Abstract
Because of severe bleeding complications, patients with acute promyelocytic leukemia (APL) have to be treated with all-trans retinoic acid immediately following diagnosis. In addition to morphology, flow cytometry contributes to a rapid detection of APL according to phenotypic characteristics of leukemic cells. In some patients, these analyses are inconclusive or even contradictory to diagnosis. Previously, we showed the clinical and functional impact of class II–associated invariant chain peptide (CLIP) in acute myeloid leukemia (AML). This study focuses on the analysis of CLIP expression on leukemic cells to characterize HLA-DR–negative AML, including APL. We demonstrate exclusive and significant CLIP expression in all cases of typical and variant APL, as compared to other HLA-DR–negative non–APL-type AML. CLIP appears to be a highly sensitive and specific flow cytometric marker, resolving discrepant identification of both genetic subgroups. Our findings show the additive value of CLIP analysis for a fast and unequivocal recognition of APL by flow cytometry in conjunction with morphology.
Acute promyelocytic leukemia (APL) is characterized by excessive proliferation of abnormal promyelocytes bearing t(15;17)(q22;q12), which results in the expression of the PML-RARα fusion transcript.1 Immediate recognition of this acute myeloid leukemia (AML) subtype is of major importance, as it is associated with coagulation disturbances leading to early death2; rapid treatment with all-trans retinoic acid is required to control or prevent these severe complications and improve the prognosis of APL patients.3, 4 At present, classical morphology showing bundles of Auer rods (ie, elongated aggregates of intracellular material found only in acute myeloid leukemia blasts) is the most commonly used, fast approach for detection before the confirmation of PML-RARα gene rearrangement for diagnosis of APL. It can also assist in the classification of hypergranular (typical) and hypogranular (variant) subforms of APL, according to the World Health Organization criteria.5 In variant APL cases, the leukocyte count is usually increased and is associated with a worse clinical outcome, as compared to typical APL cases.6, 7 In some patients and in less experienced hands, however, morphological examination is inconclusive and could lead to a delayed or even a false-negative or false-positive recognition of APL.
Another sensitive technique to identify APL is multicolor flow cytometry, which provides distinctive immunophenotypes of leukemic promyelocytes. For both typical and variant APL, bright and homogeneous expression of MPO, CD33, and CD117, heterogeneous expression of CD13, and low expression or absence of CD34, HLA-DR, CD11b, and CD15 are classically found.5, 8 Still, there are examples of HLA-DR–negative non–APL-type patients that display a similar immunophenotype. CD56 expression has been described for APL, but only for ∼20% of cases, with low specificity.9, 10 Increased side scatter (SSC) is only a feature of typical APL,5 whereas the expression of CD2 is mostly restricted to variant APL.11, 12
HLA-DR is a molecule that plays a major role in HLA class II antigen presentation; in our previous studies regarding this issue in AML, APL patients were excluded because of their low or absent HLA-DR expression.13, 14 One of the key mediators of HLA class II antigen presentation is the invariant chain. A small remnant of this protein, the class II–associated invariant chain peptide (CLIP), can be presented by HLA-DR complexes on the surface of leukemic blasts, resulting in decreased immunogenicity.15 Interestingly, in APL cases, we observed that CLIP was abundantly expressed at the surface of HLA-DR–negative leukemic promyelocytes. Here, we studied the application of the analysis of CLIP expression in the discrimination of HLA-DR–negative AML subgroups, including APL, by flow cytometry.
