To determine if therapy-related acute myeloid leukemia (t-AML) with t(8;21)(q22;q22) [t-AML-t(8;21)] harbors similar characteristic clinicopathologic features as de novo AML-t(8;21) (q22;q22), we studied 13 cases of t-AML-t(8;21) and 38 adult cases of de novo AML-t(8;21) diagnosed and treated at our hospital (1995–2008). Of 13 t-AML-t(8;21) cases, 11 had previously received chemotherapy with or without radiation for malignant neoplasms and 2 received radiation alone. The median latency to t-AML onset was 37 months (range, 11–126 months). Compared with patients with de novo AML-t(8;21), patients with t-AML-t(8;21) were older (P = .001) and had a lower WBC count (P = .039), substantial morphologic dysplasia, and comparable CD19/CD56 expression. The AML1-ETO (RUNX1-RUNX1T1) fusion was demonstrated in all 10 cases assessed. Class I mutations analyzed included FLT3 (0/10 [0%]), RAS (0/10 [0%]), JAK2 V617 (0/11 [0%]), and KIT (4/11 [36%]). With a median follow-up of 13 months, 10 patients with t-AML-t(8;21) died; the overall survival was significantly inferior to that of patients with de novo AML-t(8;21) (19 months vs not reached; P = .002). These findings suggest that t-AML-t(8;21) shares many features with de novo AML-t(8;21)(q22;q22), but affected patients have a worse outcome.
Therapy-related acute myeloid leukemia
Acute myeloid leukemia (AML) associated with the t(8;21)(q22;q22) [henceforth referred to as AML-t(8;21)] represents approximately 5% to 12% of de novo AMLs.1 Morphologically, most cases are characterized as M2 according to the French-American-British (FAB) classification.2 At the molecular genetic level, this neoplasm is defined by the presence of the t(8;21)(q22;q22) involving the AML1 (RUNX1) gene on chromosome 21q22.3 and the ETO (RUNX1T1) gene on chromosome 8q22. The AML1-ETO (RUNX1-RUNX1T1) fusion product disrupts the core binding factor transcription complex, which affects cell differentiation, proliferation, apoptosis, and self-renewal3 and, thus, initiates leukemogenesis.4 Patients with de novo AML-t(8;21) have a high complete remission rate and relatively long disease-free survival, especially adults treated with high-dose cytarabine in the consolidation phase.5–9
Therapy-related AML (t-AML) is a known complication of cytotoxic chemotherapy and radiation therapy, which are known to be mutagenic.10 Causative agents include alkylating agents and topoisomerase-II inhibitors and platinum drugs. Radiation therapy, especially at low to intermediate doses, is also known to induce t-AML. The outcomes for patients with t-AML have been historically poor compared with those for patients with de novo AML.11 t-AML-t(8;21) is uncommon, with the many sporadic cases reported in the English literature.2–17 Quesnel et al18 reviewed a total 26 cases of t-AML with t(8;21) from 7 centers previously reported in the literature and concluded that patients with t-AML-t(8;21) had hematologic characteristics and responses to therapy very similar to patients with de novo disease. As part of an international workshop in 2002, 44 cases of t-AML-t(8;21) were reported,19 11 with subsequent morphologic review and 3 with immunophenotypic analysis.20 These cases were obtained from multiple institutions (17 centers), but the patients’ treatment information was not obtained as part of the central review. This report showed that patients with t-AML-t(8;21) had a median overall survival (OS) of 19 months; however, this survival was not compared with that for de novo AML-t(8;21).
Although these earlier studies are helpful, a number of questions remain to be addressed about t-AML-t(8;21). Is the AML1-ETO fusion always present in t-AML-t(8;21)? Does this translocation in the setting of t-AML correlate with the presence of the characteristic morphologic features and immunophenotype observed in de novo AML-t(8;21)? Does the presence of t(8;21)(q22;q22) imply an excellent response to chemotherapy and a favorable prognosis in patients with t-AML? Are class I mutations that are commonly involved in leukemogenesis, such as RAS, FLT3, JAK2, and KIT, involved in the pathogenesis of t-AML-t(8;21)?
In this study, we reviewed a total of 13 cases of t-AML-t(8;21) diagnosed and treated at our hospital from May 1995 to July 2008. We report in detail the previous treatment and exposure to cytotoxic agents; describe the morphologic, immunophenotypic, and cytogenetic characteristics; and review the treatment modalities, response to treatment, and outcomes. We also compared this group with patients with de novo AML-t(8;21) diagnosed and treated at the same institution during the same time period. Lastly, we analyzed the potential role of important class I genetic mutations in the pathogenesis of t-AML-t(8;21), including JAK2 V617, FLT3, RAS, and KIT.
