Fax: 011 +886-3-3286697
Clinical relevance of internal tandem duplication of the FLT3 gene in childhood acute myeloid leukemia
Article first published online: 12 JUN 2002
Copyright © 2002 American Cancer Society
Volume 94, Issue 12, pages 3292–3298, 15 June 2002
How to Cite
Liang, D.-C., Shih, L.-Y., Hung, I.-J., Yang, C.-P., Chen, S.-H., Jaing, T.-H., Liu, H.-C. and Chang, W.-H. (2002), Clinical relevance of internal tandem duplication of the FLT3 gene in childhood acute myeloid leukemia. Cancer, 94: 3292–3298. doi: 10.1002/cncr.10598
- Issue published online: 12 JUN 2002
- Article first published online: 12 JUN 2002
- Manuscript Accepted: 10 JAN 2002
- Manuscript Revised: 9 JAN 2002
- Manuscript Received: 11 OCT 2001
- Childhood Cancer Foundation of R.O.C.. Grant Number: 90-3
- National Science Council, Taiwan. Grant Number: NSC90-2314-B-182-086
- childhood acute myeloid leukemia;
- FLT3 internal tandem duplication;
- French-American-British subtype;
- fusion transcript
Recently, an internal tandem duplication of the FLT3 gene (FLT3/ITD) was found in approximately 20% of adult acute myeloid leukemia (AML) cases and associated with a poor outcome. However, there are few studies on FLT3/ITD in childhood AML, and the clinical significance of FLT3/ITD is thus unclear.
FLT3/ITD was analyzed in 80 children with de novo AML. The genomic DNA polymerase chain reaction (PCR) assay was performed to identify FLT3/ITD. Genescan analysis to determine the allelic distribution was then performed for those PCR products with aberrant bands. Direct sequencing of PCR products was also carried out in each sample with FLT3/ITD.
The incidence of FLT3/ITD was 11.3% (9 out of 80 patients) in AML, with 25% (3 out of 12 patients) in acute promyelocytic leukemia (APL) and 8.8% (6 out of 68 patients) in non-M3 AML. The size of duplicated fragments varied from 21 base pairs (bp) to 75 bp, and the mutant to wild type ratio of FLT3 ranged from 0.28 to 16.60 in the nine patients with FLT3/ITD. The incidence of FLT3/ITD in childhood AML in patients > 10 years of age was 24%, compared to 5% of those patients ≤ 10 years of age (P = 0.011). The six non-M3 AML patients with FLT3/ITD were all older than 10 years of age. In APL, FLT3/ITD was found in 2 of 2 patients with S-form PML/RARα, compared with 1 in 10 patients with non-S form PML/RARα(P = 0.045). There were no cytogenetic abnormalities or fusion transcripts derived from common specific translocations found in non-M3 AML patients with FLT3/ITD. There was no significant difference in treatment outcome between APL patients with FLT3/ITD and those without FLT3/ITD. The authors failed to find a correlation between the treatment outcome and the presence of FLT3/ITD in non-M3 AML patients. Instead, the authors found that all three patients with a mutant FLT3 to wild type ratio of greater than 2.0 died within eight months after diagnosis; two of them failed to achieve complete remission.
The current study shows that the mutant FLT3 to wild type ratio, but not the presence of FLT3/ITD itself, may serve as a potential marker to improve risk-assessment in childhood AML. Cancer 2002;94:3292–8. © 2002 American Cancer Society.
Modern chemotherapy has significantly improved the outcome of childhood acute myeloid leukemia (AML).1 Cytogenetic analysis has revealed the heterogeneity of AML and the prognostic impact of clonal chromosomal abnormalities.2 In addition, it is well recognized that the response to chemotherapy is significantly affected by the presence of certain molecular genetic abnormalities. However, only 40% of childhood AMLs have specific fusion transcripts resulting from chromosomal translocations which are of prognostic significance.3 Thus, it would be very helpful if new molecular genetic abnormalities other than the specific fusion transcripts could predict the outcome of childhood AML.
Recently, an internal tandem duplication of the FLT3 gene (FLT3/ITD) was identified.4FLT3/ITD was found in 20% of adult AMLs.5FLT3/ITD was also found to be associated with leukocytosis in acute promyelocytic leukemia (APL) and the AML transformation of myelodysplastic syndrome.6, 7 These observations suggested that FLT3/ITD might be associated with poor prognosis in AML patients.