Materials and Methods
Patients
From a cohort of 297 patients who subsequently presented with AML between 2004 and 2010, 16 cases were diagnosed with APL (5.4%) and 23 with HLA-DR–negative non–APL-type AML (7.7%). HLA-DR–negative non–APL-type AML patients were defined as cases with less than 20% of HLA-DR–expressing myeloid blasts. Blood and bone marrow samples were collected after obtaining informed consent as part of routine diagnostic procedures at our department and by the rules of our institute and the Helsinki Declaration. All patients were classified according to the World Health Organization 2001 criteria,5 and routinely screened for cytogenetics and molecular aberrancies, including t(15;17), PML-RARα transcripts (bcr1/bcr3), and internal tandem duplication of the FLT3 gene (FLT3-ITD). Cytogenetic risk groups were defined as favorable [t(8;21), t(15;17), or inv(16)], standard (neither favorable nor adverse), or adverse [complex karyotype, −5 or −7, del(5q), abnormality 3q or 11q23], as previously described.16
Flow Cytometry
The following monoclonal antibodies were used for immunophenotyping of AML cases: PE-conjugated c-MPO (Dako, Glostrup, Denmark), CD13 (BD Biosciences, San Jose, CA), CD33 (BD Biosciences), CD117 (BD Biosciences), CD11b (BD Biosciences), CD56 (BD), anti-CXCR4 (12G5 clone; BD PharMingen, San Diego, CA), and anti-CLIP (cerCLIP.1 clone; Santa Cruz Biotechnology, Santa Cruz, CA); FITC-conjugated CD15 (BD Biosciences), CD11a (BD Biosciences), CD18 (BD Biosciences), CD2 (Dako), and anti–HLA-DR (L243; BD Biosciences); PerCP-conjugated CD45 (BD Biosciences); and APC-conjugated CD34 (BD Biosciences). Staining of cells was performed by following local standard operating protocols. Leukemic cells were analyzed by gating for the immature myeloid cell compartment, defined as CD45dim/SSClow-int; in this compartment, the percentage of marker-positive cells was assessed as compared to unstained cells. For comparison, we also investigated normal promyelocytes (SSChigh/CD34−/CD117+) from bone marrow aspirates of four different healthy donors. Expression data were acquired on a FACSCalibur flow cytometer and analyzed by using CellQuestPro software (both BD Biosciences).
Statistical Analysis
We compared the differences in marker expression between patient groups by the Mann-Whitney U-test. We regarded P ≤ 0.05 as significant.
Results and Discussion
The main clinical and immunophenotypic characteristics of APL and HLA-DR–negative non–APL-type AML patients are demonstrated in Table 1. All of the patients with APL harbored t(15;17)(q22;q12), whereas no rare variants such as t(11;17)(q23;q21) or t(5;17)(q35;q21) were encountered. None of the HLA-DR–negative non–APL cases had cytogenetic abnormalities. There were no significant differences in age, sex, or FLT3-ITD (Table 1). Strikingly, FLT3-ITD was present in all variant APL cases, and in none of the typical APL cases (Table 2). As expected, standard immunophenotypic analysis of leukemic cells in APL showed significantly increased SSC and CD2 expression, and decreased expression of CD11b and CD56, as compared to HLA-DR–negative non–APL-type AML (Table 1). Also, consistent with previous reports,6, 11, 17 we found a significantly low SSC and high expression of CD34, CD2, and CD56 in variant APL, in comparison with typical APL (Table 2).
Table 1.
HLA-DR–Negative AML
| Clinical parameters | APL⁎ | Non-APL† | P value |
|---|---|---|---|
| Number of patients | 16 | 23 | |
| Male/female | 6/10 | 13/10 | |
| Age in years at diagnosis, mean | 45 (21–68) | 57 (22–75) | 0.0079 |
| WHO 2001 classification | |||
| AML with recurrent genetic abnormalities | |||
| AML with t(15;17)(q22;q12) | 16 (100) | 0 | |
| AML with multilineage dysplasia | 0 (0) | 4 (17) | |
| AML not otherwise categorized | |||
| AML without maturation | 0 (0) | 5 (22) | |
| AML with maturation | 0 (0) | 7 (30) | |
| Acute monoblastic and monocytic leukemia | 0 (0) | 1 (4) | |
| Acute megakaryoblastic leukemia | 0 (0) | 2 (9) | |
| AML of ambiguous lineage | |||
| Biphenotypic acute leukemia | 0 (0) | 1 (3) | |
| AML (not classifiable) | 0 (0) | 3 (13) | |
| FLT3-ITD | |||
| Normal | 8 (50) | 11 (48) | |
| Heterozygous | 8 (50) | 9 (39) | |
| Not done | 0 (0) | 3 (13) | |
| Cytogenetic risk group | |||
| Favorable | 16 (100) | 0 (0) | |
| Standard | 0 (0) | 15 (65) | |
| Adverse | 0 (0) | 0 (0) | |
| Not applicable | 0 (0) | 8 (35) | |
| Immunophenotype, mean‡ | |||
| SSC (ratio)§ | 3.2 (1.6–4.5) | 1.9 (1–3.5) | 0.0003 |
| CD34 (%) | 27 (0–90) | 26 (0–99) | 0.3301 |
| c-MPO (%) | 94 (84–99) | 76 (0–99) | 0.6473 |
| HLA-DR (%) | 10 (0–44) | 4 (0–17) | 0.5191 |
| CD13 (%) | 89 (55–99) | 61 (1–99) | 0.1228 |
| CD33 (%) | 94 (37–99) | 91 (33–99) | 0.5826 |
| CD117 (%) | 74 (21–97) | 75 (1–99) | 0.5392 |
| CD11b (%) | 6 (1–33) | 21 (0–86) | 0.0024 |
| CD15 (%) | 3 (0–11) | 10 (0–99) | 0.8078 |
| CD2 (%) | 25 (0–78) | 2 (0–25) | 0.0005 |
| CD56 (%) | 10 (0–64) | 38 (0–99) | 0.0079 |
| CXCR4 (%) | 22 (1–97) | 48 (8–100) | 0.0033 |
| CLIP (%) | 90 (82–98) | 8 (0–73) | <0.0001 |
Clinical parameters were not significantly different between groups unless otherwise indicated. P values in italics are significant.