Materials and Methods
We searched the files of the Clinical Cytogenetics Laboratory, Department of Hematopathology, The University of Texas M.D. Anderson Cancer Center, Houston, from 1995 to 2008 for cases of t-AML-t(8;21). Clinical information was obtained by review of the medical records in accordance with an institutional review board–approved protocol. Diagnosis and classification were based on World Health Organization classification criteria.1 In t-AML, the latency was defined as the time from first chemotherapy or radiation treatment for the primary disease to the development of t-AML.19 OS was measured from the day of diagnosis of t-AML until death of any cause, censored for patients known to be alive at the last follow-up. The treatment response and prognosis for these patients were compared with those for a group of 38 patients with de novo AML-t(8;21) diagnosed and treated at our institution during the same period. These 38 cases were included in a previously published project studying the risk factors in de novo AML with t(8;21).7
All cases had representative trephine bone marrow core biopsy specimens and bone marrow aspirate smears available for evaluation. Cytochemical staining for myeloperoxidase (MPO) was performed on aspirate smears in all 12 cases assessed. The smears were specifically assessed for Auer rods, granulocytic differentiation, morphologic evidence of dysplasia in maturing elements, and percentage of eosinophils. The diagnostic criteria for AML-t(8;21), de novo or t-AML, were based on World Health Organization criteria, in which no minimum number of blasts is required for the diagnosis of AML in the presence of t(8;21)(q22;q22).1 The criteria of the FAB classification were also used to describe blast morphologic features and differentiation.
Flow Cytometric Immunophenotypic Analysis
We performed 3- or 4-color flow cytometric analysis using a FACScan or FACSCalibur flow cytometer (BD Biosciences, San Jose, CA) on peripheral blood or bone marrow aspirate specimens collected in EDTA. After incubation of the cells with monoclonal antibodies for 10 minutes at 4°C, the erythrocytes were lysed with NH4CL for 10 minutes, followed by 2 washing steps using phosphate-buffered saline solution. The cells then were resuspended and fixed with 1% paraformaldehyde. A panel of monoclonal antibodies was used, including reagents specific for CD2, CD3, CD5, CD7, CD10, CD13, CD14, CD15, CD19, CD20, CD33, CD34, CD64, CD117, CD56, HLA-DR, MPO, and terminal deoxynucleotidyl transferase (BD Biosciences). Cases were considered CD56+ if CD56 was expressed on 15% or more7 of the cells in the leukemia cell gate and if coexpression of CD56 with CD34 and CD117 on the surface of leukemia cells was demonstrated.
Conventional Karyotyping and Fluorescence in Situ Hybridization
Conventional cytogenetic analysis was performed on metaphase cells prepared from bone marrow aspirates cultured for 24 or 48 hours without mitogens using standard techniques as previously described.7 Giemsa-banded meta-phases were analyzed, and the results were reported using the International System for Human Cytogenetic Nomenclature. Fluorescence in situ hybridization (FISH) for AML1/ETO rearrangement was performed on interphase nuclei using the LSI AML1/ETO dual-color, dual-fusion translocation probe (Vysis, Downers Grove, IL) at the time of diagnosis as part of the routine clinical workup.
Molecular Diagnostic Studies
For the detection of AML1-ETO (RUNX1-RUNX1T1) fusion transcripts by reverse transcriptase–polymerase chain reaction (RT-PCR), RNA was extracted from bone marrow aspirate samples using Trizol reagent (Gibco-BRL, Gaithersburg, MD) according to the manufacturer’s instructions. Complementary DNA was synthesized by using SuperScript II Reverse Transcriptase (Gibco-BRL) as described previously.21 The resulting complementary DNA was subjected to PCR amplification with the following oligonucleotide primers and probe: 5′ primer (AML1 gene, 5′-AGC-TTC-ACT-CTG-ACC-ATC-AC-3′); 3′ primer (ETO gene, 5′-TGA-ACT-GGT-TCT-TGG-AGC-TCCT-3′), and 821 probe (ETO gene, 5′-TTC-ACA-AAC-CCACCG-GAA-GTA-3′). The Kasumi-1 and HL-60 cell lines were used as positive and negative control samples, respectively. Sequencing analysis of exons 8 and 17 of the c-KIT gene was performed using genomic DNA extracted from bone marrow aspirate samples and methods described previously.22 Methods for detecting FLT3 internal tandem repeat (ITD) or D835 point mutation and assays to detect RAS (K- and N-) mutations were described previously.23 Assessment for JAK2 V617 mutations was performed according to a previously described protocol.24 RT-PCR, RAS, and FLT3 studies were performed at the time of diagnosis as part of the routine clinical workup; JAK2 V617 and KIT mutation studies were specifically performed in this study.
