Article Text

Concordant chromosome 3 results in paired choroidal melanoma biopsies and subsequent tumour resection specimens
  1. Sarah E Coupland1,
  2. Helen Kalirai1,
  3. Vivian Ho2,
  4. Sophie Thornton1,
  5. Bertil E Damato2,3,
  6. Heinrich Heimann2
  1. 1Department of Pathology, Royal Liverpool and Broadgreen University Hospital Trust (RLBUHT), University of Liverpool, Liverpool, UK
  2. 2Department of Ophthalmology, Royal Liverpool and Broadgreen University Hospital Trust (RLBUHT), Liverpool, UK
  3. 3Ocular Oncology Service, University of California, San Francisco, California, USA
  1. Correspondence to Prof Sarah E Coupland, Molecular and Clinical Cancer Medicine, Institute of Translational Medicine, University of Liverpool, 6th Floor Duncan Building, Daulby Street, Liverpool L69 3GA, UK; s.e.coupland{at}


Background/aim The study's aim was to compare chromosome 3 aberrations of choroidal melanoma (CM) as determined by multiplex ligation dependent probe amplification (MLPA) or microsatellite analysis (MSA) in intraocular tumour biopsies with those results obtained from subsequent endoresection/enucleation of the same CM.

Methods A retrospective cohort of 28 patients with CM seen between 2007 and 2014 at the Liverpool Ocular Oncology Centre was analysed. Prognostic genetic testing, for chromosome 3 status, was performed on all tumour specimens, either by MLPA or MSA, depending on DNA yield. In nine cases genetic testing was performed on a sample taken after radiotherapy; four of these had genetic information pre- and post-radiotherapy.

Results Fourteen biopsy specimens were analysed by MLPA and 14 by MSA. Twenty-seven endoresection or enucleation specimens were analysed by MLPA, and a single enucleation specimen by MSA. Chromosome 3 data showed prognostic concordance for the patient-matched samples in all 28 cases including 4 cases where samples were taken pre pre- and post radiotherapy. Thirteen cases were classified as monosomy 3 and 12 as disomy 3. Two cases had a loss of chromosome arm 3q in both samples and a single case showed loss of 3p in the biopsy sample with complete monosomy 3 in the subsequent enucleation sample taken 5 months later.

Conclusions Intraocular biopsy of CM yields similar prognostic information to larger surgical specimens. Initial evidence, that genetic testing can be successfully conducted post radiotherapy, is also provided.

Trial registration number NITRO trial, ISRCTN35236442.

  • Choroid
  • Neoplasia
  • Pathology

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Almost 50% of patients with choroidal melanoma (CM) develop metastatic disease to the liver after successful local treatment of the primary tumour. Metastatic risk stratification of patients with CM takes into account clinical, histopathological and genetic features of their tumours. It allows for more intensive hepatic screening in ‘high-risk’ metastatic patients with CM leading to earlier detection of their metastases. Such measures have been shown to prolong survival in these patients.1 ,2 Chromosomal and gene expression profile (GEP) analyses of CM are routinely applied to determine their genetic characteristics for patient risk stratification and prognostication purposes.3–6 At the Liverpool Ocular Oncology Centre (LOOC), we routinely perform genetic testing on CM specimens obtained from consented patients undergoing intraocular tumour biopsy, local tumour resection, endoresection or enucleation. Our preferred genetic technique is multiplex ligation dependent probe amplification (MLPA), which detects abnormalities on chromosomes 1p, 3, 6 and 8, reserving microsatellite analysis (MSA) of chromosome 3, only if the DNA yield is <100 ng.3–6 Currently, most patients with CM are treated at the LOOC by eye-sparing techniques, such as ruthenium plaque radiotherapy (Ru106-RT) or proton beam radiotherapy (PBRT), on whom we perform biopsy for genetic analyses. Studies describing intratumoral genetic heterogeneity in CM7–12 have called into question the accuracy of genetic results obtained from such biopsies.

The aim of this study was to compare chromosome 3 aberrations as determined by MLPA or MSA in intraocular biopsy specimens with those of subsequent endoresection/enucleation specimen from the same CM.

