ORIGINAL ARTICLE Unexpected findings in cancer predisposition genes detected by array comparative genomic hybridisation: what are the issues? Gabriella Pichert, 1 Shehla Nilofer Mohammed, 1 Joo Wook Ahn, 2 Caroline Mackie Ogilvie, 2 Louise Izatt 1 ABSTRACT Objective To calculate and discuss the percentage of imbalance for selected cancer predisposition genes in patients referred for routine diagnostic array comparative genomic hybridisation (CGH). Design Audit of findings from application of array CGH for patients referred for developmental delay, behavioural abnormalities and birth defects in 4805 patients referred to Guy’s and St Thomas’ NHS Foundation Trust for cytogenetic investigation from South East London, Kent and East Sussex and other genetic centres across the UK. Results 29 of 4805 (0.6%) patients examined by array CGH had genomic imbalance of <5 Mb involving cancer predisposition genes. Six patients were referred for syndromes involving cancer predisposition genes; none of the other 23 patients with array CGH findings in cancer predisposition genes had any symptoms/family history stated on their referral form suggestive for the respective syndrome. Twelve whole gene deletions, two partial deletions, 12 duplications, two partial duplications, and one mosaic duplication were observed. In 17/29 patients (59%), inheritance could not be established, eight imbalances were de novo, and four inherited. Conclusions This new technology raises the possibility of unexpected findings in cancer predisposition genes. Therefore, the possibility of such findings has to be addressed in pre-test and post-test counselling by genetically trained healthcare professionals. As many of these findings have not been described previously, their clinical significance is unknown and patients need long- term follow-up to determine their clinical relevance. This will enable genetic healthcare professionals to advise such people about their cancer risks and appropriate cancer risk management options. BACKGROUND G-banded chromosome analysis has, until recently, been the preferred method of examining the genetic background of developmental delay, learning diffi- culties, behavioural abnormalities and birth defects; this technique has a maximum resolution of approximately 3e5 megabases (Mb). Additionally, more recent tests such as fluorescence in situ hybridisation and multiplex ligation-dependent probe amplification are available to target specific sequences. In contrast, array comparative genomic hybridisation (CGH) can detect regions of imbal- ance as small as 20 kb, although the actual resolu- tion will depend on the array platform used and the region of the genome harbouring the imbalance. Therefore, CGH arrays, while detecting changes causative of the patient’s presenting phenotype with a much higher sensitivity than conventional cytogenetic methods, may also identify unexpected, clinically significant, changes in genes associated with late onset or cancer predisposition conditions. We have described the introduction of array CGH in our clinical genetics service as a sensitive first-line diagnostic test for patients with suspected genomic imbalance. 1 Here, we report the detection of copy number variations (CNVs) involving known cancer predisposing genes in 29 of 4805 patients tested as part of our diagnostic service and discuss the issues raised by these findings. DESIGN, PARTICIPANTS AND SETTING Four thousand eight hundred and five patients were tested by Agilent oligonucleotide arrays 44 K plat- form (designs 014950 and 017457) with a total imbalance detection rate of 24%. Testing was carried out at a regional cytogenetics CPA accredited labo- ratory, where samples were referred from regional paediatricians and other health specialists, as well as from genetics centres both in and outside the region. All CGH arrays were initiated to determine the causes of developmental delay, neurocognitive disability, learning difficulties, behavioural abnor- malities or birth defects or to confirm a clinical diagnosis of a suspected syndrome. Forty-seven cancer genes were selected where there is published evidence 2 linking germline muta- tions in dominantly inherited genes (both tumour suppressor genes and oncogenes) to oncogenesis or disorders predisposing to childhood neoplasia (see table 1). Autosomal recessive cancer predispo- sition genes and cancer genes where only somatic mutations have been identified were excluded from the list. The array CGH database was interrogated to identify patients with imbalance involving these genes. Abnormalities with imbalance size >5 Mb were excluded from the dataset. Although imbal- ances of this size, which include cancer predisposi- tion genes, were detected such as in trisomy 21, these were not included in our analysis as children with such significant imbalances are already closely monitored. Furthermore, as such imbalances are very unlikely to have been inherited, they will not predispose other family members to any significant cancer risks. Similarly, in our molecular genetics 1 Department of Clinical Genetics, Guy’s and St Thomas’ NHS Foundation Trust, London, UK 2 Department of Cytogenetics, Guy’s and St Thomas’ NHS Foundation Trust, London, UK Correspondence to Dr Gabriella Pichert, Clinical Genetics Department, Guy’s and St Thomas’ NHS Foundation Trust, London SE1 9RT, UK; gpichert@bluewin.ch Received 24 November 2010 Revised 12 January 2011 Accepted 24 January 2011 Published Online First 23 March 2011 J Med Genet 2011;48:535e539. doi:10.1136/jmg.2010.087593 535 Cancer genetics group.bmj.com on October 10, 2014 - Published by jmg.bmj.com Downloaded from