The Mann-Whitney test was used for numerical comparisons between 2 groups. The Fisher exact and χ2 tests were applied for categorical variables. Patient survival was estimated by using the Kaplan-Meier method; survival was calculated from the date of bone marrow diagnosis until death of any cause or until the last patient follow-up. Survival curves were statistically compared by using the log-rank test. Differences between 2 groups were considered statistically significant if the P values were less than .05 in a 2-tailed test.
A total of 13 cases were identified in which patients had a clinical history of chemotherapy, radiation therapy, or both, and t-AML-t(8;21) had developed Table 1. Of the patients, 12 had a history of a malignant neoplasm, including breast carcinoma (n = 4), classical Hodgkin lymphoma (n = 2), mucosa-associated lymphoid tissue lymphoma (n = 1), anaplastic large cell lymphoma (n = 1), diffuse large B-cell lymphoma (n = 2), squamous cell carcinoma of the lung (n = 1), and prostate carcinoma (n = 1). One patient (case 12) had nonmalignant thyroid disease treated with radiation therapy. In 6 cases, chemotherapy and radiation therapy were used (cases 1, 2, and 8–11), in 5 cases only chemotherapy was used (cases 4–7 and 13), and in 2 cases, only radiation therapy was used (cases 3 and 12). In 10 cases, the chemotherapy regimens included alkylating agents and DNA topoisomerase II inhibitors. One patient (case 11) received radiation 156 months before the diagnosis of t-AML, and docetaxel, a drug that disrupts the microtubule network in dividing cells, 31 months before the diagnosis of t-AML. The chemotherapy regimens are listed in Table 1.
The median latency interval between the date of therapy and the t-AML diagnosis was 37 months (range, 11–156 months). One patient (case 2) had refractory cytopenia with multilineage dysplasia and ringed sideroblasts (RCMD-RS) without t(8;21)(q22;q22) initially, but the disease evolved to refractory anemia with excess blasts (RAEB; FAB) after 11 months with the acquisition of t(8;21)(q22;q22). The other 12 patients had no antecedent myelodysplastic syndrome. At the time of the t-AML diagnosis, 2 patients had active primary carcinoma (cases 9 and 10), and the others had no detectable evidence of the primary malignant neoplasm. The patients’ previous diseases, treatment modalities, and intervals to the diagnosis of t-AML are summarized in Table 1.
Compared with 38 patients with de novo AML-t(8;21) Table 2, the patients with t-AML-t(8;21) were older (P = .001), had a comparable male/female ratio, and had a lower WBC count (P = .039). The median follow-up was 13 months (range, 2–41 months) for t-AML-t(8;21) and 37 months for de novo AML-t(8;21) (range, 4–104 months).
Morphologic Features and Immunophenotype of t-AML With t(8;21)(q22;q22)
At the time of the t-AML diagnosis, the median blast count was 46% (range, 16%-96%) in bone marrow aspirate smears and 13% (range, 0%–89%) in peripheral blood smears Table 3. The median estimated bone marrow cellularity was 40% (range, 10%–100%). In 8 of 13 cases, the blasts were large with indented nuclei, a perinuclear hof, salmon-colored cytoplasmic granules, and one to several long slender Auer rods (M2 using FAB criteria) Image 1. Two cases (cases 4 and 7) with blasts in the range of 20% to 30% would be classified as RAEB in transformation (RAEB-t), and the case with 16% blasts (case 2) would be RAEB-2 using the FAB classification. One case showed M1 and another M5b morphologic features. Morphologic evidence of dysplasia was assessed in the cases with sufficient maturing elements present in each respective lineage. Dysgranulopoiesis was observed in 9 (100%) of 9 cases and dyserythropoiesis in 7 (88%) of 8 (Image 1B), with ringed sideroblasts present in 1 case (case 2). The numbers of megakaryocytes were markedly decreased in all cases. In the 2 cases with sufficient megakaryocytes to assess, both showed dysmegakaryopoiesis. Eosinophils ranged from 0% to 4% and averaged 1%. Cytochemical stains for MPO were strongly positive in all 12 cases assessed.
