genotype fluorescence in situ hybridization

Cancer Genetics and Cytogenetics 158 (2005) 99–109

Lead article

Targeting plasma cells improves detection of cytogenetic aberrations in multiple myeloma: phenotype/genotype fluorescence in situ hybridization Marilyn L. Slovaka,*, Victoria Bedella, Kristen Pagela, Karen L. Changa, David Smithb, George Somloc Divisions of aPathology, bBiostatistics, and cHematology/Bone Marrow Transplantation, City of Hope National Medical Center, 1500 East Duarte Road, Duarte, CA 91010 Received 15 November 2004; received in revised form 5 January 2005; accepted 10 January 2005

Abstract

Standard fluorescence in situ hybridization (FISH) easily detects nonrandom karyotypic abnormalities in multiple myeloma (MM) at disease presentation, when tumor burden is high. In contrast, the detection of residual MM using the standard 200 unselected nonmitotic nuclei FISH approach correlates poorly with residual disease detected by morphology, flow cytometry, immunohistochemistry, or reverse-transcription polymerase chain reaction (RT-PCR). We have used sequential MayGru¨nwald Giemsa stain to identify plasma cell populations, followed by FISH analyses (target FISH or T-FISH) to detect immunoglobulin heavy-chain gene (IGH) rearrangements, 13q or 17p deletions, or hyperdiploidy. In this study, 115 samples were collected from 100 patients with MM regardless of treatment status. In this proof-of-principle prospective study, T-FISH detected MM in 52 samples (45%), a percentage similar to that obtained by pathology. Disease detection increased from 5.6% with standard FISH to 48% with T-FISH, and cell culture experiments showed that T-FISH consistently detected a clonal abnormality at dilutions of 10⫺3. In five patients, T-FISH further identified myelodysplastic-associated karyotypic changes restricted to myeloid cells. Our observations suggest that T-FISH identifies cell lineage involvement of cytogenetic abnormalities, improves detection of low-level or residual MM, and may define the coexistence of hematologic karyotypic changes in individual patients. 쑖 2005 Elsevier Inc. All rights reserved.

1. Introduction Multiple myeloma (MM), a B-cell malignancy characterized by the clonal proliferation of plasma cells in the bone marrow, has been associated with unique clinicopathologic features, genetic abnormalities, and response to therapy [1– 3]. Because terminally differentiated plasma cells usually have a low proliferative capacity, informative conventional cytogenetics are reported in only one third of newly diagnosed cases, mostly in cases with an exceptionally high plasma cell proliferative rate [4]. In contrast to conventional cytogenetics, standard fluorescence in situ hybridization (FISH) analysis and flow cytometry have identified aneuploidy in the majority of newly diagnosed MM patients [5,6].

* Corresponding author. Tel.: (626) 256-4673, ext. 62313; fax: (626) 301-8877. E-mail address: [email protected] (M.L. Slovak). 0165-4608/05/$ – see front matter 쑖 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.cancergencyto.2005.01.006

Despite the disappointing frequency of conventional cytogenetic aberrations, -13/del(13q), the 14q32/immunoglobulin heavy chain gene (IGH), and del(17p) have diagnostic and prognostic significance in MM. Combined cytogenetic and FISH studies have reported 14q32/IGH rearrangements with multiple partner chromosomes in 60–90% of cases, establishing IGH rearrangements as a genetic hallmark of MM and an early pathogenic event, especially in nonhyperdiploid MM [2,7,8]. Losses or deletions of chromosome band 13q14.3, with a reported frequency of 30–50% in MM, are associated with specific clinicopathologic features, including a higher frequency of lambda-type MM, high plasma cell labeling index, female predominance, and inferior survival after standard chemotherapy [3,4,9–12] and doseintensive therapy [13]. Although TP53/17p13 deletions are less frequent (reported range, 9–33%), they are powerful independent predictors of shortened survival in MM and are associated with stage III disease, clonal evolution of disease, drug resistance, and genetic instability [3,14–16].

100

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109

Alternatively, hyperdiploid MM, characterized by a count of 48–74 chromosomes, has less than 30% IGH gene rearrangements, less frequent chromosome 13 deletions or other structural chromosomal rearrangements, and a more favorable response to therapy [8,17,18]. Taken together, cytogenetic aberrations are prognostically relevant in MM and should be incorporated into clinical trials to facilitate therapeutic trial comparisons and to better define prognosis. Monitoring the clinical course of a MM patient for residual disease by standard cytogenetics and FISH assays for specific karyotypic aberrations, however, is inferior to the pathology-based assays of morphologic examination, immunohistochemistry, and flow cytometry. Moreover, when a patient is referred to a tertiary setting after treatment, defining cytogenetic aberrations becomes a great challenge. Falsenegative chromosomal results are common, due in part to the low proliferative capacity of plasma cells in a setting of spontaneously mitotic myeloid elements, resulting in a “karyotypically normal” bone marrow study. Furthermore, previously treated patients may have less than 10% plasma cells in their bone marrow aspirates, and involvement is usually focal, compromising the detection of residual disease in a study of unselected interphase nuclei by FISH assay. Several limited acute leukemia or MM studies have reported the advantage of a combined May-Gru¨nwald Giemsa (MGG) phenotype/FISH genotype approach to detect a specific population [19–22]. In this prospective study, we show the feasibility and clinical utility of a targeted FISH strategy using a sequential MGG/FISH technique to detect MM and enable direct comparison of cytogenetic status (genotype) with morphologically identified plasma cells (phenotype) in 100 patients. These results were correlated with conventional cytogenetics (CC), pathology, and standard FISH analyses. Our results underscore the importance of phenotype/genotype correlation to detect MM throughout the clinical course of the disease.

2. Materials and methods 2.1. Patients Between August 2003 and March 2004, 115 bone marrow or stem cell aliquot samples were collected from 100 consecutive patients with the working diagnosis of MM, regardless of treatment status. Patient characteristics are outlined in Table 1. The age range was 32–84 years, with a median age of 57 years and male predominance (60%). Approval was obtained from the City of Hope Institutional Review Board, and the experiments were performed according to the guidelines for research with human subjects. 2.2. Pathologic evaluation The pathologic evaluation consisted of morphological examination, flow cytometry, immunohistochemistry, and

Table 1 Patient and disease characteristics Characteristics Age (y), Median Range Sex Treatment (samples ⫽ 115)

Pathologya

Cytogeneticsb

No. of patients Percentage 57 100 32–84 Female 40 Male 60 Chemotherapy only 46 Stem cell transplant 54 MGUS/no previous 15 treatment or newly diagnosed Abnormal 50 Normal 60 No assessment 5 Normal 90 Abnormal 15 Not done 10

100 40 60 40 47 13

43.5 52.2 4.3 78 13 9

a

Pathology assessment included morphologic examination, immunohistochemistry, flow cytometry, and IGH RT-PCR data. b Cytogenetics with any clonal abnormality: hyperdiploid (9); hypodiploid(1); del(13q)(2), del(20q)(2); dup(1q)(1).