Patients and methods

A database query of patients with CM seen at LOOC between 1 September 2007 and 31 December 2014 was performed to identify patients who had undergone an intraocular biopsy for routine, prognostic testing, and subsequently underwent endoresection or enucleation for varying reasons. The aim was to compare chromosome 3 aberrations detected in the biopsy specimen with the ‘paired’ tumour sample obtained from the same patient. Clinical and histopathological data were also collected. The routine processes undertaken for clinical management, sample processing and genetic testing are described below. This study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice Guidelines. The service evaluation was approved by the Royal Liverpool and Broadgreen University Hospital Trust (RLBUHT).

Clinical management

All patients with CM included in this study were new referrals to LOOC. Full ophthalmological examination was performed together with B-scan echography (Ellex Medical, Adelaide, Australia) to measure the largest basal diameter (LBD) and thickness of the tumour. Routine preoperative liver ultrasonography was performed if the LBD >16 mm. Primary CMs were treated with either Ru106-RT, PBRT, endoresection or enucleation, depending on tumour size, location and patient preference. All patients had consented to undergo tumour biopsy for diagnostic and/or prognostic purposes. Biopsies were performed either transretinally or transsclerally, depending on the size and location of tumour, treatment methods and surgeons’ preference. Briefly, transretinal biopsies were performed under local anaesthesia with a 25-gauge vitreous cutter using a three-port sutureless vitrectomy kit, as previously described.13 ,14 Transscleral biopsies were performed either as fine needle aspiration biopsy with a 25-gauge needle attached to a 20 mL syringe by flexible plastic tubing; or as an incisional scleral flap biopsy, where a lamellar scleral flap was created, and tumour samples were obtained with Essen forceps as previously described.15 Four tumours treated with Ru106-RT underwent biopsy immediately before the insertion of plaque; in a single case, the patient underwent biopsy 2 weeks after Ru106-plaque insertion (patient 2). Patients whose CMs were treated with PBRT underwent biopsy on the last day of treatment. During the study period, seven patients with CM had participated in a phase II trial (NITRO trial, ISRCTN35236442) to assess the safety and efficacy of neoadjuvant intravitreal injections of ranibizumab in patients with a large CM. Ranibizumab is not a routine primary treatment for CM. The primary objective of the NITRO trial was to determine response rate of intravitreal ranibizumab, in the neoadjuvant setting, aiming to reduce tumour size, in order to increase the possibility of using eye-preserving treatments as an alternative to enucleation. In these patients, a prognostic intraocular biopsy was performed before the first ranibizumab injection.

Sample processing

Enucleation, endoresection and biopsy specimens were transported immediately without fixative to the Cellular pathology laboratory in RLBUHT. All samples were examined histologically and analysed following standard clinical protocols as described briefly below:

  1. Enucleation: Following transillumination, the freshly enucleated globe was bisected through the tumour. A thin slice of the whole tumour was subsequently taken from the apex to the base for DNA extraction. The remaining material was fixed in 10% neutral buffered formalin, processed, and paraffin wax embedded for histological assessment, as described previously.16

  2. Endoresection: Tumour fragments were extracted with a Pasteur pipette for DNA extraction. The remaining material was fixed in 10% neutral buffered formalin, processed and paraffin wax embedded for histological assessment.

  3. Biopsy: A single cytospin slide was prepared according to standard protocols for May-Grunwald-Giemsa staining, and the tumour cells assessed for their cytological features. DNA was extracted from the remaining sample as described below.

Genetic testing

MLPA analysis of chromosomes 1p, 3, 6 and 8 and MSA analysis of chromosome 3 are the currently validated tests for CM prognostication used within the clinical pathology accredited laboratories at the RLBUHT.