Immunophenotypic analysis by flow cytometry was performed in 12 of 13 cases (Table 3). All cases expressed CD13, CD34, CD117, HLA-DR, and MPO and were negative for terminal deoxynucleotidyl transferase, CD2, CD3, CD5, and CD7. Of the 12 cases, 5 (42%) cases showed dim CD33 or negative CD33 expression. Expression of CD19 was seen in 10 (83%) of 12 cases of t-AML and in 35 (95%) of 37 de novo AML cases and showed no statistically significant difference (P = .248; Table 2). CD56 expression was positive in 9 (82%) of 11 t-AML and 10 (59%) of 17 de novo AML cases and showed no statistically significant difference (P = .249; Table 2).
Conventional karyotypic analysis performed on bone marrow aspirate specimens obtained at the time of initial examination demonstrated the t(8;21)(q22;q22) in all 13 cases. Five cases showed t(8;21)(q22;q22) as a sole cytogenetic abnormality. The remaining cases showed additional karyotypic abnormalities (Tables 1 and 3). Additional abnormalities commonly associated with t(8;21) included loss of a sex chromosome, seen in 3 cases (23%; cases 4, 6, and 8), and trisomy 4 in 2 cases (15%; cases 9 and 12). No cases showed del(9q), a relatively common additional abnormality in de novo AML with t(8;21). The cytogenetic findings in the t-AML cases compared with the de novo cases were not statistically significantly different (P = 1.00; Table 2).
FISH for AML1-ETO rearrangement demonstrated the fusion gene in all 7 cases assessed (Table 1).
RT-PCR to assess for AML1-ETO (RUNX1-RUNX1T1) fusion transcripts was performed on bone marrow aspirate material in 8 cases. The fusion transcript was detected in all 8 cases (100%; Table 1). PCR-based DNA sequencing analysis of exons 8 and 17 of the KIT gene was performed in 11 cases and was positive in 4 (cases 5, 6, 9, and 12). All 4 positive cases showed mutations in exon 17 at various sites, including 2 cases of D816V, 1 case of D816Y, and 1 case with D816A and I817L (case 6). Of the 4 cases, 2 showed trisomy 4 as an additional cytogenetic abnormality (cases 9 and 12). There was no evidence of K-RAS (n = 10), N-RAS (n = 10), JAK2 V617 (n = 11), or FLT3 mutation, the latter including ITD and point mutations at codon 835 (n = 10). In contrast, in the de novo AML-t(8;21) cases, FLT3 mutations were identified in 2 (11%) of 19 cases, N-RAS mutations in 2 (11%) of 19 cases, and KIT exon 17 mutations in 2 (17%) of 12 cases. The 2 patients with positive results for the KIT mutation died at 6 and 8 months. JAK2 V617 was not tested in any of the de novo AML cases. The differences in any mutations (class I) were not statistically significant between t-AML-t(8;21) and de novo AML-t(8;21).
(Case 6) Bone marrow aspirate smears. Cases of therapy-related acute myeloid leukemia (AML) associated with t(8;21)(q22;q22) with characteristic perinuclear hof and cytoplasmic granules, including cells with large pink granules (salmon-colored) (A) as often seen in de novo AML-t(8;21). B, Dysplastic erythrocytes are present (A and B, Wright-Giemsa, x1,000).
Treatment information and responses are listed in Table 1. Of the patients with t-AML-t(8;21), 11 received induction chemotherapy: 6 received idarubicin and cytarabine, and 6 received fludarabine-based regimens. In 10 (91%) of 11 patients, complete remission was achieved; however, 7 experienced relapse of AML and received reinduction therapy using various salvage protocols, including bone marrow transplantation (cases 1, 4, and 10). With a median follow-up of 13 months, 10 patients died and 3 patients were alive at 7, 7, and 36 months after t-AML diagnosis (Table 1). In comparison (Table 2), in the de novo AML-t(8;21) group, all treated with similar chemotherapy regimens at our institution, complete remission was achieved in 36 (95%) of 38 patients, and 14 (39%) of 36 experienced disease relapse. However, the lower relapse rate was not statistically significantly different from the rate for the t-AML-t(8;21) group (P = .150). The median OS for patients with t-AML-t(8;21) was 19 months, whereas the OS for patients with de novo AML-t(8;21) was not reached in a median follow-up of 37 months (range, 4–104 months). This difference was statistically significant (P = .0021; Kaplan-Meier log-rank) Figure 1.
