Uveal melanoma – testing of abnormalities of chromosome 3 and 8 in the Czech Republic

Aim. The aim of this cytogenetic study is to confirm the significance of chromosome 3 loss (monosomy 3) and of the gain of chromosome 8 as prognostic markers in histopathological samples of enucleated eyes with uveal melanoma in the Czech population. Methods. This is a retrospective study of 52 enucleated eyes. Chromosome 3 and 8 status were tested by CISH, and in a few samples FISH was used. The correlation between monosomy 3 and gain of chromosome 8 and clinical features (histopathological type, size of the tumour) were evaluated. A follow up for the detection of metastases was conducted in all patients. The statistical significance of chromosomal abnormalities as a prognostic factor for the development of metastases was calculated. Results. There were 52 patients, 27 men (51.9%) and 25 women (48.1%) enrolled in our study group. The mean age was 63 ± 14 years. Loss of the one copy of chromosome 3 (monosomy 3) was detected in 26 (50.0%) patients, monosomy 8 was present in 34.6% of patients with monosomy 3. After 5 years there were no metastases in 82% of patients without monosomy 3 as opposed to 40% of patients with monosomy 3. We confirmed a statistically significant association between progression free survival (PFS) and the presence of monosomy 3 (P=0.017). The association between PFS and gain of chromosome 8 was significant as well (0.010). Conclusions. Our data showed the association of progression-free survival with the presence of monosomy 3 in uveal melanomas. We provided a good prognostic value of monosomy 3 in uveal melanoma.


INTRODUCTION
Although uveal melanoma (UM) is a rare cancer, it is, however, the most frequent primary intraocular malignant tumour in adults 1 . The incidence of uveal melanoma in Europe was analysed in a European Cancer Registry-based study (EUROCARE) where it was estimated that there were 1.3-8.6 cases per million per year in Europe 2 . These data are almost equal to a study by the United States National Cancer Institute (SEER) where the overall incidence of uveal melanoma was 5.1 per million per year in the United States 3 . The incidence of uveal melanoma has remained stable since the 1970s with approximately 7000 new cases of uveal melanoma worldwide 4 .
The mean age at presentation is 60 years. There is a progressive increase with age, peaking at 70-75 years and then reaching a plateau. There is a slight predominance in males 5 . Uveal melanoma typically affects Caucasians in 98% of cases. The risk factors are light skin, inability to tan and light iris colour 6 . This is the reason we can see a north-to-south decreasing gradient of melanoma incidence in Europe 2 . The role of exposure to sunlight has been broadly discussed, but has not been proven 7 . Other predisposing conditions include choroidal nevus and ocular melanocytosis 8 .
Ocular melanoma arises from melanocytes that are differentiated from pluripotent neural crest cells 1 . Uveal melanoma can be located in the iris (4% of patients), in the ciliary body (6% of patients) and in the choroid (90% of patients) (ref. 9 ).
Despite a high success rate of therapy for the primary tumour, 50% of patients die from metastases 10,11 . It is presumed that this is due to the presence of micrometastases already at the time of the primary diagnosis 12 . For a long time, effective metastases treatment had been lacking. In recent years, immunotherapy has begun to be used. There are currently several studies in this area (https://app. emergingmed.com/mrf/trials/). Some immunostimulants have been approved by the FDA: Imlygic (T-VEC), Yervoy + Opdivo, Opdivo (nivolumab), Keytruda (pembrolizumab), Yervoy (ipilimumab), Interleukin-2 (IL-2) and Interferon alpha 2-b.
Of great importance is to identify high risk patients who may develop metastases. Clinical features predisposing a higher risk for metastasis are large tumour size, ciliary body involvement and extrascleral extension of the tumour. Shields evaluated the risk of tumour size for metastasis where every millimetre increases the risk of metastasis by 5% (ref. 13 ). The above factors are the basis of the Tumour, Node, Metastasis (TNM) description 14 . The risk of metastasis is also dependent on the histopathological characteristics of the tumour. The first Callander classification was modified by McLean who divided melanoma cell type into spindle, epithelioid and mixed 15 . Risk factors for metastatic spread are epithelioid cell type, high mitotic activity, increased tumour-infiltrating lymphocytes, the presence of networks of closed microvascular loops, and vascular invasion 16 .
The presence of chromosomal aberrations provides the possibility of a more accurate assessment of the risk of metastasis. Frequently observed non-random gross chromosomal changes are specific for uveal melanoma. They are of prognostic significance for the progression of the tumour. Prescher described the loss of one copy of chromosome 3 (monosomy) in tumours with a high tendency for metastatic spread 17 . This strong correlation was shown in several clinical studies [18][19][20][21] . The other chromosomal mutation associated with reduced life survival is a gain of 8q. Other unfavourable chromosomal alternations are a loss of 6q or 1p; contrariwise, a gain of 6p is a sign of a better prognosis 22 .
Prognostic biopsy evaluating metastatic risk is standard in some oncology centres in the management of patients with uveal melanoma in the United States 23 and in Europe 18,21,[24][25][26][27] . Presently, chromosomal analysis is not used in the Czech Republic.
The aim of this cytogenetic study is to confirm the significance of chromosome 3 loss (monosomy 3) and of the gain of chromosome 8 as a prognostic marker in histopathological samples of enucleated eyes with uveal melanoma. We want to map the genomic profile in the Czech Republic and we want to report the first results from the Czech population. Finally, we want to discuss the clinical implications of our cytogenetic findings. The aim of the study is to show the effectiveness of introducing a cytogenetic examination into clinical practice in patients with uveal melanoma in the Czech Republic.