Extracted DNA was examined for chromosome 3 aberrations by MLPA (MRC-Holland, Amsterdam, The Netherlands) if the DNA yield was ≥100 ng, or by MSA, if the DNA yield was less than this concentration. Our methods of DNA extraction, quantification, quality control and MLPA analysis have been described in detail elsewhere.5 Based on the dosage quotients (DQ) generated, classification of the chromosome 3 status of the tumour was made as follows; (1) M3=≥75% of probes on 3p and 3q with a DQ <0.85 and (2) D3=≥75% of the probes on 3p and 3q with a DQ 0.85–1.15. Loss of either the 3p or 3q arm alone was classified by the same criteria and cases in which this was not possible were deemed unclassified.17

The MSA protocol was based on previously published methods.18 Briefly, DNA was extracted from tumour and blood samples from the same patient, using the Qiagen DNeasy blood and tissue kit according to the manufacturer's instructions (Qiagen GmbH, Hilden, Germany). Primer pairs flanking eight microsatellites on chromosome 3 (four on 3p and four on 3q) were optimised for use in two separate multiplex PCR reactions. The PCR products were then analysed using the ABI 3500 genetic analyser and DNA sizing was performed with Genemapper The allele peak height in the tumour DNA was compared with normal DNA obtained from the same patient to determine the allele ratio and hence the presence or absence of loss of heterozygosity in the tumour DNA.


Demographic information, baseline tumour features and histopathological information are summarised in online supplementary table S1. Twenty-eight patients (9 female and 19 male) with a median age of 63.5 years (range 41–89 years) were included. The median follow-up time from the date of primary CM diagnosis at LOOC until the end of study was 23 months (range 4–87 months). The CM had a median LBD of 15.1 mm (range 9.2–20.4 mm) and a median thickness of 6.9 mm (range 2.2–12.7 mm). Histopathological examination of the CM specimens obtained after endoresection or enucleation showed epithelioid cells in 17 tumours; periodic acid-Schiff-positive closed loops in 14; and the mitotic rate >5/40 high power fields (HPF) in eight CMs.

In the context of this patient cohort, primary CM treatment consisted of Ru106-RT (n=5, 18%), PBRT (n=4, 14%), intravitreal ranibizumab injection(s) (n=7, 25%), enucleation (n=11, 39%) and endoresection (n=1, 4%).

Of the 28 intraocular tumour biopsy specimens, 14 yielded sufficient DNA for analysis by MLPA and the remainder were analysed by MSA. One biopsy was taken 2 weeks after Ru106-RT insertion (patient 2; tables 1 and 2) and was classified as disomy 3. Four CMs treated with PBRT underwent biopsy on the last day of treatment (patients 18, 24, 25 and 27; tables 1 and 2): two were classified as disomy 3 CMs, and two as monosomy 3. Of the 28 secondary solid tumour specimens, 27 were analysed by MLPA, and a single enucleation specimen by MSA, due to a low yield of DNA from the small recurrent tumour (table 2). Chromosome 3 analysis showed concordance between the biopsy specimen and the subsequent endoresection/enucleation specimen in 27 cases. One case showed loss of chromosome arm 3p in the biopsy versus a complete loss of chromosome 3 in the subsequent enucleation sample taken 5 months later. In total, nine patients had postradiotherapy genetic analysis; five after Ru106-RT and four after PBRT. Of these, four patients 1, 8, 21 and 23, (tables 1 and 2), had a biopsy prior to Ru106-plaque insertion and a subsequent secondary solid tumour sample was analysed at 9 months, 7 months, 37 months and 79 months, respectively, following radiotherapy. Concordant genetic results for chromosome 3 status of the preradiotherapy biopsy sample and the postradiotherapy tumour were recorded: two were classified as disomy 3 and two as monosomy 3 (table 2).

Table 1

Treatment information

Table 2

Genetic results and survival information

On the basis of the prognostic intraocular tumour biopsy specimen, CMs were classified as having a complete loss of one copy of chromosome 3, ie, monosomy 3 (n=13), partial loss of chromosome 3 (n=3) or the remainder were classified as disomy 3 (n=12). These results were analysed using the Liverpool Uveal Melanoma Prognostication Online model to plan patient management (ie, liver screening).19 During the follow-up period of the study, 2/14 patients with a tumour classified as monosomy 3 died. One patient had developed metastatic CM, and the other, metastatic prostate cancer. Examples of data from two patients are provided in figures 1 and 2.