In this study, we analyzed in detail the findings for 13 adults with t-AML-t(8;21) and compared them with the findings for 38 adults with de novo AML-t(8;21) diagnosed and treated at the same institution in the same period. We showed that the blasts in t-AML and de novo AML associated with t(8;21)(q22;q22) shared characteristic morphologic and immunophenotypic features, and affected patients had a comparable initial response to induction chemotherapy. Patients with t-AML-t(8;21), however, had a shorter median survival of 19 months. The median OS of our patients was essentially the same as that reported by Slovak and coworkers,19 who reviewed 44 cases from 17 centers.
Comparison of overall survival in 13 therapy-related acute myeloid leukemia (AML) cases and 38 de novo AML cases associated with t(8;21)(q22;q22) (P = .0021; Kaplan-Meier log rank). t-AML, therapy-related AML.
One question we had, a priori, was do balanced translocations between 8q22 and 21q22 in t-AML always lead to AML1-ETO (RUNX1-RUNX1T1) fusion? The complete 260-kb genomic sequence of AML1 on chromosome 21 contains 9 exons, and the genomic breakpoints in AML1 are often located in intron 5 in t(8;21).25 Ten in vivo topoisomerase II DNA cleavage sites have been mapped to AML1, with 2 sites mapped to ETO,26 and DNA damage at these sites due to topoisomerase II inhibitor therapy is thought to lead to chromosome translocations. Slovak and coworkers19 showed that patients who had t-AML with other 8q22 abnormalities had survival inferior to that of patients with t-AML-t(8;21). In addition, they also showed that t(8;21)(q22;q22) did not necessarily incur an AML1-ETO fusion because cryptic 3-way translocations can occur. We were able to show AML1-ETO fusion in all 10 cases tested by FISH, RT-PCR, or both. It is noteworthy, however, that transcript heterogeneity for AML1-ETO gene fusion has recently been found as a result of differential promoter usage, internal deletions and insertions, and alternate splicing involving exons 2 through 5 of AML1 and exons 2, 3, and 9a of ETO. These results suggest that synthesis of alternative AML1-ETO transcript and protein forms can significantly affect the regulation of AML1 responsive genes and may account for the clinical heterogeneity.27
According to the 2-hit model of leukemogenesis proposed by Gilliland,28 the development of AML needs the cooperation of at least 2 different types of mutations: class I mutations induce myeloproliferation via activating mutations involving signal transduction pathways, and class II mutations are frequently represented by reciprocal gene fusions and lead to a block in differentiation. Case 2 in our study had RCMD-RS initially, but the disease evolved to AML with acquisition of t(8;21)(q22;q22), consistent with this model.
We hypothesized that the adverse prognosis of patients with t-AML-t(8;21) may be attributed to different molecular genetic mutations, especially involving class I molecules. We therefore tested t-AML-t(8;21) cases for K-RAS, N-RAS, and FLT3 mutations but detected no mutations. In contrast, in the de novo AML-t(8;21) cohort, FLT3 and N-RAS mutations were each identified in 2 (11%) of 19 cases. These findings agree with the observation by Kuchenbauer and coworkers29 that common class I gene mutations, such as FIT3 and RAS, are infrequent in t-AML-t(8;21).
Recently, Schnittger and coworkers30 analyzed JAK2 V617F mutation status in 24 consecutive cases of AML with t(8;21) AML1-ETO and found 2 positive cases. It is interesting that both cases with a positive mutation were therapy-related. They proposed that the JAK2 V617 mutation might be a class I mutation involved in t-AML-t(8;21). We therefore assessed for the JAK2 V617 mutation in our cases of t-AML-t(8;21); however, we detected no mutations in any of the 11 cases assessed. Because JAK2 V617F mutation in de novo AML with t(8;21) was not the focus of this study, it was not tested in de novo cases.