The FLT3 gene belongs to the receptor tyrosine kinase class III family, which plays a central role in hematopoiesis.8 This gene locates at chromosome 13q12 and predominantly expresses in hematopoietic stem cells.9FLT3 is also expressed in human leukemia or lymphoma cell lines.10, 11 The length mutation of FLT3 occurs at the juxtamembrane (JM) domain through the first tyrosine kinase (TK) domain.8 The duplicated sequences occur mostly within exon 11, some involving intron 11, and first part of exon 12, with or without insertion; locations and lengths are different in every case.4, 5 The altered FLT3 gene is always transcribed in-frame and encodes a mutant FLT3 with a long JM domain.4, 5
To our knowledge, in childhood AML, there have only been four studies on FLT3/ITD;12–15 the incidences of FLT3/ITD varied from 5.3% to 16.5%, which were lower than those of adult AML.5 In these previous studies, however, data on the correlation between the molecular subtypes of specific fusion transcripts and FLT3/ITD were not available, and the level of FLT3/ITD was not measured. The current study, using genomic DNA polymerase chain reaction (PCR) with direct sequencing of the FLT3/ITD, and Genescan analysis to measure the ratio of mutant to wild type FLT3, aimed to determine the incidence of FLT3/ITD; to correlate the FLT3/ITD with clinical features, French-American-British (FAB) subtypes, cytogenetics, and common specific fusion transcripts; and to define the prognostic significance of FLT3/ITD in childhood AML.
MATERIALS AND METHODS
Eighty children (age ≤ 18 years) with de novo AML diagnosed at Mackay Memorial Hospital and Chang Gung Children's Hospital were studied at initial diagnosis. Informed consent was obtained from the guardians. The morphologic subtypes were classified according to the FAB Cooperative Group.16–18 Immunophenotyping was performed by flow cytometry using monoclonal antibodies against HLA-DR, CD13, CD14, CD15, CD33, CD41, CD61, CD34, CD2, CD3, CD5, CD7, and CD19. The mononuclear cells from diagnostic marrow samples were enriched by Ficoll-Hypaque (1.077 gm/mL; Pharmacia Biotech, Uppsala, Sweden) density gradient centrifugation and cryopreserved in 10% dimethyl sulfoxide and 20% fetal bovine serum in liquid nitrogen until testing. Before 1997, all AML patients were treated with the nationwide Taiwan Pediatric Oncology Group (TPOG) AML 901 protocol, consisting of induction therapy with two courses of epidoxorubicin 20 mg/m2/day, cytosine arabinoside (Ara-C) 100 mg/m2/day (3 + 7), and 6-thioguanine 80 mg/m2/day for seven days, consolidation therapy with two courses of etoposide 200 mg/m2/day for three days and cyclophosphamide 300 mg/m2/day for three days, and then maintenance therapy with five 8-week courses of epidoxorubicin 20 mg/m2/day, Ara-C 100 mg/m2/day (3 + 7) and 6-thioguanine 80 mg/m2/day for seven days, followed by etoposide and cyclophosphamide (doses as above described) and then by 6-mercaptopurine 80 mg/m2/day and methotrexate (MTX) 20 mg/m2/week for two weeks. After 1997, those with APL were treated with TPOG-APL-97 protocol. Basically, all-trans retinoic acid 30 mg/m2/day, idarubicin 9 mg/m2/day, and Ara-C 100 mg/m2/day (3+7) were given if the white blood cell (WBC) count was > 5,000/UL before treatment, > 6,000/UL on the fifth day, > 10,000/UL on the tenth day, or > 15,000/UL on the fifteenth day of induction therapy. The post-remission therapy for APL consisted of six courses of idarubicin and Ara-C (3 + 7; doses as above described). The non-APL patients were randomized to receive either: the TPOG-AML-97A protocol consisting of induction therapy with idarubicin and Ara-C (3 + 7; doses as APL-97) q3w, postremission therapy with four courses of Ara-C 1 gm/m2 q12h × 8 + etoposide 100 mg/m2/day × 5 alternated with Ara-C 1 gm/m2 q12h × 8 + mitoxantrone 10 mg/m2/day × 4, and then four courses of idarubicin 9 mg/m2/day and Ara-C 200 mg/m2/day (1 + 5); or the TPOG-AML-97B protocol, which was modified from MRC AML10 protocol19 with idarubicin substituted for daunorubicin, and cyclophosphamide substituted for amsacrine. Intrathecal (IT) MTX injection, 6, 8, 10, and 12 mg for patients aged < 1, 1-2, 2-3, and > 3 years, respectively, was given at the beginning of each chemotherapy course, except for high-dose Ara-C containing courses in 901, APL-97, and AML-97A protocols. In AML-97B, IT MTX and IT hydrocortisone were used instead.19 The clinicohematologic characteristics including age, gender, WBC count, FAB classification, karyotype, specific fusion transcript at diagnosis, event-free survival (EFS), and overall survival were analyzed.
