 [12]. Overtime, CGH has been recognized as an effective method for detection of chromosome 3 abnormalities and is especially useful for localization of small chromosomal aberrations to specific chromosomal regions [13]. Additionally, CGH has been demonstrated to effectively identify genomic copy variants that can distinguish benign nevi from melanomatous lesions [14]. An alternative form of CMA, single nucleotide polymorphism (SNP)-based chromosomal microarrays, uses the absolute intensity of a single fluorescent signal at a polymorphic chromosomal position to detect a gain or loss of genomic material at that chromosomal location [15]. Both SNP-based chromosomal microarray and FISH were implemented for both cases in this study.
Although widely used in practice today, these methods each have a number of limitations.
Based on a review published in 2012, the mean detection rate of monosomy 3 using various FISH techniques studied was only 48% [16]. Reliability is also influenced by tissue sampling technique, with highest rates of detection of monosomy 3 associated with fresh tissues when compared to fine needle aspiration biopsy [17]. Perhaps the most significant factor contributing to the limited success of the FISH technique is tumor heterogeneity seen in uveal melanoma. When FISH is used to detect chromosomal 3 aberrations in fine needle aspirations, it is possible the sampled area will not contain enough cells with chromosomal 3 loss to meet the threshold for detection [18]. Thus, FISH results are heavily influenced by both sample area and threshold levels [19, 20]. An additional limitation of the FISH technique is its limited ability to detect partial chromosome loss. For example, centromeric probes may be effective at detecting monosomy 3, but often miss partial deletions and thus underestimate chromosomal loss [21].
CMA also has limitations in its ability to detect chromosome 3 aberrations. Similar to FISH, CMA detection of monosomy 3 in uveal melanoma is limited by tumor heterozygosity as well overall malignant cell content. As discussed above, uveal melanomas exhibit tumor heterozygosity with respect to chromosome 3 copy number. Thus, CMA results are significantly influenced by detection thresholds used during analysis. Previous research has demonstrated cases in which CMA has miss-reported chromosome 3 status using a conventional cutoff of 20% of cells for detection [22]. Lastly, sample type and preparation have also been shown to affect CMA results. Previous research has demonstrated a significant amount of background signal present in material obtained from samples embedded in paraffin which can affect the ability of CMA to detect chromosomal aberrations in these samples [23].