The frequency of KIT mutations has been reported to range from 5% to 48% in AML with t(8;21) depending on the method and patients’ ethnic background and has been reported to negatively affect OS and event-free survival.22,31,32 However, these studies did not distinguish between de novo and t-AML cases. In our previously reported study of 53 cases of de novo AML-t(8;21)(q22;q22) (38 newly diagnosed and 15 relapsed), we detected KIT mutations in 4 (44%) of 9 patients who survived fewer than 14 months and in only 1 (10%) of 10 patients who survived more than 41 months,7 suggesting that KIT mutation is an adverse prognostic marker in de novo AML-t(8;21). Mutations in KIT exon 17 were detected in 4 (36%) of 11 cases of t-AML-t(8;21); 3 patients died within 9 months of diagnosis, and 1 survived for 36 months. Two patients with KIT mutations had trisomy 4 as an additional chromosomal abnormality, supporting the hypothesis that trisomy 4 (KIT is located on chromosome 4) leads to an increased gene dosage of KIT.33 We also showed that additional chromosomal abnormalities in t-AML-t(8;21) and de novo AML-t(8;21) were not significantly different. Nevertheless, we and others have shown previously that additional cytogenetic alterations do not predict an inferior outcome in de novo AML-t(8;21)7 or t-AML with t(8;21).19
Morphologically, we confirmed the observations of others that t-AML and de novo AML associated with t(8;21) (q22;q22) share common morphologic features including cytoplasmic chunky granules, a perinuclear hof, slender Auer rods, and increased promyelocytes.17,20 However, other morphologic variants can be seen occasionally. In our study group of 13 t-AML-t(8;21) cases, 1 showed M1 and 1 showed M5b morphologic features. In contrast with de novo AML-t(8;21), we observed that t-AML-t(8;21) almost always had pronounced morphologic dysplasia in the maturing elements; an increase in eosinophils was uncommon. Immunophenotypically, the myeloblasts in de novo AML-t(8;21) characteristically express CD19 and CD56.7,34 This distinct immunophenotype in de novo AML cases suggests the possibility of t(8;21)(q22;q22), which requires confirmation by cytogenetic or molecular genetic studies. Unlike earlier studies, the results of immunophenotypic analysis by flow cytometry were available for most of the cases in our study group and showed that t-AML-t(8;21) has an immunophenotype similar to that of de novo AML-t(8;21). Our findings suggest that the AML1-ETO (RUNX1-RUNX1T1) fusion is primarily responsible for the distinctive immunophenotype, as well as the characteristic morphologic features, regardless of the presence of additional molecular genetic abnormalities or previous cytotoxic exposure.
In summary, we have found that t-AML-t(8;21) shares morphologic and immunophenotypic features with de novo AML-t(8;21), and both are characterized by the AML1-ETO (RUNX1-RUNX1T1) fusion. However, AML1-ETO fusion in t-AML does not predict a favorable outcome in the therapy-related setting. The poor prognosis and aggressive clinical behavior of t-AML-t(8;21) may be explained, in part, by the older age of patients with t-AML and the presence of active primary cancer at the time of t-AML diagnosis in a subset of patients. Although we could not demonstrate statistical significance owing to the small sample, the data suggest that t-AML-t(8;21) may have a higher frequency of KIT mutations, an adverse prognostic indicator. In addition, cases of t-AML-t(8;21) showed significant morphologic evidence of dysplasia in almost all cases assessed, indicating the likelihood of antecedent biologic alterations. It is noteworthy that de novo AML-t(8;21) seems to be a heterogeneous disease group, with a poor survival in a subset of patients.7 It is possible that the underlying mechanisms that lead to the unfavorable prognosis in this subset are shared by t-AML and contribute to the poor survival of affected patients. Nevertheless, the current findings indicate that t-AML in general has a poor prognosis35 compared with de novo AML, even when they share the same cytogenetic alterations and exhibit morphologic and immunophenotypic features. Future studies may shed more light on the underlying biology.
. Acute myeloid leukemia with recurrent genetic abnormalities. In: SwerdlowSH, CampoE, HarrisNL, et al.(eds). WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. 4th ed. Lyon, France: IARC Press; 2008:110–111.
. Disease features in acute myeloid leukemia with t(8;21)(q22;q22): influence of age, secondary karyotype abnormalities, CD19 status, and extramedullary leukemia on survival. Leuk Lymphoma. 2000;40:67–77.
. RAS mutations are rare events in Philadelphia chromosome–negative/bcr gene rearrangement–negative chronic myelogenous leukemia, but are prevalent in chronic myelomonocytic leukemia. Blood. 1990;76:1214–1219.
. Phenotypical characteristics of acute myelocytic leukemia associated with the t(8;21)(q22;q22) chromosomal abnormality: frequent expression of immature B-cell antigen CD19 together with stem cell antigen CD34. Blood. 1992;80:470–477.
Steven A.Gustafson, PeiLin, Su S.Chen, LeiChen, Lynne V.Abruzzo, RajyalakshmiLuthra, L. JeffreyMedeiros, Sa A.WangAm J Clin Pathol(2009)131 (5):
647-655DOI: http://dx.doi.org/10.1309/AJCP5ETHDXO6NCGZFirst published online: 1 May 2009 (9 pages)