|Home | About | Journals | Submit | Contact Us | Français|
Cytogenetic studies were performed as part of all diagnostic and surveillance bone marrow examinations in 956 newly diagnosed patients with multiple myeloma (MM) receiving total therapy (TT) protocols and in 1085 previously treated patients enrolled in non-TT protocols. In both groups, cytogenetic abnormalities (CA) were present in one-third at baseline and persisted in 14% prior to first and 10% prior to second transplant (TT, 5%; non-TT, 15%); post-transplant detection rates increased progressively with time, from 7% within 6 months to 21% within 24 months to 28% at relapse. According to multivariate analyses, overall survival was adversely affected by the presence of CA at baseline (hazard ratio (HR)=7.20, P<0.001) and the development of CA both prior to (HR=3.28, P<0.001) and after first transplant (HR=6.24, P<0.001), whereas suppression of CA pretransplant was favorable (HR=0.38, P<0.001). The presence of CA at relapse further distinguished patients with a short median post-relapse survival of only 11 versus 47 months in those without CA (P<0.0001). Post-relapse survival was independently adversely affected by the detection of CA both at baseline (HR=1.35, P=0.044) and relapse (HR=2.47, P<0.001). Collectively, these results underscore the importance of monitoring for CA and attest to the favorable prognostic consequences of CA suppression with effective therapy.
The presence of metaphase cytogenetic abnormalities (CA) has long been recognized as an independent adverse feature in multiple myeloma (MM), which is present in one-third of newly diagnosed subjects.1–8 The detection of CA not only reflects high proliferative activity but also stroma-cell independence of in vitro cell division of malignant plasma cells.9 Virtually all cases of MM and even monoclonal gammopathy of undetermined significance are aneuploid when examined by DNA flow cytometry10 or by interphase fluorescence in situ hybridization.11,12 In head-on comparisons of metaphase karyotyping and interphase fluorescence in situ hybridization analyses, the former technique was a stronger predictor of poor outcome,13 so that the perceived shortcomings of its dependence on mitotic transition with lower CA yield actually emerged as an important functional biologic property of MM.
Recently, gene expression profiling analysis of highly purified plasma cells has been shown to identify distinct molecular entities14 and approximately 15% of patients with very high risk of early treatment failure and death.15 In a comprehensive analysis of an entire battery of baseline parameters that included metaphase cytogenetics, interphase fluorescence in situ hybridization, gene expression profiling and MRI examinations, gene expression profiling based high risk emerged as the strongest adverse parameter conferring short event-free survival (EFS) and overall survival (OS) among patients treated on our total therapy 2 (TT2) regimen.16 Moreover, while achieving complete remission status was an independent favorable event in the context of standard prognostic models for both good- and poor-risk MM, complete remission was beneficial only in high-risk MM in the context of the gene expression profiling-defined prognostic model.17 Collectively, the above findings and observations by many other investigators support the notion of an overriding role of genetic abnormalities for patient outcome.
In earlier studies, we had observed that the presence of CA was an adverse variable not only when detected at baseline prior to initiation of therapy, but also when present at relapse for subsequent survival and when detected after initiation of induction therapy.18 With access to a vast body of data in our patient population due to sample submission for cytogenetic analysis with each bone marrow examination, we have investigated the implications for clinical outcome of CA when detected during the entire clinical course.
Two major groups of patients were considered: (1) those with newly diagnosed MM enrolled in successive front-line trials of total therapy 1 (TT1),19–21 total therapy 2 (TT2)22 and total therapy 3 (TT3)23–25 (TT group); and (2) all other patients (non-TT group) receiving at least one cycle of melphalan-based high-dose therapy with autograft at our institution.26 A further requirement was the availability of serial cytogenetic data at the following time points: at baseline prior to protocol enrollment and prior to first transplant. Patients used in the post-relapse survival analysis had to meet the requirements above, and were also required to have cytogenetic data available at relapse.