Genomic DNA PCR Assay
Genomic DNAs were extracted from frozen bone marrow mononuclear cells collected at diagnosis from January 1995 to March 2001 by using a DNA extraction kit (Puregene Gentra Systems, Minneapolis, MN) according to the manufacturer's instruction. The PCR reaction was carried out with a mixture of 50 μL containing 150 ng genomic DNA, 200 μM dNTP, 1X gold PCR buffer, 1.5mM MgCl2, 1U Taq gold polymerase, 0.001% gelatin, 50 mM tetramethylammonium chloride, 6% dimethyl sulfoxide, and 30 pmol for each primer 11F (5′-CAA TTT AGG TAT GAA AGC C-3′) and 12R (5′-GTA CCT TTC AGC ATT TTG AC-3′), which completely covers the JM domain through the TK I domain4, 5 on a DNA thermal cycler (ABI 9600; Applied Biosystems, Foster City, CA) using a program consisting of denaturation at 94 °C for 30 seconds, annealing at 59 °C for 1 minute, and extension at 72 °C for 2 minutes for 35 cycles, with an initial preheating at 95 °C for 12 minutes and a final extension at 72 °C for 10 minutes. The PCR products were then run on a 3% Nusieve (BioWhittaker Molecular Applications, Rockland, ME) agarose gel, detected by ethidium bromide staining, and visualized under an ultraviolet lamp.
Genescan Analysis and Determination of Allelic Distribution
The genomic PCR assay was performed again from the DNA samples positive for mutant FLT3 as described above except that one primer was labeled at 5′ end with fluorescin. Four μL of PCR products were mixed with 5 μL of formamide (95%) and loading buffer (5% blue dextran, 25 mM ethylenediaminetetraacetic acid) containing 0.55 μL Rox-1000. Then 1.5 μL of this mixture were loaded on a 5% long Ranger/6M urea gel with 1X Tris 0.089M, Borate 0.089M, EDTA 0.002M (TBE) running buffer. Electrophoresis was performed at 200W for two and a quarter hours and analyzed with an automated DNA sequencer (ABI 377; Perkin-Elmer, Foster City, CA) and quantified using the Genescan 3.1 software (Perkin Elmer). The level of mutant FLT3 was calculated from the ratio of FLT3/ITD to wild type FLT3.
Sequencing of Longer PCR Products
Direct sequencing of abnormal PCR products was carried out in both directions with the Dye Terminator Cycle Sequencing Ready Reaction kit containing Ampli Taq DNA polymerase FS (Perkin-Elmer) on an automated DNA sequencing system (ABI 377) according to the manufacturer's instructions.
Cytogenetic Analysis and Detection of Common Fusion Transcripts
Cytogenetic analysis was performed using a direct method or short-term unstimulated cultures. G-banded metaphases were analyzed. The karyotypes were interpreted according to the International System for Human Cytogenetic Nomenclature.20 The detection of common fusion transcripts by reverse transcriptase (RT) PCR assays for AML1/ETO, PML/RARα, CBFβ/MYH11, and 11q23 abnormalities, including MLL-AF4, MLL-AF9, MLL-ENL, and MLL-ELL, followed by Southern blot analysis, was preformed as previously described.21
Associations of FLT3/ITD with clinical and biologic features were examined by Fisher exact test. Induction failure was defined as failure to achieve remission after induction therapy and early death before achieving remission. For the calculation of EFS, all deaths and relapses occurring after the start of chemotherapy were classified as events, and induction failure was assigned a time of zero. The univariate analysis between variables and survival or EFS was done using the log-rank test. A statistically significant difference was defined as a P value < 0.05. Because of the small sample size, multivariate analysis was not performed.
