Detection of Nonrandom Chromosomal Changes in Multiple Myeloma by Comparative Genomic Hybridization

To the Editor:

Recently, Cigudosa et al1 used comparative genomic hybridization (CGH) for the analysis of patients with either multiple myeloma (MM) or monoclonal gammopathy of undetermined significance (MGUS). Whereas cytogenetics demonstrated clonal abnormalities in 43% of patients, CGH identified chromosomal imbalances in 70% of cases.

In MM, cytogenetic studies are hampered by the low proliferative index of malignant plasma cells. A technique enabling identification of chromosomal abnormalities without requiring obtention of metaphases within the malignant clone would be of major interest for the analysis of plasma cell malignancies. CGH fulfills some of these theorical criteria. Using this technique, it is possible to identify chromosomal imbalances throughout the whole genome, even on frozen or fixed material, independently of the mitotic index of malignant cells. However, CGH presents some limitations, the most important being that it does not allow identification of balanced chromosomal abnormalities, that it requires at least 50% of tumor cells in the sample, and that its resolution to detect a deletion is at best 10 Mb. We and others have demonstrated that this technique may be useful in hematologic malignancies, including MM.2-7 

Taking into account these limitations, the study of Cigudosa et al1 deserves some comments. First of all, 14 of the 29 patients had less than 50% plasma cells in the sample and then are theoritically not analyzable with this technique.8Moreover, in 8 of these 14 patients with less than 50% malignant plasma cells, Cigudosa et al1 found abnormalities with CGH. For these cases, comparison of the CGH results with cytogenetics showed major discrepancies. Unfortunately, these investigators did not perform any fluorescence in situ hybridization (FISH) experiment to validate the CGH results. Furthermore, they emphasized that trisomy 19 could be a primary event in MM, because it was the sole detected abnormality in 2 patients. These results have to be tempered, because many previously reported studies have shown that false-positive abnormalities are often found with this chromosome8,9 (possibly related to the high density of small, interspersed repetitive elements present in this chromosome). Finally, these investigators interpreted normal CGH profiles as corresponding to tumors with deletions smaller than 2 kb and thus undetectable by CGH. Many studies previously analyzed CGH resolution and demonstrated that the threshold for deletion detection is close to 10 Mb.10 

In conclusion, CGH is probably an important tool for the analysis of chromosomal imbalances in MM. However, its usefulness in MM will probably be limited because of the frequent low plasma cell infiltrate in patients, preventing the detection of abnormalities. An alternative strategy could be to include a first plasma cell purification step. Nevertheless, in patients with infiltrates close to 50%, abnormalities should be validated by other techniques such as interphase FISH.

In their letter to the editor, Drs Avet-Loiseau and Bataille make several comments about our report cited above. The limitations of the CGH technique that they point out, namely its inability to detect structural rearrangements and the need for a reasonable number of tumor cells in the specimen analyzed, are well known and well recognized, as documented by them in the several publications that they cite. In any case, we have only reported CGH abnormalities in this series, indicating that we have indeed sampled the tumor cells. With regard to their concern about chromosome 19 being a false-positive, as noted in our report, we have performed the appropriate normal × normal hybridizations and did not see any positive results. In addition, trisomy 19 has previously been reported as a recurring change in conventional cytogenetic studies of multiple myeloma.1-1-1-3Our statement about the power of resolution of CGH to detect deletions as 2 kb is an unfortunate typographical error. It should have said 2 Mb.1-4 We regret that we did not pick this up in the page proofs.