AML, acute myeloid leukemia; APL, acute promyelocytic leukemia; FLT3-ITD, internal tandem duplication of the FLT3 gene; SSC, side scatter; WHO, World Health Organization.
Classification according to morphology and presence of PML-RARα (5).
PML-RARα–negative AML patients with less than 20% myeloid cells expressing HLA-DR.
Numbers in parentheses represent the percentage or range of each variable within the patient group.
Median SSC ratio between immature myeloid cells and CD45bright lymphocytes in the same sample.
Table 2.
Typical and Variant APL
| Clinical parameters | Typical APL | Variant APL⁎†‡ | P value |
|---|---|---|---|
| Number of patients | 8 | 8 | |
| Male/female | 3/5 | 5/3 | |
| Age in years at diagnosis, mean (range) | 43 (21–65) | 47 (37–68) | |
| t(15;17) | 8 | 8 | |
| PML-RARα transcript | |||
| bcr1 | 6 (75) | 0 (0) | |
| bcr3 | 1 (13) | 6 (75) | |
| Not done | 1 (13) | 2 (25) | |
| FLT3-ITD, heterozygous | 0 | 8 | |
| Immunophenotype, mean† | |||
| SSC (ratio)‡ | 3.7 (2.6–4.5) | 2.6 (1.6–3.4) | 0.0281 |
| CD34 (%) | 5 (0–17) | 48 (2–90) | 0.0011 |
| c-MPO (%) | 94 (81–98) | 95 (84–99) | 0.3823 |
| HLA-DR (%) | 1 (0–5) | 19 (1–44) | 0.0030 |
| CD13 (%) | 88 (55–99) | 89 (68–99) | 0.9591 |
| CD33 (%) | 97 (94–99) | 91 (37–99) | 0.7984 |
| CD117 (%) | 82 (64–97) | 74 (51–95) | 0.5054 |
| CD11b (%) | 3 (1–5) | 8 (2–33) | 0.1949 |
| CD15 (%) | 4 (1–11) | 2 (0–4) | 0.3823 |
| CD2 (%) | 2 (0–5) | 48 (10–68) | 0.0002 |
| CD56 (%) | 2 (0–8) | 19 (1–64) | 0.0148 |
| CXCR4 (%) | 17 (1–60) | 27 (4–97) | 0.5728 |
| CLIP (%) | 89 (83–98) | 91 (82–98) | 0.3823 |
Genetic aberrancies were determined according to routine cytogenetics and molecular analyses. P values in italics are significant.
APL, acute promyelocytic leukemia; bcr, breakpoint cluster region, an isoform of PML-RARα; FLT3-ITD, internal tandem duplication of the FLT3 gene; PML, promyelocytic leukemia; RAR, retinoic acid receptor; SSC, side scatter.
Classification of APL subforms according to morphology.5
Numbers in brackets represent the percentage or range of each variable within the patient group.
Median SSC ratio between immature myeloid cells and CD45bright lymphocytes in the same sample.