In 9 of 80 patients (11.3%), additional longer PCR products of the FLT3 gene were detected. Sequence analysis of the abnormal PCR products confirmed the presence of tandemly duplicated fragments. All duplication involved exon 11 of the FLT3 gene. The length of the duplicated fragments varied from 21 to 75 bp with a median of 36 bp; eight had simple internal duplication, and one had 24 bp internal duplication plus a 6 bp insertion. The FLT3/ITD found in our patients were all in-frame. The ratios of mutant FLT3/ITD to wild type FLT3 ranged from 0.28 to 16.60, with a median of 0.65. All but one patient had a ratio of > 0.5, and three patients (Patients 4, 5, and 9) had a ratio of >2.0.
Clinical Characteristics of Patients with FLT3/ITD
The clinical characteristics of the nine patients with FLT3/ITD are summarized in Table 1. FLT3/ITD was frequently associated with age > 10 years (P = 0.011). There was a trend for those with initial WBC count of ≥ 50 × 109/L to have a higher frequency of FLT3/ITD (P = 0.076). Since APL is a unique subgroup in AML both clinicobiologically and therapeutically, we separated AML patients into APL and non-M3 AML groups.
|Patient||Age (year)/Gender||FAB subtype||WBC (× 109/L)||Karyotype||Fusion transcripta||FLT3 duplicated sequence||Allelic ratio of mutant/wt FLT3||Treatment protocol||Induction failure||Relapse||Overall survival (months)|
|1||10/F||M3||277.1||t(15;17)||PML-RARα(L)||nt 1753-1788||0.65||No therapy||–||–||–|
|4||11/M||M5a||204.0||46,XY||Not found||nt 1729-1794||2.27||97B||Yes||–||8|
|5||14/F||M1||24.1||46,XX||Not found||nt 1793-1837||16.60||97B||Yes||–||1|
|6||14/M||M1||214.0||46,XY||Not found||nt 1794-1817||0.67||97B||No||No||39+|
|7||15/F||M5a||505.0||46,XX||Not found||nt 1757-1795||0.60||97B||No||No||32+|
|8||12/M||M2||171.0||ND||Not found||nt 1789-1809||0.51||97B||No||No||11+|
|9||11/F||M4||65.8||46,XX||Not found||nt 1779-1802 +6-bp||13.23||97A||No||No||5|
FLT3/ITD in APL
One fourth of patients (3 out of 12) with APL had FLT3/ITD. Of the 12 APL patients, 8 had L-form (bcr-1) of PML/RARα, 2 had V-form (bcr-2), and 2 had S-form (bcr-3). FLT3/ITD was found in two patients with S-form and in one of the eight patients with L-form of PML-RARα. There seemed to be no differences in age, gender, and WBC counts between patients with FLT3/ITD and those with wild type FLT3. Patient 1, who presented with hyperleukocytosis and intracerebral hemorrhage, died on the first hospital day before the initiation of anti-leukemic treatment. The remaining two APL patients had an EFS of 4+ and 14+ months, respectively, as of August 2001.
FLT3/ITD in Non-M3 AML Patients
Of the 68 non-M3 AML patients, only 6 patients (8.8%) had FLT3/ITD. The correlations of FLT3/ITD with clinical and hematologic features are summarized in Table 2. None had FAB subtypes of M0, M6, or M7. Subtypes M1 and M5 were more prevalent. There was no significant difference in the incidence of FLT3/ITD between the genders (P = 1.000). There was a statistically significant difference in the incidence of FLT3/ITD between ages > 10 years and ≤ 10 years (P = 0.0002). All six of the non-M3 AML patients with FLT3/ITD were older than 10 years. The WBC counts of these six patients ranged from 24.1 to 505.0 × 109/L, with a median of 187.5 × 109/L. The difference in the incidence of FLT3/ITD between WBC counts ≥ 50 × 109/L and < 50 × 109/L was statistically significant (P = 0.047). None of the non-M3 AML patients with FLT3/ITD had MLL rearrangement, fusion transcripts of AML1/ETO and CBFβ/MYH11, or other nonrandom chromosomal translocations, whereas in non-M3 AML patients without FLT3/ITD, t(8;21)/AML1-ETO, inv (16)/CBFβ-MYH11, and 11q23 abnormality were found in 14, 5, and 7 patients, respectively. The incidence of FLT3/ITD in non-M3 AML patients was lower than APL patients (8.8% vs. 25%), but the difference did not reach statistical significance (P = 0.129).