IgH reverse-transcription polymerase chain reaction (RTPCR). Morphologic criteria for the diagnosis of MM included the following: (1) marrow plasmacytosis (⬎30% plasma cells) at the time of diagnosis, (2) 10% or more plasma cells in the marrow after treatment, or (3) 5% or more plasma cells plus immunophenotypic evidence of light chain restriction after treatment. A pathology assessment was possible for 110 bone marrow samples but this information was not available for 5 specimens (two bone marrows and three stem cell aliquots). Immunohistochemistry was available for 33 cases, and 24 samples had RT-PCR results. 2.3. Conventional cytogenetics CC results were available for 105 (91.3%) samples. The samples were processed using three-cell culturing methods: 72-hour unstimulated culture, 72-hour culture supplemented with interleukin-4, or 72-hour culture supplemented with interleukin-6 and granulocyte-macrophage colony stimulating factor (GM-CSF). A minimum of 20 metaphase cells was analyzed per sample. Clonal abnormalities followed standard definitions, and chromosome abnormalities were described according to the International System for Human Cytogenetic Nomenclature (1995) [23]. 2.4. FISH analysis The experimental design for this two-phase molecular cytogenetics study is outlined in Fig. 1. In phase I (54 samples), both standard FISH and T-FISH analyses were performed on fresh bone marrow or stem cell aliquots according to standard instructions using DNA probes to detect IGH rearrangements (locus specific identifier [LSI] IGH dual-color break apart), del(13q)/⫺13 (LSI D13S319 13q14.3/13q34 LAMP1; Vysis, Downers Grove, IL), and del(17p)/TP53 deletions (indirect labeled) BAC 199f11

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109

Fig. 1. Experimental design of the two-phase prospective T-FISH feasibility study. To determine if a phenotype/genotype strategy using sequential MGG morphology/FISH to target plasma cells would improve the detection of MM in comparison to standard “area” FISH studies, a two-phase study with an interim analysis was performed. In the first 54 samples (phase I) using a three- probe DNA FISH panel, standard FISH detected residual MM in three cases (5.6%), whereas T-FISH analysis of the same samples showed genomic aberrations in 14–100% of plasma cells in 26 cases, increasing disease detection from 5.6 to 48%. To increase disease detection in phase 2 of the study, the standard FISH arm was discontinued and a fourth hybridization to detect hyperdiploidy was added to the FISH panel. A total of 115 samples from 100 patients were evaluated.

(TP53)/BAC 62n23 (ERBB2) (obtained from Dr. Pieter J. deJong, Children’s Hospital, Oakland Research Institute). The test samples were not enriched for plasma cells. The normal pattern for the IGH (CH/VH) break-apart probe was two red/green paired signals, whereas the presence of a red and green signal separated by at least three signal lengths was considered positive (Fig. 2, A–D). Deletions were detected by the loss of the 13q14.3/D13S319 or 17p13.1/TP53 probe signals (red), with retention of the reference probe (green) for the respective chromosome (Fig. 2, E and F). Chromosome loss was defined as the loss of both the test probe and the reference probe, with retention of both signals for the second chromosome homolog. In each case, a minimum of 200 unselected white blood cell nuclei was evaluated for standard FISH. Four samples that were TFISH negative yet positive for persistent disease in the interim study were hybridized with a probe to detect hyperdiploidy of ⫹5, ⫹9, and/or ⫹15 (Vysis). For the remaining 61 samples (phase II), the hyperdiploid hybridization was performed prospectively when the original three-probe strategy failed to yield a positive T-FISH result. 2.5. Plasma cell T-FISH Cytospin slides (four to six slides per patient) were made using a ThermoShandon cytocentrifuge (Cheshire, England) and 200 µL of the same bone marrow or stem cell aliquot used for standard FISH or CC. The cytospin slides were

101

stained with MGG for morphologic classification by brightfield scanning (Duet Image Analyzer; Bioview Ltd., Rehovot, Israel), a system equipped with an Axioplan 2 imaging microscope (Carl Zeiss, Gottingen, Germany) which allows rapid scanning of cells by morphology and cytogenetics [24]. At least three slides were scanned per patient (mean, 6753 white blood cells per slide; range, 855–22,828); however, only the plasma cells were captured and mapped for FISH analysis (range, 2–167 cells; median, 37 cells). The slides were then destained and used for one but no more than two FISH hybridizations without increased background fluorescence. Two scorers evaluated the results independently. Z-stacking and interactive capture of the plasma cell image allowed for exact signal location and pattern. If the scorers disagreed, the cell was re-reviewed by a third scorer. In each experiment, normal cells were run in parallel to check the efficiency of the target procedure. A 20q12 probe (LSI D20S108) was used to detect 20q deletions, and the LSI 1p36/ LSI 1q25 dual-color probe set (Vysis) was used to detect 1q25 duplication. Validation of the BioView Duet was performed by comparing manual FISH scoring with the BioView Duet assessment for 20 samples for each of the 4 probe strategies described using exact binomial confidence intervals with 0.001 alpha levels. With an average number of 240 cells detected per sample, each probe in BioView was within 3 standard deviations (SD) of the cutoff specified for manual scoring. If target scan results were atypical, an area scan was performed to verify that the abnormality was actually in the plasma cells. 2.6. Sensitivity assay To determine the sensitivity of BioView Duet, serial dilutions of two cell lines, lambda-positive U266 myeloma cell line with a t(11;14) (American Type Culture Collection, Rockville, MD) and the lambda-negative Kasumi t(8;21) cell line, were made and cytospin slides were prepared in duplicate. The slides were then treated with EnVision⫹ System-Horseradish Peroxidase (HRP) 3-amino-9-ethylcarbazole (AEC) for use with mouse primary antibodies (Dako, Carpinteria, CA) to detect lambda-positive cells. The BioView Duet scanned the slides to select lambda-positive cells. The sensitivity of the assay was confirmed by verifying the target cell genotype by FISH using the IGH break-apart 14q32 probe (Vysis). Positive and negative controls were run with the dilutions of U266 cells in Kasumi-1 at dilutions of 10⫺2, 10⫺3, and 10⫺4. 2.7. Biostatistics For standard FISH studies, cutoff limits were determined independently for each FISH probe considered. Falsepositive signals were estimated by evaluating 20 known negative samples (200 nuclei per sample) and constructing 99.9% exact binomial confidence intervals around the error rate intervals using StatXact version 4.01 (Cytel Software