In search of new immunophenotypic markers to distinguish APL from HLA-DR–negative non–APL-type AML, researchers recently described reduced CXCR4 expression on leukemic promyelocytes,18 which we were able to confirm in our patient cohort (P = 0.0042, Table 1; 69% sensitivity and 74% specificity). Furthermore, as previously shown,17 reduced expression of CD11a and CD18 was seen in our cohort of APL cases, as compared to non-APL cases (data not shown). Thus, decreased expression of certain markers is prevalently associated with APL. In contrast, we observed a significantly high expression of CLIP on all leukemic promyelocytes (both typical and variant subforms) when compared to HLA-DR–negative non–APL-type blasts (means of 90% ± 6% and 8% ± 17%, respectively; P < 0.0001, Table 1). Only in 2 of 23 HLA-DR–negative non-APL cases was an increased CLIP expression of 44% and 73% seen; nevertheless, all APL cases revealed a CLIP expression of at least 82% (Figure 1A). The large mean difference in expression implicates a potential role of CLIP to separate APL from other HLA-DR–negative AML patients. Using CLIP as an additional marker, a correct classification of APL by flow cytometry was obtained for all cases (100% sensitivity and 95% specificity). Despite the significant differences in HLA-DR (P = 0.0030), CLIP expression was equally high between typical and variant APL subforms (Figure 1A). In control experiments, normal promyelocytes from healthy donors revealed only very low surface CLIP expression (mean 3% ± 0.3%, P = 0.0029 as compared to APL, Figure 1A and data not shown). Normally, CLIP associates with HLA class II molecules, but we observed that in APL, the absence of HLA-DR was accompanied by a remarkable lack of all HLA class II isoforms (n = 6, data not shown). The underlying mechanism for this HLA class II–independent CLIP presentation is currently not understood and is under investigation.
Figure 1.
The use of CLIP for discrimination of APL and HLA-DR–negative non-APL by flow cytometry. A: Overview of CD34, HLA-DR, and CLIP expression on leukemic cells from patients diagnosed with typical APL (APL, n = 8), variant APL (APLv, n = 8), or HLA-DR–negative non–APL-type AML (non-APL, n = 23). We assessed CD45dim-gated cells using flow cytometry. Normal promyelocytes (defined as CD34 −CD117+ cells and depicted as normal, n = 4) from healthy bone marrow were also analyzed. B: Representative examples of patients whose diagnosis is not conclusive according to standard flow cytometric analysis. Patients showed the following immunophenotypes: MPO+, CD13+, CD33+, CD117+, CD11b−, CD15−, CD2−, and CD56−. Dotplots are shown for the forward scatter (FCS) versus side scatter (SSC), and CD34 versus HLA-DR and CD34 versus CLIP expression. For each patient, the latter two plots are displayed for CD45dim-gated cells. Final diagnosis of these patients is depicted and made according to morphology and analysis of the PML-RARα gene. Note that patient 2 has increased SSC of leukemic cells, a characteristic more frequently observed in patients with typical APL.
To illustrate the additive value of CLIP for the flow cytometric recognition of APL, we will discuss three representative AML cases (Figure 1B). Leukemic cells of these cases revealed a similar expression pattern of routinely analyzed markers; ie, high levels of MPO, CD117, and CD33, heterogeneous levels of CD13, low levels of CD11b, CD15, CD2, and CD56 (data not shown), and absence of CD34 and HLA-DR (Figure 1B). The increased SSC in patients 1 and 2 reflecting granular blasts or granulocytic differentiation suggests typical APL for both cases. Because the lower SSC in patient 3 could also fit the immunophenotype of the variant APL form, this case might be diagnosed with either APL or HLA-DR–negative non–APL-type AML. According to cell morphological scrutiny, patient 1 was classified as typical APL, patient 2 as HLA-DR–negative non-APL, and patient 3 as variant APL, demonstrating a discrepancy between flow cytometric and morphological analysis for patients 2 and 3. After subsequent examination of the PML-RARα gene and conventional cytogenetics, patient 1 was indeed diagnosed with typical APL, but patients 2 and 3 were negative for this translocation and thus regarded as HLA-DR–negative non–APL-type AML. Consistent with this, leukemic cells of patients 2 and 3 (non-APL) did not express CLIP in contrast to the high CLIP expression found for patient 1 (typical APL; Figure 1B). This shows that standard immunophenotyping and morphological examination can be misleading, but that incorporation of CLIP analysis can distinguish APL, including variant APL, from HLA-DR–negative AML.
