A threshold of 10% for myeloperoxidase by flow cytometry is valid to classify acute leukemia of ambiguous and myeloid origin†
How to cite this article: Van den Ancker W, Westers TM, de Leeuw DC, van der Veeken YFCM, Loonen A, van Beckhoven E, Ossenkoppele GJ, van de Loosdrecht AA. A threshold of 10% for myeloperoxidase by flow cytometry is valid to classify acute leukemia of ambiguous and myeloid origin. Cytometry Part B 2013; 84B: 114–118.
According to WHO 2008 guidelines, an important role is designated for cytoplasmic myeloperoxidase (cMPO) as measured by flow cytometry for classifying acute leukemia of myeloid or ambiguous origin (AML or MPAL). However, no threshold with respect to expression level and percentage positive cells is provided. Since the expression of solely cMPO can change the diagnosis from acute lymphoid leukemia into MPAL in the current WHO 2008, a consensus is needed for the cut-off for cMPO.
In this study, we investigated whether or not a cut-off of 10% positivity for cMPO equally defines an acute leukemia as AML or MPAL as compared to a cut-off for cMPO of 20% and compared this with results obtained for Sudan Black B (SBB) staining by cytomorphology.
Cell lineage-defining markers and SBB staining were analyzed retrospectively in a cohort of 198 patients who presented with acute leukemia. Eight patients were positive for SBB (>3%), but were considered negative for cMPO (<10%); six patients were negative for SBB (≤3%) and positive for cMPO (≥10%) staining. In six patients, we found 10–20% cMPO positive leukemic cells. Five of these cases were SBB positive; the sixth patient showed a clear myeloid phenotype without positivity of any lymphoid marker. Using a 10% cut-off instead of 20% would have changed diagnosis from ALL into MPAL in two patients; both cases were SBB positive by morphology.
We conclude that a 10% cut-off is a secure lower limit for cMPO expression and can be used independently from SBB expression. © 2013 International Clinical Cytometry Society
Acute leukemias (AL) are classified by flow cytometry according to their commitment to either the myeloid or lymphoid lineage (1). In about 4% of AL, blast cells express both myeloid and lymphoid markers directing the diagnosis toward a mixed phenotype AL (MPAL) rather than acute myeloid leukemia (AML) or acute lymphoid leukemia (ALL) (2). Based on WHO 2008, a MPAL has to fulfill criteria based on expression of immunological markers that assign a blast cell to at least two lineages (1). In case no directive conclusions can be drawn from molecular diagnostics, cytogenetics, and morphology, flow cytometry has a decisive role when leukemic cells have both lymphoid and myeloid features as described by Bain et al. (3). Expression of solely CD3 and flow cytometrically detected cytoplasmic myeloperoxidase (cMPO) is sufficient to assign a blast cell to the T cell and the myeloid lineage, respectively; for the B cell lineage one particular B cell marker in addition to expression of CD19 is required (4). Cell lineage specific markers CD19 and CD3 should be highly expressed on blasts; however, for cMPO no thresholds with respect to expression level or percentage positive cells are provided. Since the expression of cMPO as an isolated marker can change the diagnosis from ALL into MPAL in the current WHO 2008, a consensus is needed for the cut-off for cMPO.
The former WHO 2001 criteria clearly described a cut-off for positivity of a marker of 20% of the whole blast population as compared to the appropriate isotype control. The 20% threshold was formerly defined based on one-color flow cytometry; this higher cut-off was necessary to exclude false positivity due to e.g. contamination of non-leukemic cells. With the development of multicolor flow cytometry and because of the high degree of specificity of cMPO, cTdT, and cCD79a, a lower cut-off of 10% was implemented when positivity was confirmed by light microscopy (EGIL criteria); in clinical practice, this is usually assessed by Sudan Black B (SBB) staining on blood and bone marrow smears. In cytomorphology, SBB representing cMPO expression is considered the hallmark of the myeloid lineage; the presence of more than 3% SBB positive leukemic progenitor cells is diagnostic for myeloid leukemias (5).
In this study, we investigated whether or not a cut-off of ≥10% positivity for cMPO by flow cytometry equally defines an AL as AML or MPAL as compared to a cut-off for cMPO of ≥20%; these results were compared with results obtained with SBB staining.