|Feature||No. of patients||FLT3/ITD(+)||P value|
|FAB Subtype||M0||4||0 (0)|
|Age||< 10 years||47||0||0.0002|
|≥ 10 years||21||6|
|WBC (× 109/L)||< 50||39||1||0.047|
Three of the six non-M3 AML patients with FLT3/ITD died (Table 1). Patients 4 and 5 failed to achieve complete remission and survived eight months and one month, respectively. Patient 9 died of fungal infection during remission with a survival of five months; these three patients had a mutant FLT3 to wild type ratio of > 2.0. Patients 6, 7, and 8, with lower mutant FLT3 to wild type ratios, had EFSs of 10+, 31+, and 38+ months, respectively. Patient 7, with a WBC count of 505.0 × 109/L treated with the AML-97B protocol, received allogeneic bone marrow transplantation (BMT) with cyclophosphamide and total body irradiation as a conditioning regimen three months after complete remission.
Prognostic Value of FLT3/ITD and Other Factors in AML
There was no difference in outcome between APL patients with FLT3/ITD and those without FLT3/ITD with respect to remission rate (P = 1.000), EFS (P = 0.355), and overall survival (P = 0.440). In APL patients, age also did not correlate with treatment outcome (P = 0.959 for overall survival and P = 0.608 for EFS). Since the only APL patient with a WBC ≥ 50 × 109/L did not receive therapy, analysis of treatment outcome with respect to WBC count was not performed in APL patients.
The complete remission rate did not differ between non-M3 AML patients with FLT3/ITD and those without FLT3/ITD (P = 0.322). In non-M3 AML patients, a WBC count ≥ 50 × 109/L had a trend of predicting poor overall survival (P = 0.069), whereas the presence of FLT3/ITD (P = 0.446) and age (P = 0.314) did not. Similarly, neither age (P = 0.310), WBC count (P = 0.938), nor presence of FLT3/ITD (P = 0.964) was identified as a poor risk factor for EFS in non-M3 AML patients. Six non-M3 AML patients had unfavorable chromosomal aberrations. Of the three non-M3 AML patients with complex chromosomal abnormalities, one died in relapse at 12 months after diagnosis, and the other two have remained in disease free for 17+ and 29+ months, respectively. Three patients had monosomy 7, only one of whom achieved remission but died of BMT-related complication at 13 months after diagnosis; the other two patients died 1 and 10 months, respectively, after diagnosis. Fourteen patients had poor response to initial therapy (more than two induction courses), and all died with a median survival of 5 months, ranging from 0 to 24 months.
The current study showed that the frequency of FLT3/ITD in childhood AML was 11.3%, which was comparable to the incidences of 5.3% to 16.5% in previous studies of childhood AML.12–15 The incidences of childhood AML were lower than those reported in adult AML, which varied between 13.2% and 23.3%.5, 22–25 We found that the incidence of FLT3/ITD was even lower in children ≤ 10 years old, 5% of the current study. Interestingly, the ages of all APL patients with FLT3/ITD were ≤ 10 years and of non-M3 AML patients > 10 years in the current series. Apart from age, patients with FLT3/ITD were more likely associated with a high WBC count, which was also observed in the reported series of adult and childhood AML.12–14, 22
Kondo et al. found that two out of three pediatric APL patients had FLT3/ITD,12 whereas the frequency of APL was 0-11% in other pediatric series.13, 14 The incidence of FLT3/ITD in our patients with M3 subtype was 25% (3 out of 12), comparable to 10-33% in adult patients with M3 subtype.6, 23–25 In the current study, the frequency of FLT3/ITD was higher in APL than in non-M3 AML but did not reach statistical significance because of the small patient number.