102

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109 Fig. 2. MGG/FISH phenotype/genotype correlations in multiple myeloma cases. (A) Plasma cell stained with MGG shows an (B) IGH rearrangement using a break-apart DNA FISH probe that signals a translocation by a split red/green signal. (C) A normal lymphocyte is negative for an IGH gene rearrangement (D), as shown by the presence of two paired red/green signals using a break-apart probe strategy. (E) Plasma cell stained with MGG shows a 17p/TP53 deletion (F) using a TP53/17p13.1-ERBB2/ 17q21.1 probe set. Loss of the TP53 (red) signal with retention of two HER2 (green) signals indicates a 17p deletion. (G) Plasma cell stained with MGG shows (H) hyperdiploidy with three copies (trisomy) of chromosomes 5, 9, and 15, revealed by a cocktail of DNA probes for these chromosomes (Vysis). (I–L) Case 52 showed an IGH gene rearrangement in the plasma cells (I and J) and a del(20q) in the myeloid cells (K and L). Case 40 showed a duplication of the long arm of chromosome 1 by CC (M–P). Using a 1p36/1q25 probe set, the duplication of 1q25 (arrow shows two green doublet signals) was observed in myeloid cells (neutrophil in M and N) and not in the plasma cells (O and P). Case 12 showed a del(13q) by CC; however, the del(13q) was detected in myeloid cells only using the 13q14.3/D13S319–13q34/LAMP probe combination (Q–T). 䉳

Corp., Cambridge, MA). The upper 99.9% confidence bound corresponds to 3SD above the error rate. For each probe, the upper confidence bounds for the error rate were strictly less than our cutoffs. A percentage of 2% or less cells was considered to be within background limits for the IGH breakapart probe (mean ⫹3SD ⫽ 0.97): 13q deletions (mean ⫹3SD ⫽ 0.34), hyperdiploidy (mean ⫹3SD ⫽ 1.29), and 17p13.1/TP53 deletions (mean ⫹3SD ⫽ 0.77). The decision rule for false positives using the BioView Duet was based on the assumption that detection probabilities of abnormalities in a given cell could be modeled by independent Bernoulli trials with a global probability. We estimated empirically that we could detect abnormalities at a sensitivity of 2% and a more conservative estimate of 5% or greater. Given these detection thresholds and a given sample size of cells, the probability of detecting a certain number of abnormal cells follows a binomial distribution with parameter p equal to 2 or 5%. To test the statistical hypothesis that we observed our BioView results by random chance, we compared the cumulative probability given by the binomial distribution for the observed number of abnormal cells versus a nominal one-sided type I error level of 0.05. In every case we observed, it made no difference in our conclusions whether we used an empirical assay sensitivity of 2% or a worst-case sensitivity of 5%. For this feasibility study, at least two targeted plasma cells had to show the identical aberrant FISH pattern to be classified as positive. We considered the case in which no abnormal cells were observed in a sample. Assuming that the detection of abnormal cells can be modeled as independent Bernoulli trials, the probability of observing no abnormal cells in a sample is the assay sensitivity raised to the power of the sample size. As before, we assumed an empirical assay sensitivity of 98% and a worst-case sensitivity of 95%. If the observed probability was less than 0.05, we concluded that abnormal cells may have been present in the sample but not detected.

103

3. Results One hundred fifteen samples from 100 patients with a diagnosis of MM or monoclonal gammopathy of undetermined significance (MGUS) were cumulatively studied in a two-phase analysis to determine the feasibility of using T-FISH to detect treated or untreated MM. Samples were collected from patients receiving diverse therapies and who were at various stages in their treatment courses (Table 1); 15 patients had 2 samples submitted. At the time of study, 46 samples were collected from patients after receiving chemotherapy only, 54 samples were collected from patients who had received a stem cell transplant, and 15 samples were collected from newly diagnosed patients with a diagnosis of either MM, “smoldering” MM, or MGUS and had received no treatment. An interim analysis of the first 54 samples was performed to compare the sensitivity of standard FISH and T-FISH with no previous treatment, after chemotherapy, and after hematopoietic stem cell transplant (Fig. 3). Standard FISH detected low-level residual disease in 3/54 (5.6%) samples with IGH rearrangements in 3–4% of cells scored, a finding just above our established background limits. Using a 200-µL aliquot of the same sample used for standard FISH, T-FISH showed an IGH gene rearrangement in 14–100% of plasma cells (mean, 77%) in 26 samples. The ability to directly correlate genotype with phenotype by targeting the appropriate cells regardless of treatment status clearly increased sensitivity. For example, standard FISH detected 3.3% IGH-positive cells, whereas T-FISH detected an IGH gene rearrangement in 13 of 15 plasma cells (87%) from the same sample. Overall, when T-FISH results were compared to standard FISH of 200 unselected interphase nuclei in the interim analysis, disease detection increased

Fig. 3. Interim analysis comparing MM detection by standard “area” FISH versus T-FISH. An interim analysis of the first 54 samples compared the sensitivity of standard FISH and T-FISH in the absence of previous treatment, after chemotherapy, and after hematopoietic stem cell transplant. Standard FISH detected low-level residual disease in 3/54 (5.6%). Using a 200-µL aliquot per slide of the same sample used for standard FISH, T-FISH showed aberrations in 26/54 (48%) samples. Based on the low detection rate of standard FISH at the interim analysis, only T-FISH was performed for the remaining 61 test samples.