MATERIAL AND METHODS
A retrospective non-randomised study of eyes of 52 patients who underwent enucleation due to the diagnosis of choroidal melanoma or ciliary body melanoma. This cytogenetic study was conducted between January 2008 and December 2017.
The patients were treated in the Department of Ophthalmology of the University Hospital at Masaryk University in Brno.
The clinical data included age and gender of the patients, tumour location (ciliary body, posterior pole, nasal part, and temporal part) and histological type. The measurement of thickness and diameter (mm) of the tumour were assessed using ultrasound, A and B scans (Compact touch, Quantel Medical, France). For statistical evaluation we used the mathematical approximation of the total size of the tumour. Other clinical features included the presence of rupture of Bruch's membrane, secondary retinal detachment and orbital extension. The staging of the disease was accessed at the time of the diagnosis. All of the patients were followed up in our Department, the staging of the disease was repeated every 3 months during the first 2 years of the following period and then every 6 months.
Other treatments, such as brachytherapy or Cyberknife, were not carried out because of the size of the tumour or because of the tumour being located too close to the optic nerve.
The follow-up data, including the time and cause of death, were obtained from clinical records and by contacting the Institute of Health Information and Statistics of the Czech Republic. Based on these data we distinguished the metastatic death (disease specific survival rate) and death due to other causes. Informed consent was obtained from all the participants. We followed Good Clinical Practice (GCP) guidelines according to the Declaration of Helsinki.
Archival formalin-fixed, paraffin-embedded tissues were used in this study. The samples were stored in the archives of the Department of Pathology of the University Hospital at Masaryk University in Brno. A histopathologic examination was conducted according to standardised protocols. Haematoxylin and eosin stained sections were reviewed by a pathologist to determine the histological type and to choose the sample for the cytogenetic study. We tested samples for monosomy 3, and in patients with monosomy 3 we also tested the gain of chromosome 8 by chromogenic in situ hybridisation (CISH) and by fluorescence in situ hybridisation (FISH).

Chromogenic in situ hybridisation (CISH)
The CISH method was performed on tissue sections in VENTANA BenchMark Ultra (Ventana Medical Systems, USA) using the VENTANA ultraView RED ISH DIG Detection Kit (Ventana Medical Systems, USA). Tissue sections were deparaffinised and pretreated with Cell Conditioning 2 (CC2) at pH 6 at 82°C. Enzymatic digestion of proteins was performed with ISH Protease 3 for 32 min. The double-stranded DNA was denatured for 12 min at 80 °C, the hybridisation of the probe proceeded at 44 °C for eight hours. Chromosome 3 DIG Probe (Ventana Medical Systems, USA) hybridising to the centromere of the chromosome was used. A stringency wash was performed at 72 °C to wash off unbound or weakly bound probes. The ultraView Red ISH DIG detection kit was used for the detection of the CEN3 signal. Tissue sections were then counterstained in hematoxylin II and bluing reagent to enhance the contrast. Images were scanned by a DM 5500 B microscope equipped with a Leica DFC 290 HD camera. Red signals were analysed using Leica LAS AF software (Leica Microsystems GmbH, Wetzlar, Germany). To define the monosomy of chromosome 3 we used a cut-off value of 20%.