To our knowledge, there have been no data focused on the correlation of FLT3/ITD with fusion transcripts derived from specific chromosomal translocations.12–15 In the current study, RT-PCR followed by Southern blot analysis for fusion transcripts of common specific translocations, which was more sensitive than the conventional cytogenetics for the detection of nonrandom chromosomal translocations, was performed in each patient. Two APL patients with S-form PML/RARα had FLT3/ITD, compared to one in eight with L-form and none of the two with V-form PML/RARα, indicating that S-form PML/RARα was more frequently associated with FLT3/ITD than non-S form APL. None of our patients with FLT3/ITD had the specific genetic abnormalities of AML1/ETO, CBFβ/MYH11, or fusion transcripts of 11q23 abnormalities. In previous studies by cytogenetic analysis, t(15;17) was found in three of the four series of childhood AML,12–14 but only one patient each with t(8;21), inv(16), or 11q23 anomaly was found in the series of Meshinchi et al.15
Several factors have been known to be associated with poor prognosis in pediatric AML, i.e. complex karyotypes, monosomy 7, and poor initial response to induction therapy. None of our patients with monosomy 7 or poor response to initial treatment had long-term survival. FLT3/ITD has recently been reported to have prognostic significance in childhood and adult AML.12–15, 23, 24 We found that the presence of FLT3/ITD did not significantly influence the remission rate and outcome of our APL patients, which was in agreement with adult APL.6 These findings indicated that the presence of FLT3/ITD in APL did not confer a poor prognosis, which was probably attributed to the greatly improved outcome of APL patients treated with all trans-retinoic acid and combination chemotherapy containing anthracyclines.26 We also failed to find that FLT3/ITD was an independent factor of poor outcome in non-M3 AML. Recently, Schnittger et al. found that FLT3/ITD had no prognostic significance regarding overall survival, remission rate, or relapse rate in a very large series of 652 adult AML patients.25 A recent Cancer and Leukemia Group B study showed no difference in complete remission rate between FLT3/ITD (+) and FLT3/ITD (−) adult AML.27 The patient number in the current pediatric series was small, which precludes us from making a firm conclusion on the limited data set. In light of the small sample size with the well known heterogeneity of pediatric AML, the prognostic implication of FLT3/ITD remains to be determined in a larger series in childhood AML.
The levels of ITD mutants were variable among our patients harboring FLT3/ITD; however, it was not likely that the ratio of FLT3/ITD to wild type allele was caused by the variable percentages of leukemic cells in different samples, since, at the time of DNA extraction, the percentage of leukemic cells after enrichment of blasts in each sample analyzed was at least 90%. The wild type allele detected could not be attributed to the contamination of a very low percentage of normal cells in the samples. It was noteworthy that a high ratio of mutant to wild type FLT3 (> 2.0) was observed to be associated with a poor outcome in our patients. Of the three patients with a high level of FLT3/ITD, two failed to attain remission, compared to none of the five patients with a lower ratio. These three patients all died within eight months after diagnosis. Very recently, Whitman et al. showed that overall survival and EFS were inferior in FLT3/ITD (+) adult AML patients who lacked an FLT3 wild type allele, which was detected by using PCR and loss of heterozygosity analysis.27 The current observation is consistent with those findings, as evidenced by the fact that Patients 5 and 9 had very high ITD mutant to wild type ratios reflecting the loss of wild type allele; both patients had an overall survival of less than six months. The current results suggest that the ratio of mutant to wild type FLT3, but not the presence of FLT3/ITD itself, may be a potential predictor for a poor prognosis in childhood AML.
The authors thank Jen-Fen Fu, Ph.D. and Ms. Hui-Chin Hsu for technical assistance and Ms. Yu-Feng Wang for secretarial assistance.
- 5Internal tandem duplication of the FLT3 gene is preferentially seen in acute myeloid leukemia and myelodysplastic syndrome among various hematological malignancies: a study on a large series of patients and cell lines. Leukemia. 1997; 11: 1605–1609., , , et al.
- 14Tandem duplication of the FLT3 gene is found in acute lymphoblastic leukaemia as well as acute myeloid leukaemia but not in myelodysplastic syndrome or juvenile chronic myelogenous leukaemia in children. Br J Haematol. 1999; 105: 155–162., , , et al.
- 19Randomized comparison of DAT versus ADE as induction chemotherapy in children and younger adults with acute myeloid leukemia. Results of the medical Research Council's 10th AML Trial (MRC AML 10). The Adult and Childhood Leukaemia Working Parties of the Medical Research Council. Blood. 1997; 89: 2311–2318., , , et al.
- 20MitelmanF, editor. An international system for human cytogenetic nomenclature. Basel: Karger, 1995.
- 25FLT3 length mutation in AML: correlation to cytogenetics, FAB-subtype, and prognosis in 652 patients [abstract]. Blood. 2000; 96( Suppl. 1): 826a., , , et al.