104

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109

from 5.6 to 48%. Based on the low detection rate of standard FISH at the interim analysis, only T-FISH was performed for the remaining 61 test samples. Furthermore, the interim analysis showed an 8% false-negative rate (4 samples) in the initial 54 samples using a three– DNA probe FISH panel to detect -13/del(13q), IGH rearrangements, and del(17p). One of the false-negative cases had a karyotypic designation of 55-57,X,-X,⫹2,⫹3,⫹5, ⫹6,⫹7,del(8)(p21p23),⫹der(9)t(1;9)(q21;q34),⫹11,-13,i(13) (q10),⫹15,⫹15,⫹19,⫹19,⫹21. This case did not show a clonal aberration that could be detected by the original three– DNA probe panel. The four false-negative cases were rehybridized and found to be hyperdiploid using a three-probe FISH cocktail (Fig. 2, G and H). To increase disease detection in phase 2 of the study, a fourth hybridization to detect hyperdiploidy was added to the FISH panel. Using the entire data set of 115 samples from 100 patients, T-FISH detected at least one genetic aberration in 52 samples (45%; Table 2). Of these, an IGH gene rearrangement was detected in 37 of 52 abnormal studies (71%; Fig. 4). IGH rearrangements were observed as the sole FISH abnormality in 15 (40.5%) of the IGH-positive samples, and 16 of 37 (43%) samples showed both IGH and -13/del(13q), three (8.1%) samples showed IGH/-13/del(13q)/del(17p)/ TP53 deletions, and two cases showed ⫹17 and IGH rearrangements. One case showed an IGH rearrangement with hyperdiploidy. Chromosome 13 aberrations were observed in 23/52 (44%) samples showing morphologic evidence of MM. Two samples (8.7%) showed -13/del(13q) with hyperdiploidy. Deletion of 17p was detected in 5/52 (9.6%), always in association with other aberrations. Four cases showed gain of chromosome 17 and were placed in the hyperdiploidy subset. Hyperdiploidy, detected by gains of chromosomes 5, 9, 15, or 17, was detected in 14/52 (26.9%) samples. Three samples positive by T-FISH were considered nonevaluable by pathology because the samples were insufficient (hemodilute) and a biopsy was not available for evaluation.

Table 2 Comparison of disease detection assays for all 115 samples from 100 patients Multiple myeloma disease MM detected Negative for MM Total Percentage

Pathology Conventional assessmenta Cytogeneticsb

Three-probe standard FISH

Four-probe T-FISH

50 60

12 90

3 51

52 63

110c 45%

105d 11.4%

54 5.6%

115 45%

a Pathology assessment included morphologic examination, immunohistochemistry, flow cytometry, and IgH RT-PCR data. b Fifteen samples had abnormal cytogenetic results, however, MDS karyotypic aberrations were detected in five patients (two with and three without MM). c Five samples did not have a morphologic review. d Ten samples did not have standard cytogenetics.

Fig. 4. Distribution of chromosomal abnormalities detected by T-FISH. TFISH detected at least one genetic aberration in 52 samples (45%). An IGH gene rearrangement was detected in 37 of 52 abnormal studies (71%), del(13q)/-13 in 23/52 (44%) samples showing morphologic evidence of MM, and del(17p) was detected in 5/52 (9.6%). Hyperdiploidy, detected by gains of chromosomes 5, 9, 15, or 17, was assayed in a subset of samples, the four samples found to be negative by the original 3-probe panel in the first 54 samples that were positive for MM by pathology, and all samples after the interim analysis that were negative by the original threeprobe panel regardless of pathology status. Hyperdiploidy was detected in 14/52 (26.9%) samples.

To determine the sensitivity of the target assay and instrumentation, mixing experiments were performed using serial dilutions of lambda-positive, t(11;14)-positive U266 cells in lambda-negative Kasumi cells at dilutions of 10⫺1 to 10⫺4. The BioView Duet easily detected 1 positive cell in 1,000 normal cells; however, consistent results were not observed in the 10⫺4 diluted samples. This finding is consistent with an average number of 6,753 analyzable cells on a cytospin slide. Sensitivity was also evaluated in patient samples. The number of cells in patient samples was found to be highly variable due to sample variability and specimen processing (i.e., specimens are lysed and washed several times before slide preparation). Nevertheless, T-FISH identified an IGH translocation in case 85 with less than 1% plasma cells per morphologic evaluation that was RT-PCR positive for an IGH gene rearrangement. In case 112, T-FISH detected an IGH translocation in a sample that had normal morphologic examination with 1–2% lambda monoclonal plasma cells detected by flow cytometry. In treated samples, a median of 8 plasma cells was observed among 6,355 (median) white blood cells per cytospin slide (0.1%). Taken together, the sensitivity of this assay is within the 10⫺3 to 10⫺4 range, a finding similar to our ability to detect an IGH rearrangement when 0.1% plasma cells were present in the bone marrow sample. A comparison with the common pathology assays is presented in Table 2. A pathology assessment, which included a combination of morphologic examination, flow cytometry, immunohistochemistry, and IgH RT-PCR analysis, or one of these analyses, was performed for 110 samples (95.7%).

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109

Five samples (three stem cell aliquots and two bone marrow samples) did not have morphologic evaluation. Pathology detected MM in 50 samples (45%) and no evidence of disease in 60 samples (55%). T-FISH was concordant with pathology for the 15 untreated samples; 11 cases were positive for MM and 4 cases were negative. A comparison of the untreated cases with CC revealed only five positive CC cases, all confirmed by T-FISH. These cases showed 18–70% plasma cells (mean, 50%). The remaining six CCnegative, T-FISH–positive cases had 10–30% plasma cells (mean, 20%). FISH results concurred with MM detection observed by pathology, with the exception of three cases. All three cases showed focal disease in the biopsy sample. Our attempts to isolate nuclei from the corresponding paraffin-embedded biopsy material for these three samples were not successful. IgH RT-PCR studies were positive in 7 of 24 samples. Of these seven IgH-positive cases, five cases were found to be positive for an IGH translocation by T-FISH. The two remaining cases were hyperdiploid and did not show a detectable IGH rearrangement by T-FISH. Of the five cases that were positive for residual disease by PCR and T-FISH, three samples were negative for residual disease by morphology and flow cytometry (3/24 ⫽ 12.5%). Only 12 of 105 (91%) samples showed common MM karyotypes by CC. Of these, nine cases were hyperdiploid, one case was hypodiploid, and two cases were pseudodiploid. A smoldering MM case, with loss of the Y chromosome as the sole karyotypic change detected by CC (regarded as “normal” based on patient age), showed monosomy 13 and an IGH rearrangement in 8 and 22 plasma cells, respectively, whereas the standard FISH of 240 untargeted cells was within normal limits. In addition, six samples from five patients showed karyotypic clonal abnormalities commonly associated with myelodysplasia, namely, del(20q) (two cases), del(13q) (two cases), and dup(1q) (one case); however, these aberrations were not detected in plasma cells by T-FISH (Table 3). T-FISH of all white blood cells on concurrent cytospin slides revealed that the myelodysplastic syndrome

Table 3 MDS-associated abnormalities detected in five MM patients MDS-associated Patient MM detected MM detected MDS detected abn detected by ID by T-FISH by pathology by pathology CC or FISHa 1 12 31 40 52

Negative Negative Positive (⫹5) Negative Positive (IGH⫹)

Negative Negative Negative

Negative Negative Negative

del(20q) del(13q) del(13q)

Negative Positive

Negative Negative

dup(1q) del(20q)

a The presence of an MDS-associated cytogenetic aberration was confirmed in myeloid lineage cells by T-FISH. Abbreviation: abn, aberration.