Figure 1

Patient 5—a 66-year-old man with a large ciliochoroidal melanoma. (A) Haemorrhagic biopsy specimen with epithelioid tumour cell morphology analysed by microsatellite analysis. Representative microsatellite profile showing loss of heterozygosity in the tumour. (B) H&E stained section of the tumour specimen in the enucleated eye showing a mixed cell phenotype. Multiplex ligation dependent probe amplification data showing loss of one copy of chromosome 3. (C) The prognostication curve placed the patient in a high metastatic risk group; unfortunately, he died 6 months later. Chr, chromosome; DQ, dosage quotient.

Figure 2

Patient 10—a 77-year-old man with a large choroidal melanoma. (A) Biopsy specimen with spindle tumour cell morphology analysed by microsatellite analysis. Representative microsatellite profile showing retention of both alleles in the tumour. (B) H&E stained section of the tumour specimen in the enucleated eye showing a spindle cell phenotype. Multiplex ligation dependent probe amplification data showing a normal chromosome 3 copy number. (C) The prognostication curve placed the patient in a low metastatic risk group. Chr, chromosome; DQ, dosage quotient.

In addition to chromosome 3 status, chromosomes 8q and 6p are also correlated with a poor and good prognosis, respectively, for CM.3–6 These data are not currently incorporated in the prognostic model used at LOOC; however, the data are available from all MLPA analyses. Of the 13 biopsy specimens and their corresponding secondary solid tumour specimen analysed by MLPA, concordant results for 8q were demonstrated in all cases; 7 had a gain of 8q and 6 had a normal diploid copy number (table 2). Chromosome 6p classification could be made in 11/13 biopsy specimens analysed by MLPA. Nine of these 11 CMs showed concordant results between the biopsy and the subsequent larger tumour specimen; 5 had a gain of 6p and 4 had a normal diploid copy number. In the remaining two cases, patient 23 demonstrated a gain of 6p in the biopsy specimen and a loss of 6p in the subsequent larger tumour specimen, while patient 28 had a normal diploid copy number for 6p in the biopsy specimen and a gain of 6p in the second larger tumour specimen (table 2).


This study's main finding is that the genetic results for chromosome 3 status from the paired samples showed prognostic concordance in all cases. These findings, which are consistent with data from several other large case series,20 ,21 demonstrate that DNA of sufficient quality and quantity for chromosome 3 analysis, can be extracted from intraocular CM biopsy specimens, and supports the use of MLPA and/or MSA for genetic analysis with no loss of prognostic accuracy compared with DNA obtained from the endoresection specimen or enucleated eyes. Another significant finding is that chromosome 3 results, as assessed by MLPA and MSA, were not affected by prior radiotherapy in this small series.

To our knowledge, this is the first study to compare MLPA or MSA results for chromosome 3 status from CM biopsy samples with subsequent endoresection or enucleation specimens from the same individual. The main weakness is the small patient number, which prevents accurate estimation of the true incidence of sampling errors; however, our results to date indicate that this is likely to be small. Another weakness is the short follow-up time, which does not allow evaluation of the results according to mortality. It would take many years to analyse the results according to survival.

Morphological heterogeneity is well recognised in CM and has led to concerns regarding the accuracy of genetic results for chromosome 3 classifications in these tumours using a biopsy sample taken from a single site. Heterogeneity of chromosome 3 copy number in CM analysed by fluorescence in situ hybridisation has been reported by several groups.7–12 However, it is difficult to draw conclusions from these studies because of: (1) inconsistent cut-off points used to determine monosomy 3; (2) signal reduction caused by melanin pigment; (3) in some studies, the use of only a centromeric probe for chromosome 3 analysis; and (4) the omission of a control probe for aneuploidy and/or truncation artefacts. Naus et al10 compared 40 fine needle aspiration biopsies of CM with the corresponding main tumour sample, reporting two cases with an inconsistency that may have led to a misclassification of chromosome 3 status. Similarly, Schoenfield et al12 reported a lack of concordance between the chromosome 3 status at the tumour apex and base of 3 out of 17 CMs.