Multicolor flow cytometry of surface markers on leukemic cells is considered one of the fastest and easiest-to-use techniques to identify genetic subgroups in patients with AML. For the identification of APL, progress has been made in intracellular analysis of PML-RARα fusion proteins by flow cytometry,19 but it is not yet common practice. In the present study, we introduce cell surface CLIP expression (>80%) as a positive marker in case of HLA-DR–negative AML to characterize APL by the use of flow cytometric analysis alone. Irrespective of the requirement for molecular diagnostic testing, inclusion of this marker in present AML immunophenotyping protocols may attribute to rapid identification of APL, thereby shortening the time to treatment of APL patients.
Acknowledgments
We thank Adri Zevenbergen for his expert technical assistance, as well as Pauline Merle and Hans van Oostveen regarding molecular diagnostics.
Footnotes
Supported in part by research funding from Stichting Vanderes (grant 125 to M.M.v.L.).
References
- 1.Melnick A., Licht J.D. Deconstructing a disease: rARalpha, its fusion partners, and their roles in the pathogenesis of acute promyelocytic leukemia. Blood. 1999;93:3167–3215. [PubMed] [Google Scholar]
- 2.Breccia M., Latagliata R., Cannella L., Minotti C., Meloni G., Lo-Coco F. Early hemorrhagic death before starting therapy in acute promyelocytic leukemia: association with high WBC count, late diagnosis and delayed treatment initiation. Haematologica. 2010;95:853–854. doi: 10.3324/haematol.2009.017962. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Tallman M.S., Andersen J.W., Schiffer C.A., Appelbaum F.R., Feusner J.H., Ogden A., Shepherd L., Willman C., Bloomfield C.D., Rowe J.M., Wiernik P.H. All-trans-retinoic acid in acute promyelocytic leukemia. N Engl J Med. 1997;337:1021–1028. doi: 10.1056/NEJM199710093371501. [DOI] [PubMed] [Google Scholar]
- 4.Kawai Y., Watanabe K., Kizaki M., Murata M., Kamata T., Uchida H., Moriki T., Yokoyama K., Tokuhira M., Nakajima H. Rapid improvement of coagulopathy by all-trans retinoic acid in acute promyelocytic leukemia. Am J Hematol. 1994;46:184–188. doi: 10.1002/ajh.2830460305. [DOI] [PubMed] [Google Scholar]
- 5.Vardiman J.W., Harris N.L., Brunning R.D. The World Health Organization (WHO) classification of the myeloid neoplasms. Blood. 2002;100:2292–2302. doi: 10.1182/blood-2002-04-1199. [DOI] [PubMed] [Google Scholar]
- 6.Foley R., Soamboonsrup P., Carter R.F., Benger A., Meyer R., Walker I., Wan Y., Patterson W., Orzel A., Sunisloe L., Leber B., Neame P.B. CD34-positive acute promyelocytic leukemia is associated with leukocytosis, microgranular/hypogranular morphology, expression of CD2 and bcr3 isoform. Am J Hematol. 2001;67:34–41. doi: 10.1002/ajh.1073. [DOI] [PubMed] [Google Scholar]
- 7.Lee J.J., Cho D., Chung I.J., Cho S.H., Park K.S., Park M.R., Ryang D.W., Kim H.J. CD34 expression is associated with poor clinical outcome in patients with acute promyelocytic leukemia. Am J Hematol. 2003;73:149–153. doi: 10.1002/ajh.10337. [DOI] [PubMed] [Google Scholar]
- 8.Paietta E. Expression of cell-surface antigens in acute promyelocytic leukaemia. Best Pract Res Clin Haematol. 2003;16:369–385. doi: 10.1016/s1521-6926(03)00042-2. [DOI] [PubMed] [Google Scholar]
- 9.Ferrara F., Morabito F., Martino B., Specchia G., Liso V., Nobile F., Boccuni P., Di N.R., Pane F., Annunziata M., Schiavone E.M., De S.M., Guglielmi C., Del V.L., Lo C.F. CD56 expression is an indicator of poor clinical outcome in patients with acute promyelocytic leukemia treated with simultaneous all-trans-retinoic acid and chemotherapy. J Clin Oncol. 2000;18:1295–1300. doi: 10.1200/JCO.2000.18.6.1295. [DOI] [PubMed] [Google Scholar]