105

(MDS)-associated karyotypic changes were restricted to myeloid cells in all five patients (Fig. 2, I–T). Interestingly, none of the five cases showed any morphologic evidence of myelodysplasia, even though CC detected a del(13q) in 5 of 20 mitotic cells in one case. Two of five patients showed different abnormalities in different cell lineages. Case 52 showed del(20q) in the myeloid cells and an IGH translocation in plasma cells (Fig. 2L). Case 31 showed trisomy 5 in plasma cells and del(13q) in myeloid cells. The remaining three patients showed FISH clonal aberrations in the myeloid cells only: del(20q) in case 1, dup(1q) in case 40 (Fig. 2, M–P) and case 12 (and confirmed at follow-up in case 41; Fig. 2, Q–T).

4. Discussion One objective of this study was to determine if a phenotype/genotype strategy using sequential MGG morphology/ FISH to target plasma cells would improve the detection of MM in comparison to pathology assays or karyotype and standard FISH studies in 100 randomly selected patients. In the first 54 samples (interim study), we used a three- probe DNA FISH panel and compared our FISH results with morphology, immunohistochemistry, CC, and/or PCR, if available. In this interim analysis, standard FISH detected residual MM in three cases (5.6%) at a level of detection (3.3–3.8%) barely above our established background limits (⭓2%), whereas T-FISH of the same samples showed genomic aberrations in 14–100% of plasma cells in 26 cases, increasing disease detection from 5.6 to 48%. However, MM was missed in four cases (8%) using a three-probe FISH panel. Interestingly, one of these four cases showed a hyperdiploid karyotype by CC, underscoring the need to expand the DNA probe set to include an additional hybridization to detect trisomy of common gains in MM. Thus, a DNA probe cocktail to detect ⫹5, ⫹9, and ⫹15 was incorporated for the remaining 61 samples and the 4 discordant cases in the first test group. Using the revised strategy and the complete test population, 52 of 115 samples showed clonal cytogenetic MM changes, despite the fact that 87% of the patients were treated previously. T-FISH detected IGH rearrangements in 71%, chromosome 13 anomalies in 44%, and del(17p) deletions in 9.6% of the aberrant cases, reflecting the clonal-specific frequencies commonly reported in the literature [2,7,25,26]. The number of hyperdiploid cases in this study (26.9%) was slightly lower than the reported frequency (30–50%) and may be underrepresented for the following reasons: (1) The hyperdiploid probe set was only tested in a subset of the 115 samples (the four false-negative samples by FISH in the interim study and those negative by the original threeprobe panel in the remaining 61 samples). (2) Hyperdiploid MM has a more favorable diagnosis and these patients may not be referred to a transplant center until the disease shows signs of progression. (3) The probe panel used in this study

106

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109

was limited to trisomy of chromosomes 5, 9, 15, and 17; however, clonal gains for all 24 chromosomes have been reported in MM [25]. Selection of the DNA probes for this analysis were based on both their current frequency of occurrence and diagnostic/ prognostic impact, with hopes of finding at least one cytogenetic marker to follow the patient’s clinical course. The finding of IGH rearrangements in 71% (37/52) of the aberrant cases in this study, mostly in the nonhyperdiploid cases, supports the supposition that IGH is an early or primary pathogenic event in MM that is highly associated with ploidy status. Smadja et al. [18] and Fonseca et al. [8] have also shown that IGH rearrangements are more common in the nonhyperdiploid cases (⬍48 or ⬎74 chromosomes). Although this study did not reflex the IGH-positive samples to identify the translocation partner chromosomes, it is certainly feasible to do so. Defining the partner chromosomes as well as the number and combination of genetic insults in malignant plasma cells in association with their characteristic gene expression profiles [27] should lead to further refinement of the clinical–pathologic subgroups of this heterogeneous disease. Furthermore, given the clear association of ⫺13/del(13q) with specific translocations in MM, particularly t(4;14)(p16.3;q32) and t(14;16)(q32;q23), such testing would provide evidence whether the poorer prognosis of nonhyperdiploid MM is due to -13/del(13q), the highrisk IGH translocations, a combination of the two, or perhaps to the complexity of the karyotypic alterations. It was also interesting that the three cases that showed -13/del(13q) without an IGH rearrangement were hyperdiploid, raising the question of whether del(13q) is a primary or secondary aberration in hyperdiploid MM, and whether del(13q) retains its poor prognostic influence in hyperdiploid MM. The varying prognostic implications of chromosome 13 abnormalities in MM, when detected by metaphase analysis and interphase FISH, with some reports indicating a poor prognosis [4,9,28,29] and another an intermediate prognosis [8], may be attributed to the lack of our ability to detect multiple cytogenetic aberrations within ploidy test populations. Moreover, as targeted agents aimed at specific genetic aberrations in MM evolve, such as short hairpin RNAs [30] or small molecule inhibitors [31,32] for patients with t(4;14) MM, it becomes essential to develop reliable and reproducible assays to detect and quantify these nonrandom aberrations, not only at disease presentation, but throughout the treatment course to evaluate efficacy. The detection of residual disease in poorly proliferative MM with regenerating bone marrow usually requires plasma cell enrichment techniques such as monotypic light chain staining by immunofluorescent microscopy, CD-138 magnetic micro-beads, or flow cytometry techniques [33–36], adding additional cost to the analysis. Dilution experiments showed that T-FISH could reliably detect a clonal abnormality in the 10⫺3 to 10⫺4 range, based on the distinct morphology of mature plasma cells using only 200 µL of sample, thus eliminating expense, time-consuming cell culturing, and