Using MLPA, we have previously shown that heterogeneity of gene loci dosage quotients is observed across different regions in a single CM for chromosomes 3, 6 and 8. Consistent with our current study, these data did not alter the overall chromosome 3 classification status, as compared with MLPA analysis of the whole tumour.22 Moreover, 14 CMs, in which two tumour regions were examined that differed phenotypically, demonstrated no intratumour heterogeneity for chromosome 3 classification, when examined with MSA.23 We previously reported one exceptional case where true heterogeneity existed within the tumour, with differing components of the CM displaying distinct morphological and genotypical features, which we believe had evolved over time.24 However, from the evidence in the literature, such CMs appear to be very rare. Concordance of chromosome arm 8q classification between the biopsy and secondary solid tumour specimen also highlights its utility in genetic testing of CM specimens. In our recent analysis of MLPA data from 602 cases, a gain of chromosome 8q was also associated with a poor outcome in patients with CM and worsened the prognosis of patients with monosomy 3.17 Although information for chromosome 8q is not currently incorporated into the prognostic model used in LOOC (,25 our data suggest that this may be of value. In order to investigate 8q further, it will be important to develop an MSA analysis for smaller samples where DNA yield does not permit MLPA. In contrast to chromosomes 3 and 8q, the data for chromosome 6p were less concordant between the matched pairs, although the reason for this is unclear.

The main implications of this study are: (1) when DNA extracted from biopsy samples of CM, taken and processed as described above, are analysed by MLPA or MSA, the results are most likely to be the same as those that would be obtained from analysis of DNA extracted from a cross-section of the entire tumour. In other words, the chances of sampling error missing chromosome 3 loss with a single site biopsy specimen are relatively small. A recently published study of 80 CMs analysed by GEP has recommended two-site biopsy sampling with independent GEP testing of each specimen, in order to lessen prognostic misclassification of metastatic risk with this test.26 It would be of interest to repeat the current study using GEP and determine the concordance of the results between the smaller and larger tumour samples. (2) The second main implication of our study is that genetic testing using MLPA or MSA is feasible in CM samples post Ru106-RT or PBRT with no alteration of chromosome 3 status when compared with a preradiotherapy specimen. These data are not only consistent with our general experience with post Ru106-RT CM biopsies (manuscript in preparation), but also with a recently published study of 15 cases of CM analysed by array comparative genomic hybridisation following Ru106-RT or gamma knife radiosurgery, which demonstrated successful determination of chromosome 3 and 8q status.27 Five of the 15 cases had tumour material collected before and after radiotherapy demonstrating concordant results for chromosomes 3 and 8q, as similarly demonstrated for 4 cases in our study.

In conclusion, our study demonstrates that biopsy of CM yields DNA of sufficient quality and quantity with no change in the chromosome 3 profile that would result in loss of prognostic accuracy compared with DNA extracted from a larger tumour specimen. Moreover, MLPA/MSA can be successfully conducted on postradiotherapy tumours with this, as yet limited number of cases, suggesting that chromosome 3 results are not affected by prior radiotherapy. While these findings are of relevance for predicting behaviour of the primary tumour, a similar analysis addressing genomic heterogeneity in CM metastases must be undertaken: it is only through the detailed characterisation of CM metastases that we will be able to identify druggable targets and improve survival outcomes for patients with CM.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Twitter Follow The Liverpool Ocular Oncology Research Group (LOORG) @icanceresearch

  • Contributors HH, SEC, HK, VH and ST conceptualised the study, analysed the data, formulated data interpretation and revised the manuscript. VH and HK collected the required follow-up clinical data with the help of Gary Cheetham. BED helped to interpret the data, and provided valuable additions and corrections to the manuscript.

  • Competing interests BED is an unpaid advisor for Impact Genetics.

  • Patient consent Obtained.

  • Ethics approval NRES 11/NW/0249.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost Jonas