- 10.Murray C.K., Estey E., Paietta E., Howard R.S., Edenfield W.J., Pierce S., Mann K.P., Bolan C., Byrd J.C. CD56 expression in acute promyelocytic leukemia: a possible indicator of poor treatment outcome? J Clin Oncol. 1999;17:293–297. doi: 10.1200/JCO.1999.17.1.293. [DOI] [PubMed] [Google Scholar]
- 11.Biondi A., Luciano A., Bassan R., Mininni D., Specchia G., Lanzi E., Castagna S., Cantu-Rajnoldi A., Liso V., Masera G., Barbui T., Rambaldi A. CD2 expression in acute promyelocytic leukemia is associated with microgranular morphology (FAB M3v) but not with any PML gene breakpoint. Leukemia. 1995;9:1461–1466. [PubMed] [Google Scholar]
- 12.Lin P., Hao S., Medeiros L.J., Estey E.H., Pierce S.A., Wang X., Glassman A.B., Bueso-Ramos C., Huh Y.O. Expression of CD2 in acute promyelocytic leukemia correlates with short form of PML-RARalpha transcripts and poorer prognosis. Am J Clin Pathol. 2004;121:402–407. doi: 10.1309/XC8P-9M8N-KQDT-38LB. [DOI] [PubMed] [Google Scholar]
- 13.Chamuleau M.E., Souwer Y., van Ham S.M., Zevenbergen A., Westers T.M., Berkhof J., Meijer C.J., van de Loosdrecht A.A., Ossenkoppele G.J. Class II-associated invariant chain peptide expression on myeloid leukemic blasts predicts poor clinical outcome. Cancer Res. 2004;64:5546–5550. doi: 10.1158/0008-5472.CAN-04-1350. [DOI] [PubMed] [Google Scholar]
- 14.van Luijn M.M., Chamuleau M.E., Thompson J.A., Ostrand-Rosenberg S., Westers T.M., Souwer Y., Ossenkoppele G.J., van Ham S.M., van de Loosdrecht A.A. Class II-associated invariant chain peptide down-modulation enhances the immunogenicity of myeloid leukemic blasts resulting in increased CD4+ T-cell responses. Haematologica. 2010;95:485–493. doi: 10.3324/haematol.2009.010595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.van Luijn M.M., van den Ancker W., Chamuleau M.E., Zevenbergen A., Westers T.M., Ossenkoppele G.J., van Ham S.M., van de Loosdrecht A.A. Absence of class II-associated invariant chain peptide on leukemic blasts of patients promotes activation of autologous leukemia-reactive CD4+ T cells. Cancer Res. 2011;71:2507–2517. doi: 10.1158/0008-5472.CAN-10-3689. [DOI] [PubMed] [Google Scholar]
- 16.Lowenberg B., Downing J.R., Burnett A. Acute myeloid leukemia. N Engl J Med. 1999;341:1051–1062. doi: 10.1056/NEJM199909303411407. [DOI] [PubMed] [Google Scholar]
- 17.Paietta E., Goloubeva O., Neuberg D., Bennett J.M., Gallagher R., Racevskis J., Dewald G., Wiernik P.H., Tallman M.S. A surrogate marker profile for PML/RAR alpha expressing acute promyelocytic leukemia and the association of immunophenotypic markers with morphologic and molecular subtypes. Cytometry B Clin Cytom. 2004;59:1–9. doi: 10.1002/cyto.b.20001. [DOI] [PubMed] [Google Scholar]
- 18.Oelschlaegel U., Mohr B., Schaich M., Schakel U., Kroschinsky F., Illmer T., Ehninger G., Thiede C. HLA-DRneg patients without acute promyelocytic leukemia show distinct immunophenotypic, genetic, molecular, and cytomorphologic characteristics compared to acute promyelocytic leukemia. Cytometry B Clin Cytom. 2009;76:321–327. doi: 10.1002/cyto.b.20475. [DOI] [PubMed] [Google Scholar]
- 19.Dimov N.D., Medeiros L.J., Kantarjian H.M., Cortes J.E., Chang K.S., Bueso-Ramos C.E., Ravandi F. Rapid and reliable confirmation of acute promyelocytic leukemia by immunofluorescence staining with an antipromyelocytic leukemia antibody: the M. D Anderson Cancer Center experience of 349 patients. Cancer. 2010;116:369–376. doi: 10.1002/cncr.24775. [DOI] [PMC free article] [PubMed] [Google Scholar]