Comparative genomic hybridization (CGH) was used for the examination of ten cases of myeloid leukemia (eight acute myeloid leukemias and two myelodysplastic syndromes). In five cases, genomic gains or losses were identified, which mapped to chromosomal regions known to be involved in this group of malignancies. In comparison to the results obtained by banding analysis, discrepancies were found in three of the ten cases; in two cases, chromosomal imbalances were not identified by CGH because they were present only in small subclones. In the other case, there were no evaluable metaphase cells for banding analysis; CGH revealed an overrepresentation of chromosome 8, which was confirmed by interphase cytogenetics with a chromosome 8-specific alphoid probe. All abnormalities revealed by CGH were confirmed by G-banding or subsequent interphase cytogenetic analysis, which demonstrates the high specificity of the method. Furthermore, in all cases, CGH identified the chromosomal imbalances present in the major clone as detected by banding analysis. The good correlation between CGH and chromosome banding results in myeloid leukemias makes this tumor a good model for the assessment of tools that are developed for automated and quantitative CGH analysis.

A series of 18 primitive neuroectodermal tumors (PNETs), the most common malignant central nervous system tumors of childhood, were analyzed with the recently developed approach of comparative genomic hybridization (CGH). In five cases, in which only small amounts of DNA were available, universal polymerase chain reaction was successfully applied to generate adequate probe material. In 15 tumors, chromosomal imbalances were elicited, most frequently involving chromosome 17 (loss of 17p and gain of 17q). Further recurrent imbalances included gains of the distal regions of 4p, 5p, 5q, 7q, 8q, and 9p. High-level amplifications were found on 2p24 (one case) and 8q24 (three cases), suggesting involvement of the protooncogenes MYCN and MYC, respectively. In one of these cases, Southern blot analysis could be performed, proving high-copy-number amplification of MYC. Interestingly, none of the three patients with high-copy-number amplifications of MYC responded to therapy.

No abstract