the need of large sample volumes required for some immunophenotyping techniques. Moreover, our observations of T-FISH indicate the method is relatively straightforward, reliable, and reproducible, improving MM disease detection with clinical implications that may direct future therapy for the following reasons. (1) Genotype–phenotype correlation (plasma cells in the test population and lambda-positive in the dilution experiments) increased assay sensitivity by restricting the analysis to a relatively small but appropriately targeted population in both newly diagnosed and previously treated MM cases. (2) Positive T-FISH results were similar in frequency to the common ancillary pathology detection assays, regardless of treatment status, and were complementary by being more informative than CC in the untreated samples when the percentage of plasma cells was greater than 30%. These results argue for T-FISH as a frontline assay when the diagnosis of MM is uncertain. Additionally, T-FISH confirmed the presence of an IGH translocation in samples that were either low-level (1–2%) positive by flow cytometry with a normal morphologic examination, or positive by IgH RT-PCR yet negative for residual disease by morphology and flow cytometry. (3) T-FISH can provide a means to correlate and quantitate specific genetic prognostic markers with outcome and therapeutic decisions, as shown by the detection of an IGH gene rearrangement in 7 of 10 plasma cells among 7,295 white blood cells in a stem cell aliquot, estimating residual disease near 0.1%, a detection frequency reported by others using a combined immunoglobulin light chain staining/IGH FISH technique [33]. (4) Finally, the assay can be performed in a cancer cytogenetics laboratory with minimal training. We found the time necessary to perform the bright-field scan of the entire slide to identify and capture plasma cells greatly reduced FISH scoring times by narrowing the analysis to the targeted plasma cells. In addition, each slide can be used for two hybridizations, and interactive capture of the plasma cell image eliminates the three-dimensional focal plane effect, permitting assessment of the exact signal location and pattern and decreasing background (false positives). Discordant cases resulted in two situations that could result in underestimation of the level of involvement: (1) plasma cells with atypical morphology and (2) sample variation due to “patchy” bone marrow involvement, estimating a false-negative rate of about 5%. A statistical approach based on the binomial distribution was used to relate analytical sensitivity to the number of observations to assay for mosaicism after treatment. For example, five IGH-positive cells in a total of 10 plasma cells with an analytical sensitivity of 95% indicates a result that is highly unlikely to occur by chance (P ⫽ 0.00003). Thus, the finding of a few positive cells after treatment is likely to be clinically significant but requires confirmation by detection of a simultaneous second cytogenetic aberration in the plasma cells, correlation with other pathology assays, or confirmation in a repeat or follow-up sample. On the other hand, calculating false-negative probability is a bit more

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109

problematic because of the low level of target cells after treatment. The probability of not detecting any abnormal cells at 0.05 necessitates a 60-cell study; however, a MM patient responsive to therapy will rarely show this number of plasma cells in his bone marrow after treatment. Thus, if this degree of significance were warranted, plasma cell enrichment techniques and large volumes would be necessary. Five MM patients showed MDS-associated cytogenetic abnormalities restricted to myeloid cells even though pathology showed no evidence of overt myelodysplasia in any of these patients. Two of these cases, however, showed persistent MM with different genetic aberrations in the plasma cells. Abnormalities of 1q and del(13q) may be observed in a variety of hematologic malignancies, including MM and therapy-related MDS/t-acute myeloid leukemia (AML). Duplication of the 1q arm as the sole anomaly was observed in a patient 10 years after autologous stem cell transplant, raising concerns of either relapsing MM or evolving myelodysplasia. T-FISH confirmed the 1q trisomy in myeloid cells and not in plasma cells. Of the two patients with del(13q) in myeloid cells, one patient was being evaluated for a tandem autologous/allogeneic transplant after completing four cycles of vincristine, doxorubicin (Adriamycin), dexamethasone (VAD) chemotherapy with radiation to the lumbar spine 6 months earlier, and the other elderly patient with del(13q) myeloid cells and trisomy 5 in the plasma cells had not received any previous chemotherapy. The remaining two cases showed del(20q). Del(20q) has been reported to occur in 5% of karyotypically abnormal MM/MGUS cases [25]; however, a recent flow-sorted/FISH study of a single MM case indicated del(20q) cells in CD34⫹CD38⫹, CD34⫹CD38-, CD15⫹, and a small population of CD19⫹ cells, indicating del(20q) is not plasma cell–specific [37]. Our data confirm that the del(20q) is a recurrent aberration in treated MM, and they lend additional support to the idea that these MDS- and AML-related cytogenetic findings may represent multipotent progenitor/stem cell damage, possibly caused by recent cytotoxic therapy. Because many MM patients are referred to hematopoietic cell transplant centers after receiving multiple rounds of chemotherapy, such observations for the possible risk of evolving secondary myelodysplasia must be viewed with caution. Recent advances in tandem autologous/allogeneic transplantation have improved the prognosis of MM patients by increasing complete response rates, event-free survival, and overall survival [26,38]. It remains unclear whether relapse is caused by residual malignant cells surviving in the patient after conditioning regimens or by malignant plasma cells reinfused at the time of autologous transplant. To answer this question, two studies detecting residual monoclonal plasma cells in stem cell aliquots by either FISH [26] or combined morphology and monotypic light chain staining [38] reported shortened relapse-free survival. In this study, residual disease was detected in plasma cells in one stem

107

cell harvest, and five patients showed MDS-associated cytogenetic aberrations, raising additional concerns about the timing of obtaining peripheral blood stem cells from patients who are candidates for autologous transplantation. A balance must be sought between the optimal preparative regimen needed to increase tumor cell kill in circulating plasma cells and ways to minimize damage to hematopoietic progenitors that could lead to therapy-related MDS. In summary, our observations suggest that T-FISH is a sensitive test for the detection of low-level presentation or residual MM disease, identification of cell lineage involvement of cytogenetic abnormalities, quantifying tumor load, and studying the coexistence of hematologic malignancies in patients with MM without the need to culture or purify plasma cells by immunomagnetic beads or flow cytometry. Even though this method may not be as sensitive as quantitative PCR, determining the underlying genomic aberrations and quantifying clonal plasma cells in bone marrow or stem cell harvests provides clinically relevant information regarding disease status and may direct future therapy. Application of the proposed FISH technique in prospective phase III clinical trials at diagnosis and at carefully selected times during the clinical course should help to refine the cellular biology of malignant plasma cells and disease progression, fine-tune the prognostic impact of genetically defined risk groups, expand the framework to stratify patients to riskadapted therapies, and define the potential risk of secondary disease in patients with MM.