Metaphase analysis was performed as previously described.27 Briefly, bone marrow was processed for chromosome studies by standard techniques. Cultures were set up in 10 ml of RPMI 1640 media (Sigma, St Louis, MO, USA) supplemented with 20% fetal bovine serum (Sigma) and 1% L-glutamine/Pen-Strep solution (Gibco, Grand Island, NY, USA). Cultures were initiated with 1–3 ml of whole bone marrow. Direct harvest and 24- and 48-h unstimulated cultures were employed on most specimens. The direct harvest procedure included the combination of 1 ml of 0.05% trypsin EDTA (Gibco), 9 ml of hypotonic solution (0.075 M KCl) and 5 μl ml−1 colcemid solution (Irvine Scientific, Santa Ana, CA, USA) for 1 h at 37 °C. The 24- and 48-h cultures received 50 μl colcemid solution for 1 and 2 h, respectively. Twenty cells were studied in each case. An abnormal clone was identified as two or more metaphases with either the same structural abnormality or the same extra chromosome, or at least three cells with the same missing chromosome.
The Kaplan–Meier method was used to estimate OS and EFS;28 group comparisons were made using the log-rank test.29 OS was defined from the date of registration until the date of death from any cause; survivors were censored at the time of last contact. EFS was defined from the date of registration until the date of progression or death, censored at the time of last contact. Univariate and multivariate analyses of prognostic factors were carried out using Cox regression.30 Development of CA and loss of CA were treated as time-dependent variables.
Table 1 summarizes CA frequencies for all patients with cytogenetic studies performed within the specified time periods. Approximately one-third of patients harbored CA at baseline whether treated on TT or non-TT protocols. Subsequently, CA were detected in 14% prior to first transplant and in 10% between first and second transplants (15% in the case of non-TT and 5% in the case of TT protocols; P<0.0001); after the second transplant, the overall detectability of CA increased from 7% within the first 6 months to 13% within 12 months to 18% within 18 months and 21% within 24 months, always at higher frequencies for non-TT versus TT protocols (all P<0.0001).
There were 940 patients who received TT regimens for whom cytogenetic data were available at baseline and prior to first transplant. As is apparent from Figure 1, the 583 patients with absence of CA at both time points enjoyed superior OS (Figure 1a) and EFS (Figure 1b) with medians of 134 and 78 months, respectively, compared to 77 and 42 months among the 228 patients who displayed CA only at baseline but not pretransplant, and compared to 50 and 21 months among the remaining 129 patients in whom CA were present pretransplant whether or not CA were present at baseline. When examined in the context of each TT protocol, the same highly significant differences observed among all TT patients also pertained to those treated with TT2 (Figures 1c and d). In the case of TT1, the absence of CA both prior to therapy and transplant identified patients with superior OS and EFS, whereas similar poor outcomes were observed for the remainder (CA at baseline only and not pretransplant versus pretransplant regardless of baseline status) (Figures 1e and f). For TT3, the only significant difference in both OS and EFS pertained to patients without CA at both time points versus those with CA prior to therapy only; those with CA pretransplant whether CA were present or absent at baseline constituted a small proportion of 7.4% (versus 20.4% in TT1 and 13.9% in TT2), so that differences were not detectable (data not shown).
Among non-TT protocol enrollees, 463 had cytogenetic data recorded at enrollment and also pretransplant. Similar to TT protocols, both OS and EFS were longest among the 291 patients without CA at both time points (60 and 40 months) compared to 108 patients with baseline CA only (39 and 24 months) and to 64 patients with CA pretransplant (15 and 9 months) (Figure 2).
