Critical Importance of Conventional Cytogenetics in Detecting Prognostically Significant Chromosome 13 Deletions in Myeloma.

Abstract

Among chromosomal abnormalities described in myeloma (MM), deletion of chromosome 13 (Δ13) has been linked to a poor outcome. However, prognosis appears to vary according to the detection method used: conventional cytogenetics (CC) or FISH. We have investigated this inconsistency in a large multi-center cytogenetic study. Bone marrow samples from 814 MM patients were provided from 95 UK hospitals. FISH on purified plasma cells (n=697) and/or CC (n=577) was performed centrally: 458 samples had both tests. Sample quality was often poor, with low cell counts (median 2x107 and low plasma cell (PC) content (median 9%), precluding both tests on all samples. FISH for Δ13, t(4;14), t(11;14), gain of chromosomes 5, 9, 15 and, where possible, 3, 7 &17 was carried out. Overall survival was used as the end point with median follow-up of 22mo (range 12–53mo). Abnormalities were detected by CC in 216 cases (37%). Non-hyperdiploidy (non-HRD) (≤47 chromosomes, n=82, 14% of total, 38% abnormal) was a stronger indicator of poor prognosis than hypodiploidy (≤45 chromosomes, n=55, 10% of total, 25% abnormal) (p<0.001 vs p=0.027). Pooled data from all FISH tests allowed us to determine ploidy status: whether hyperdiploid (HRD) or non-HRD with specificity >96% and sensitivity >92%. There was complete agreement between CC and FISH for individual abnormalities. CC showed a highly significant difference in prognosis between abnormal karyotypes with (n=95, 44%) and without Δ13 (median 15 months vs not reached, p<0.001), including non-HRD karyotypes with (n=57, 70%) and without Δ13 (median 13mo vs not reached, p<0.001). Thus the poor prognosis appeared to be entirely due to the presence of Δ13. A similar difference was seen for HRD with (n=38, 28%) and without Δ13 (median 18 mo vs 45mo, p=0.001). FISH showed Δ13 (in >20% PC, n=322, 46%) to have the worst outcome among the individual abnormalities (median 30mo vs not reached, p<0.001 cf 30mo vs 45mo, p=0.032 for t(4;14)) (Fig1A). This adverse prognosis was not altered in those cases with Δ13 in >50% or >80% PC, or by considering only Δ13 in non-HRD cases (median 29mo vs 49mo, p=0.001), consistent with our CC results. When patients with results from both CC and FISH were examined, the outcome for Δ13 by CC was significantly worse than by FISH only (median 15mo vs 33mo, p<0.001). The survival figures were identical for the total data set, despite some patients having had one test only (Fig1B). Comparison of the two survival curves showed that, despite both Δ13 by FISH and CC having p values of <0.001, there was a large difference in actual survival between these two groups (30mo vs 15mo) and multivariate analysis of all genetic markers showed only Δ13 by CC to be an independent variable. This multi-center study has clearly implicated Δ13 seen by CC as an independent indicator of poor prognosis, with Δ13 by FISH probably only informative within a prognostic index, regardless of modifications by ploidy or proportion of abnormal cells. Thus, we recommend CC as the most appropriate prognostic test for myeloma. [Display omitted]