Acknowledgments We are grateful to Dr. Sandra Wolman for her critical review of the manuscript and Diana Weigel for manuscript preparation. We are also indebted to the physicians, nurses, and staff of the Division of Hematology and HCT for assistance with obtaining samples for these studies. This work was supported in part by NIH CA-33572, CA-30206, and a private donation from the family of Bernie and Pearl Ruttenberg.

References [1] Willis TG, Dyer MJ. The role of immunoglobulin translocations in the pathogenesis of B-cell malignancies. Blood 2000;96:808–22. [2] Avet-Loiseau H, Facon T, Grosbois B, Magrangeas F, Rapp MJ, Harousseau JL, Minvielle S, Bataille R. Oncogenesis of multiple myeloma: 14q32 and 13q chromosomal abnormalities are not randomly distributed, but correlate with natural history, immunological features, and clinical presentation. Blood 2002;99:2185–91. [3] Fonseca R, Blood E, Rue M, Harrington D, Oken MM, Kyle RA, Dewald GW, Van Ness B, Van Wier SA, Henderson KJ, Bailey RJ, Greipp PR. Clinical and biologic implications of recurrent genomic aberrations in myeloma. Blood 2003;101:4569–75. [4] Zojer N, Konigsberg R, Ackermann J, Fritz E, Dallinger S, Kromer E, Kaufmann H, Riedl L, Gisslinger H, Schreiber S, Heinz R, Ludwig H, Huber H, Drach J. Deletion of 13q14 remains an independent adverse prognostic variable in multiple myeloma despite its frequent detection

108

[5]

[6] [7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109 by interphase fluorescence in situ hybridization. Blood 2000;95: 1925–30. Drach J, Schuster J, Nowotny H, Angerler J, Rosenthal F, Fiegl M, Rothermundt C, Gsur A, Jager U, Heinz R. Multiple myeloma: high incidence of chromosomal aneuploidy as detected by interphase fluorescence in situ hybridization. Cancer Res 1995;55:3854–9. Zandecki M, Lai JL, Facon T. Multiple myeloma: almost all patients are cytogenetically abnormal. Br J Haematol 1996;94:217–27. Nishida K, Tamura A, Nakazawa N, Ueda Y, Abe T, Matsuda F, Kashima K, Taniwaki M. The Ig heavy chain gene is frequently involved in chromosomal translocations in multiple myeloma and plasma cell leukemia as detected by in situ hybridization. Blood 1997;90:526–34. Fonseca R, Debes-Marun CS, Picken EB, Dewald GW, Bryant SC, Winkler JM, Blood E, Oken MM, Santana-Davila R, Gonzalez-Paz N, Kyle RA, Gertz MA, Dispenzieri A, Lacy MQ, Greipp PR. The recurrent IGH translocations are highly associated with nonhyperdiploid variant multiple myeloma. Blood 2003;102:2562–7. Facon T, Avet-Loiseau H, Guillerm G, Moreau P, Genevieve F, Zandecki M, Lai JL, Leleu X, Jouet JP, Bauters F, Harousseau JL, Bataille R, Mary JY. Chromosome 13 abnormalities identified by FISH analysis and serum beta2-microglobulin produce a powerful myeloma staging system for patients receiving high-dose therapy. Blood 2001;97: 1566–71. Fonseca R, Harrington D, Oken MM, Dewald GW, Bailey RJ, Van Wier SA, Henderson KJ, Blood EA, Rajkumar SV, Kay NE, Van Ness B, Greipp PR. Biological and prognostic significance of interphase fluorescence in situ hybridization detection of chromosome 13 abnormalities (delta13) in multiple myeloma: an Eastern Cooperative Oncology Group study. Cancer Res 2002;62:715–20. Konigsberg R, Zojer N, Ackermann J, Kromer E, Kittler H, Fritz E, Kaufmann H, Nosslinger T, Riedl L, Gisslinger H, Jager U, Simonitsch I, Heinz R, Ludwig H, Huber H, Drach J. Predictive role of interphase cytogenetics for survival of patients with multiple myeloma. J Clin Oncol 2000;18:804–12. Zojer N, Konigsberg R, Ackermann J, Fritz E, Dallinger S, Kromer E, Kaufmann H, Riedl L, Gisslinger H, Schreiber S, Heinz R, Ludwig H, Huber H, Drach J. Deletion of 13q14 remains an independent adverse prognostic variable in multiple myeloma despite its frequent detection by interphase fluorescence in situ hybridization. Blood 2000;95: 1925–30. Barlogie B, Jagannath S, Desikan KR, Mattox S, Vesole D, Siegel D, Tricot G, Munshi N, Fassas A, Singhal S, Mehta J, Anaissie E, Dhodapkar D, Naucke S, Cromer J, Sawyer J, Epstein J, Spoon D, Ayers D, Cheson B, Crowley J. Total therapy with tandem transplants for newly diagnosed multiple myeloma. Blood 1999;93:55–65. Drach J, Ackermann J, Fritz E, Kromer E, Schuster R, Gisslinger H, DeSantis M, Zojer N, Fiegl M, Roka S, Schuster J, Heinz R, Ludwig H, Huber H. Presence of a p53 gene deletion in patients with multiple myeloma predicts for short survival after conventional-dose chemotherapy. Blood 1998;92:802–9. Avet-Loiseau H, Li JY, Morineau N, Facon T, Brigaudeau C, Harousseau JL, Grosbois B, Bataille R. Monosomy 13 is associated with the transition of monoclonal gammopathy of undetermined significance to multiple myeloma. Intergroupe Francophone du Myelome. Blood 1999;94:2583–9. Elnenaei MO, Gruszka-Westwood AM, A’Hernt R, Matutes E, Sirohi B, Powles R, Catovsky D. Gene abnormalities in multiple myeloma: the relevance of TP53, MDM2, and CDKN2A. Haematologica 2003;88: 529–37. Smadja NV, Bastard C, Brigaudeau C, Leroux D, Fruchart C. Hypodiploidy is a major prognostic factor in multiple myeloma. Blood 2001;98:2229–38. Smadja NV, Leroux D, Soulier J, Dumont S, Arnould C, Taviaux S, Taillemite JL, Bastard C. Further cytogenetic characterization of multiple myeloma confirms that 14q32 translocations are a very rare event in hyperdiploid cases. Genes Chromosomes Cancer 2003;38:234–9.