We next performed multivariate analyses to determine the independent prognostic significance of baseline and subsequent CA in the context of other parameters for all patients. The patient population assessed in this analysis contains the subset of patients included in the graphs who also met the additional requirement of the availability of all clinical variables considered. In addition to a strong adverse role of CA at baseline (OS: hazard ratio (HR)=7.20, P<0.001; EFS: HR=3.82, P<0.001), development of CA prior to first transplant affected both clinical outcomes negatively (OS: HR=3.28, P<0.001; EFS: HR=2.40, P<0.001) (Table 2). Loss of CA prior to first transplant favored better outcomes (OS: HR=0.38, P<0.001; EFS: HR=0.46, P<0.001). Acquisition of CA after first transplant was a grave parameter for both OS (HR=6.24, P<0.001) and EFS (HR=3.52, P<0.001). Additional significant parameters included pretransplant β-2-microglobulin levels ≥3.5 mg l−1, hemoglobin <10 g per 100 ml and immunoglobulin-A isotype (see Table 2). Patients receiving front-line therapy on TT protocols had superior OS (HR=0.57, P<0.001) and EFS (HR=0.65, P<0.001).
The median post-relapse survival was superior at 47 months among the 525 patients without CA at relapse compared to 11 months for the 248 patients with CA (data not shown, P<0.0001). When examined in the context of the presence of CA both at baseline and relapse, results revealed best outcomes in patients lacking CA at baseline and at relapse; an intermediate survival was observed when CA was present only at baseline and not at relapse; the poorest outcomes were noted among patients who presented with CA at relapse regardless of baseline CA status (Figure 3). Indeed, as shown in Table 3, significant variables associated with poor survival after relapse included the presence of CA at baseline as well as the development of CA at relapse (which was the sole indicator of loss of disease control in 107 of 366 patients (29%)), along with impaired renal function, whereas TT protocol administration was a favorable feature.
While genetic abnormalities are universal in MM and can be detected by fluorescence in situ hybridization in inter-phase nuclei,8,11 the presence of CA in one-third of patients with MM is a reflection of tumor cells’ inherent proliferative potential as well as their ability to divide in vitro outside the context of their bone marrow environment. Here, we confirm and extend previous observations of the adverse consequences of CA detection not only at baseline but also prior to and after transplant, signifying MM cells’ survival of such therapies. Conversely, effective suppression of CA pretransplant had highly favorable prognostic implications. The worst outcome pertained to patients with detectable CA pretransplant whether or not such property was present pretreatment. Importantly, the adverse consequences of CA extended to their presence at relapse, so that post-relapse survival was collectively affected by the entire CA history.
Examining the prognostic importance of CA at baseline and prior to transplant in the context of individual TT protocols, we noted that patients treated with TT2 shared the pattern observed for all TT and non-TT patients. In the case of TT1, only patients without CA at both time points fared well, whereas those with CA detected at baseline or prior to transplant had similar poor outcomes. In the context of TT3, the presence of CA at baseline imparted poor risk, whereas CA detection pretransplant was too infrequent to affect outcome significantly (7.4%); this constellation was present in 13.9% of patients enrolled in TT2 and was highest among those treated on TT1 (20.4%). The progressively lower detection rate of CA at the end of induction therapy attests to the improved efficacy of this protocol phase with the transition from TT1 to TT2 to TT3 protocols.
Many laboratory characteristics have been shown to impact the clinical course of patients with MM, which include markers of tumor burden (β-2-microglobulin, bone marrow plasmacytosis, hemoglobin) and MM interaction with bone marrow microenvironment (albumin, C-reactive protein) (for review, see Greipp et al.31). However, the detection of CA at any time in the disease or treatment course identified a parameter with ominous consequences even in remission as defined by standard diagnostic criteria, emphasizing the prominent role of MM genetics in prognosis, which has only been superseded by gene array studies.16,17
In today’s era of greater therapeutic options and vastly improved prognosis, we strongly advocate that metaphase karyotyping be performed as part of all bone marrow examinations for initial staging and follow-up of MM, as the effective suppression of an abnormal karyotype should be a major therapeutic objective toward long-term survival.
This work was supported in part by CA55819 from the National Cancer Institute, Bethesda, MD, USA.
Author contributions: BB and JS designed the study; VA and BB wrote the paper; JS performed and supervised cytogenetic analyses; CB and JG collected data into myeloma database; VA, FvR, EA and BB treated patients on protocols and AH and JC performed statistical analyses.