Fluorescence in situ hybridisation (FISH)
For the detection of chromosome 3 monosomy, the ZytoLight ® CEN3 Probe was used (ZytoVision GmbH, Bremerhaven, Germany). Hybridisation was performed according to the manufacturer's instructions. Images were scanned by a DM 5500 B microscope equipped with a Leica DFC 290 HD camera. Fluorescence signals were analysed using Leica LAS AF software (Leica Microsystems GmbH, Wetzlar, Germany). 100 cells per case were analysed. To define the monosomy of chromosome 3 we used a cut-off value of 20%.
For the analysis of chromosome 8q status, ZytoLight ® SPEC FGFR1/CEN 8 Dual Color Probe and ZytoLight ® SPEC MYC/CEN 8 Dual Color Probe were used. Gain of the long arm of chromosome 8 was defined by amplification of the MYC gene (located in the chromosomal region 8q24.21) and copy loss of the FGFR1 gene (located in the chromosomal region 8p11.23-p11.22). To avoid uncertainty, another probe was used located in the long arm of chromosome 8 -RUNX1T1 located in the region 8q21.3-q22.1.

Statistical analysis
Continuous variables were described using mean (standard deviation) and median (25th and 75th percentile), and the statistical significance of differences was tested by the Mann-Whitney test. Categorical variables were described by absolute and relative frequencies. The statistical significance of association between categorical variables was assessed with Fisher's exact test. Disease specific survival and progression free survival were estimated using the Kaplan-Meier method with the log-rank test as a measure of statistical significance of different survival rates between groups. The Cox PH regression model was applied to estimate hazard ratios (HR) between groups. The level of statistical significance was set to 0.05 in all hypothesis testing.

RESULTS
There were 52 patients, 27 men (51.9%) and 25 women (48.1%) enrolled in our study group. The mean age of the patients was 63 ± 14 (median 67, 56-71 years). Table 1 shows the patients' principal clinical characteristics. Posterior uveal melanoma was present in 40 patients (76.9%) and ciliary body melanoma in 12 (23.1%) patients. The average thickness of the melanoma was 8.3 ± 3.7 mm and the average size 14.6 ± 4.0 in diameter. We observed the rupture of Bruch's membrane in 23 (44.2%) and secondary retinal detachment in 39 (75.0%) cases. Extraocular extension was diagnosed in 2 (3.8%) patients. The most common histological cell type was mixed, in 19 (38.0%) patients; epithelioid cell type was present in 7 patients (14.0%); and spindle cell type in 24 (48.0%)   patients. The median follow-up was 4.3 ± 3.2 years. Metastases after 5 years were present in 27.4%, and after 5 years 40.7% of patients had died. In all the cases distant metastases were present, in one patient local recurrence was also observed. The first metastases were diagnosed in the liver in 17.3% of patients. Disease specific survival was not statistically significant when associated with the size of the tumour (P=0.064), the thickness of the tumour (P=0.053) and the size in diameter (P=0.134). Progression free survival was statistically significant when associated with the diameter size above 14 mm (P=0.008) (Fig. 1); this was not statistically significant regarding the thickness of the melanoma (P=0.303) and with the size of the melanoma (P=0.250).
Loss of one copy of chromosome 3 (monosomy 3) was detected in 26 (50.0%) patients. A more detailed description of the subgroup with monosomy 3 is shown in Table 2. Patients with monosomy 3 were a little older and they had a larger tumour with a greater thickness as well as diameter, and rupture of Bruch's membrane was more frequent; but the results are not statistically significant. Disease specific survival and progression free survival were longer in patients without monosomy 3. Disease specific survival did not quite reach a statistically significant level (P=0.053), but the difference of progression free survival between both groups was statistically significant (P=0.017), Fig. 2. When we evaluate five-year survival and monosomy 3, then patients with monosomy 3 lived longer than 5 years (61% of patients) and without monosomy 3 (93% of patients). The strongest association is between monosomy 3 and progression free survival, HR 3.21 (Table 3). After 5 years there are no metastases in 82% of patients without monosomy 3 as opposed to 40% of patients with monosomy 3. The risk of progression       to metastatic disease is 3.2 fold higher in patients with monosomy 3 ( Table 3). The gain of chromosome 8 was present in 9 patients with monosomy 3 (34.6%). The subgroup of patients with monosomy 3 and gain of chromosome 8 is described in Table 4. We did not find a statistically significant association of this subgroup with disease specific survival (P=0.317) (Fig. 3). But progression free survival was statistically significant (P=0.010) (Fig. 4).