[19] van Lom K, Hagemeijer A, Smit EM, Lowenberg B. In situ hybridization on May-Grunwald Giemsa-stained bone marrow and blood smears of patients with hematologic disorders allows detection of cell-lineagespecific cytogenetic abnormalities. Blood 1993;82:884–8. [20] Bernell P, Arvidsson I, Jacobsson B, Hast R. Fluorescence in situ hybridization in combination with morphology detects minimal residual disease in remission and heralds relapse in acute leukaemia. Br J Haematol 1996;95:666–72. [21] Nickenig C, Lang NK, Schoch C, Hiddemann W, Haferlach T. New insights into the biology of multiple myeloma using a combination of May-Gru¨nwald-Giemsa staining and fluorescence in situ hybridization techniques at the single cell level. Ann Hematol 2001;80:662–8. [22] Hardan I, Rothman R, Gelibter A, Cohen N, Shimoni A, Sokolovsky M, Reichart M, Ishoev G, Amariglio N, Rechavi G, Nagler A, Trakhtenbrot L. Determination of chromosome 13 status in bone marrow cells of patients with multiple myeloma using combined morphologic and fluorescence in situ hybridization analysis. Exp Hematol 2004; 32:254–60. [23] ISCN. An international system for human cytogenetic nomenclature. In: Mittelman F, editor. Basel: S. Karger, 1995. [24] Shimoni A, Nagler A, Kaplinsky C, Reichart M, Avigdor A, Hardan I, Yeshurun M, Daniely M, Zilberstein Y, Amariglio N, Brok-Simoni F, Rechavi G, Trakhtenbrot L. Chimerism testing and detection of minimal residual disease after allogeneic hematopoietic transplantation using the BioView (Duet) combined morphological and cytogenetical analysis. Leukemia 2002;16:1413–22. [25] Debes-Marun CS, Dewald GW, Bryant S, Picken E, Santana-Davila R, Gonzalez-Paz N, Winkler JM, Kyle RA, Gertz MA, Witzig TE, Dispenzieri A, Lacy MQ, Rajkumar SV, Lust JA, Greipp PR, Fonseca R. Chromosome abnormalities clustering and its implications for pathogenesis and prognosis in myeloma. Leukemia 2003;17:427–36. [26] Dascalescu CM, Callanan M, Chauvet M, Le Baccon P, PegourieBandelier B, Garban F, Sotto JJ, Leroux D. Interphase FISH: a rapid method for detecting malignant plasma cells in multiple myeloma patients submitted to autologous transplantation. Bone Marrow Transplant 1999;23:687–94. [27] Dring AM, Davies FE, Fenton JA, Roddam PL, Scott K, Gonzalez D, Rollinson S, Rawstron AC, Rees-Unwin KS, Li C, Munshi NC, Anderson KC, Morgan GJ. A global expression-based analysis of the consequences of the t(4;14) translocation in myeloma. Clin Cancer Res 2004;10:5692–701. [28] Tricot G, Sawyer JR, Jagannath S, Desikan KR, Siegel D, Naucke S, Mattox S, Bracy D, Munshi N, Barlogie B. Unique role of cytogenetics in the prognosis of patients with myeloma receiving high-dose therapy and autotransplants. J Clin Oncol 1997;15:2659–66. [29] Fassas AB, Spencer T, Sawyer J, Zangari M, Lee CK, Anaissie E, Muwalla F, Morris C, Barlogie B, Tricot G. Both hypodiploidy and deletion of chromosome 13 independently confer poor prognosis in multiple myeloma. Br J Haematol 2002;118:1041–7. [30] Zhu L, Somlo G, Zhou Bn, Shao J, Bedell V, Slovak ML, Liu X, Luo J, Yen Y. Fibroblast growth factor receptor 3 inhibited by shorthairpin RNAs leads to apoptosis in multiple myeloma. Mol Cancer Ther 2005. In press. [31] Plowright EE, Li Z, Bergsagel PL, Chesi M, Barber DL, Branch DR, Hawley RG, Stewart AK. Ectopic expression of fibroblast growth factor receptor 3 promotes myeloma cell proliferation and prevents apoptosis. Blood 2000;95:992–8. [32] Chesi M, Brents LA, Ely SA, Bais C, Robbiani DF, Mesri EA, Kuehl WM, Bergsagel PL. Activated fibroblast growth factor receptor 3 is an oncogene that contributes to tumor progression in multiple myeloma. Blood 2001;97:729–36. [33] Boersma-Vreugdenhil GR, Peeters T, Bast BJ, Lokhorst HM. Translocation of the IGH locus is nearly ubiquitous in multiple myeloma as detected by immuno-FISH. Blood 2003;101:1653. [34] Zojer N, Schuster-Kolbe J, Assmann I, Ackermann J, Strasser K, Hubl W, Drach J, Ludwig H. Chromosomal aberrations are shared by malignant plasma cells and a small fraction of circulating CD19⫹

M.L. Slovak et al. / Cancer Genetics and Cytogenetics 158 (2005) 99–109 cells in patients with myeloma and monoclonal gammopathy of undetermined significance. Br J Haematol 2002;117:852–9. [35] Fiserova A, Hajek R, Holubova V, Buchler T, Sobotka J, Kovarova R, Musilova R, Bourkova L, Bulikova A, Mareschova I, Janackova Z, Vanova P, Kuglik P, Vorlicek J, Penka M. Detection of 13q abnormalities in multiple myeloma using immunomagnetically selected plasma cells. Neoplasma 2002;49:300–6. [36] Fonseca R, Bailey RJ, Ahmann GJ, Rajkumar SV, Hoyer JD, Lust JA, Kyle RA, Gertz MA, Greipp PR, Dewald GW. Genomic abnormalities in monoclonal gammopathy of undetermined significance. Blood 2002;100:1417–24.

109

[37] Nilsson T, Nilsson L, Lenhoff S, Rylander L, Astrand-Grundstrom I, Strombeck B, Hoglund M, Turesson I, Westin J, Mitelman F, Jacobsen SE, Johansson B. MDS/AML-associated cytogenetic abnormalities in multiple myeloma and monoclonal gammopathy of undetermined significance: evidence for frequent de novo occurrence and multipotent stem cell involvement of del(20q). Genes Chromosomes Cancer 2004;41:223–31. [38] Gertz MA, Witzig TE, Pineda AA, Greipp PR, Kyle RA, Litzow MR. Monoclonal plasma cells in the blood stem cell harvest from patients with multiple myeloma are associated with shortened relapse-free survival after transplantation. Bone Marrow Transplant 1997;19: 337–42.