Chromosomal copy number (CN) is an important prognostic index in childhood acute lymphoblastic leukaemia (ALL). High hyperdiploidy (51–65 chromosomes) is associated with a good prognosis and near haploidy (23–29) a poor outcome. Conventional cytogenetics produces accurate estimates of CN in the majority of cases but sometimes cytogenetic fails or the morphology of chromosomes in ALL is too poor for accurate identification. We have used single nucleotide polymorphism arrays (SNPA) to determine CN in 86 childhood ALL patients and have compared the results with conventional cytogenetic analysis. Comparison of hybridisation intensity to a reference set of 110 normal individuals was used to derive CN estimates on 34 samples obtained during remission from ALL. Twelve were analysed using GeneChip® Human Mapping 10K Array Xba 131 and 22 using Xba 142 version 2.0 arrays (Affymetrix). Gene copy number estimates for individual SNPs were averaged across the autosomes and X chromosome. Results confirmed that SNPA gave accurate estimates of CN in samples of normal cells. Analysis by SNPA was successful in all 86 presentation leukaemic samples. The results of conventional cytogenetic analysis was available in 75, 11 having a failed cytogenetic result. Estimates of CN were identical in 34 and within 1 chromosome in a further 28. High hyperdiploidy was observed in 28 cases. In 23 of these the cytogenetic and SNPA results were concordant, while in the remaining five cases with a normal (n=2) or a failed (n=3) cytogenetic result, there was evidence from interphase FISH analysis of a hidden, non-dividing high hyperdiploid clone. One patient had a SNPA result indicative of near-haploidy (23–29 chromosomes) and another low hypodiploidy (33–39 chromosomes) with the SNPA confirming the appropriate chromosomal losses. There was marked discordance in one case of near tetraploidy (a total underestimate of 33 chromosomes) but this was not unexpected as the four copies of most chromosomes comprised no allelic imbalance. Analysis of concordance for individual chromosomes showed that, of the 1679 chromosomes compared, there was concordance of copy number in 1517 (90%). The lowest concordance was for the X chromosome (79%) and chromosome 21 (77%). This level of discordance, at least for chromosome 21, may be explained by multiple copies with no allelic imbalance. For all other chromosomes the concordance level was above 80%. We have shown that SNPA can reliably estimate CN in high hyperdiploidy and near haploid cases. This is of great potential value for genome-wide screening in cases in which cytogenetics is not available. It also provides validation of the applicability of the technique for the analysis of DNA from the majority of tumour types which do not yield metaphases.

Disclosure: No relevant conflicts of interest to declare.

Author notes

*

Corresponding author

Sign in via your Institution