DISCUSSION
One of the typical markers of the cancer is the presence of genetic mutations. Uveal melanoma has only a few specific mutations we can identify. A specific non-random occurrence of chromosome 3 abnormalities in melanoma was published in the early 1990s (ref. 28 ). A few patients showed chromosomal abnormalities of chromosome 3, 6 and 8. Complete loss of chromosome 3 (monosomy 3) was present in 43% of patients and was often connected with chromosome 8 abnormalities. These changes are probably secondary in uveal melanoma in contrast to cutaneous melanoma and they are probably the reason for the genetic difference between cutaneous and uveal melanoma 17 . These observations of monosomy 3 were confirmed by other authors 29,30 . The first observation of the loss of one copy of chromosome 3 and its association with metastatic spread of UM was published by Prescher in 1996 (ref. 17 ). In 54 patients they found monosomy 3 in 56% of enucleated eyes; this abnormality was the most significant risk factor for metastatic disease after enucleation 17 . During the following years, several clinical trials of enucleated eyes were conducted to confirm the association between certain genetic mutations and the risk of developing metastases 18,24,27,31,32 . Some clinical work carried out has dealt with the association of genetic abnormalities and tumour characteristics. Sisley showed a significant association of cytogenetic characteristics with ciliary body involvement 18 . However, this observation was probably influenced by the composition of the study group in which 57% of the patients had ciliary body melanoma. Kilic and Damato emphasised the connection with tumour size 20,33 .
Cytogenetic testing of chromosomal abnormalities in fine-needle aspiration biopsy (FNAB) samples from enucleated eyes was presented by Sisley 34 . In 2002 Naus reported that genetic testing of samples obtained by fineneedle aspiration biopsy could be as reliable as testing by open biopsy after enucleation in melanomas 35 . They performed fine-needle aspiration of tumours in 40 enucleated eyes; for determining genetic abnormalities the FISH method was used. The first published study of FNAB in vivo before brachytherapy of uveal melanoma included eight patients 36 . The most extensive work presenting results with genetic testing of FNAB samples in patients with uveal melanoma is the study of 500 cases published by Shields et al. They described complete monosomy 3 in 25% of patients and partial monosomy 3 in 27% of patients. Shields tried to calculate the cumulative probability according to the tumour thickness and cytogenetic    characteristics. At 3 years the cumulative probability of metastases for tumours with monosomy 3 was 0% for small melanomas, almost 25% for medium and 57.5% for large melanomas. In contrast the cumulative probability of tumours with disomy was 0% for small, 1.4% for medium and 23.1% for large tumours. Shields suggested that the higher rate of incidence of metastases in disomy 3 tumours could be caused by a sampling error of FNAB in larger tumours 23 .
The most important predictive factors of monosomy 3 included increasing tumour thickness and increasing distance to the optic disc. According to the observation of some authors, monosomy 3 is related to another prognostic factor, such as diameter size and ciliary body involvement 28,33 .
The basic characteristic of our study group corresponds to other study groups. The study group contains fewer patients, because this genetic test is not routinely used in the Czech Republic. Our Department of Ophthalmology is the oncology centre for the South Moravian region, which has 1 200 000 inhabitants. The frequency of monosomy 3 in our study group is comparable with other studies. When comparing the incidence of monosomy with these studies, we should take into account that we have examined enucleated eyes and this implies that the tumours are larger. Le Guin studied monosomy 3 in the subgroup treated with brachytherapy and by enucleation. The proportion of monosomy 3 is higher in tumours treated by enucleation (71% vs. 32%) (ref. 12 ).
In our study group tumours with monosomy 3 were a little higher, more often localised on the posterior pole and more often associated with the rupture of Bruch's membrane, but this was not statistically significant. Alterations of chromosome 3 and 8 were found occurring together in 17.3% of patients. These melanomas, with both chromosomal abnormalities, were less frequent in women, and these tumours were more frequently present in the posterior pole; they were also higher and larger. These associations were not statistically significant. At the edge of statistical significance, despite the smaller study group size, was the connection between epithelioid histological type and presence of both chromosomal abnormalities. We observed a statistically significant association between monosomy 3 and progression free survival. There are no metastases in 82% of patients without monosomy 3 as opposed to 40% of patients with monosomy at a 5-year follow-up. We can compare our study group with the subgroup of large tumours in the study of Shields. The cumulative probability of metastases in patients with disomy 3 after 3 years 23 was 23%, and in patients with monosomy 58%. The risk of progression to metastatic disease is 3.2 fold higher in patients with monosomy 3 in our study group where larger tumours are involved. Prescher already reported a 50% metastases rate in patients with monosomy 3 after 3 years 17 . Le Guin showed a significant association between monosomy 3 and tumour-related survival not only in patients after enucleation but also in the subgroup after brachytherapy 12 .
Damato published that 5% to 20% of disomy 3 uveal melanoma patients unexpectedly develop metastases 33 . There were 3 patients without monosomy 3 with the presence of metastasis at the initial examination in our study group. They were women, the diameter of the basis of the tumours were 15-20 mm. In one patient we observed extrascleral spread of melanoma. It has been suggested that translocations or partial deletions of chromosome 3 are the cause of this 37,38 . Attempts have been made to identify metastasis-suppressor genes, but without success 31 .
A limiting factor of this cytogenetic examination is a group of tumours with the loss of only a part of the chromosome, referred to as partial monosomy 3. It is probably due to a mis-sampling of tumour samples with monosomy 3 or the incorrect interpretation of a sample with partial monosomy 3 (ref. 25 ). Lake described Multiplex Ligation-Dependent Probe Amplification (MLPA) that can detect monosomy 3 in these "unfavourable" uveal melanomas, whereas FISH could not, especially in small tumours 39 . This is probably due to a higher sensitivity of the MLPA method and by cellular heterogeneity in the sample. The frequency of this abnormality differs greatly, from 0%, over 0.6% (ref. 32 ) to 4% (ref. 25 ) or 5.3% (ref. 23 ). The mortality rate in these patients is comparable with monosomy 3 patients according to Damato 33 , in contrast to the observation of Shields and Thomas 23,25 .
Vaarwater demonstrated that specificity was equal for FISH and MLPA (ref. 40 ). We did a retrospective study, thus only formalin-fixed paraffin-embedded (FFPE) tissue sections were available. Therefore, in situ hybridisation was a better choice than MLPA, which requires preferably fresh tissue. Monosomy 3 was tested by CISH by allowing the examination of CISH signal and tissue morphology simultaneously. Nevertheless, we did not detect any signals in the few samples, thus FISH was used in those cases.
The gain of chromosome 8 influences the metastatic progression of tumours with monosomy 3, but not of tu-mours with disomy 3. This coincidence of chromosome 3 and 8 abnormalities was referred by Patel 41 . The gain of chromosome 8 is known as late change during melanoma development and it is related to large tumour size 20 . The higher risk of metastatic death was present in patients with monosomy 3 with chromosome 8 abnormalities 18,25,42 . The combination of the anatomic and genetic characteristic as a prognostic factor has been shown by Damato 33 , Dogrusöz 27 and van den Bosch 26 . Even a slightly stronger high correlation between tumour diameter and gain of chromosome 8 than with monosomy 3 was presented by van den Bosch 26 . In all the patients with low percentage monosomy 3 who died of metastasis during a follow-up period, they detected a high percentage gain of chromosome 8q. This is probably due to the great importance of chromosome 8 abnormalities -in low percentage monosomy 3 tumours, the high percentage gain of chromosome 8 is an indicator of worsening patient survival. Patel and Bronkhorst presume the cut-off limit to be 30%, because of the sensitivity of the probe 41,43 . On the other hand, van den Bosch did not consider the threshold 30% as the limit that can be used to exclude patients from the high risk group; he showed that a significant number of patients with 30% of monosomy 3 died of metastasis in his study group 26 .
It is suggested that monosomy 3 tumour cells acquire their metastatic potential at an early stage in contrast to tumours with disomy 3, which acquire metastatic potential later 25 . Damato supported the idea of the transformation model. It was proposed that a natural cause is responsible for the subsequent loss of disomy 3 during the progression of the tumour 33 . However, later detailed genetic expression profiling (GEP) of monosomy 3 and disomy 3 tumours revealed that the development of monosomy 3 melanoma from disomy 3 melanoma is not probable 44 .
In summary our data showed an association of progression free survival with presence of monosomy 3 and gain of chromosome 8 in uveal melanomas. We provided a good prognostic value of monosomy 3 in uveal melanoma. Only enucleated tumours were included, which implied that we studied larger tumours only. A limiting factor of our study group was the small number of patients, however, this was the first ever larger cytogenetic testing carried out in the Czech Republic. A cytogenetic study of a smaller study group of 11 patients was published by Sobotova 45 .

CONCLUSION
Genetic testing in many countries is a standard diagnostic procedure in patients with uveal melanoma. According to the established risk of metastasis, we can adjust the interval of regular examinations. Another reason is that patients want to know their risk of developing metastases 46 . The importance of genetic testing will probably continue to increase due to the progressive development of immunotherapy, which would influence the development of metastases in patients with uveal melanoma.