MATERIAL AND METHODS
Patient Samples and Standard Diagnostic Procedures
Cell lineage-defining markers and SBB staining were analyzed retrospectively in a cohort of 198 patients who presented with AL in our hospital between 2001 and 2009; part of these patients was discussed in Ref. 2. Informed consent was obtained from all patients in accordance with the Declaration of Helsinki and institutional guidelines. Only those patients were selected in which cytochemistry and immunophenotyping were performed on the same samples drawn at equal time points. Diagnostic procedures were performed according to standard operating procedures for classifying AL as described previously (2,3). For bone marrow aspirates, first needle aspiration was used for morphology; second and third needle aspirations were used for immunophenotypic, molecular, and cytogenetic analysis. SBB and May–Grünwald–Giemsa (MGG) staining were performed on blood and bone marrow smears by standard morphologic and cytochemical procedures. All patient samples were screened for t(9;22), t(8;21), inv(16), t(15;17), and FLT-3ITD by routine molecular and cytogenetic diagnostics (6).
Flow Cytometric Analysis
Blasts cells were analyzed for expression levels of B cell, T cell, and myeloid defining- and specific-markers with Fluorescein isothiocyanate (FITC) or Phycoerythrin (PE)-conjugated antibodies with ubiquitous CD45-Peridinin-chlorophyll-protein complex (PerCP) and CD34-Allophycocyanin (APC) by performing a four-color flow cytometric analysis on a FACS Calibur flow cytometer (BD Biosciences, San Jose, CA), as described earlier (2,7). Regarding cMPO expression, we compared the impact of previously described cut-offs levels of 10 and 20%. Results were analyzed using CellQuestPro software (BD Biosciences). Percentage of marker positivity of blast cells and mean fluorescence index (MFI) were calculated as compared to appropriate isotype controls.
By construction of receiver operating characteristic (ROC) curves using the SPSS 15.0 software the specificity and sensitivity of various cut-offs for cMPO were compared with SBB results.
In 121 out of 198 (61%) of AL in our cohort, cMPO expression was found to be higher than 20% (Table 1). These cases were classified as AML (n = 115) or MPAL (n = 6), the latter depending on co-expression of lymphoid markers. The vast majority of patients (116/121, (96%) showed no discrepancies in SBB and cMPO staining (Table 1). By applying a cut-off of ≥10%, six more cases are recognized as cMPO positive (UPN 1–6, 3% of cases, Table 2a). Two of these patients should be reclassified as MPAL instead of B-ALL or T-ALL according to their immunophenotypic profile (UPN 1 and 2; simultaneous expression of lymphoid markers and cMPO≥10%). Notably, both patients scored positive for SBB and were thereby classified as AML by morphology. Cytomorphology classified UPN 3–6 as myelomonocytic or monoblastic and monocytic leukemias (former FAB M4 and M5, respectively). Of note, UPN4 was classified as AML with inv(16) which is classified as an AML with genetic abnormalities according to the WHO 2008. These leukemias often show a low to intermediate SBB and cMPO expression; thus, changing cMPO threshold did not affect the diagnosis.
Table 1. Comparison of Sudan Black B and Flow Cytometric Myeloperoxidase Staining
Table 2a. Marker Expression of Acute Leukemias with 10–20% cMPO
|B cell markers|| || || || || || |
|T cell markers|| || || || || || |
|Myeloid markers|| || || || || || |
| MFI cMPO||9.5||4||8.9||9.3||3.1||5.1|
|Cytohistochemical|| || || || || || |
| Histomorphological|| || || || || || |
|Conclusion|| || || || || || |
|WHO2008 10% cMPO||MPAL T/myeloid NOS||MPAL with t(9;22) (b2a2)||Acute monoblastic and monocytic leukemia||AML with inv(16)||Acute monoblastic and monocytic leukemia||Acute monoblastic and monocytic leukemia|
|WHO2008 20% cMPO||T-ALL||B-ALL||Acute monoblastic and monocytic leukemia||AML with inv(16)||Acute monoblastic and monocytic leukemia||Acute monoblastic and monocytic leukemia|
Regarding discordant results, eight patients (UPN 7–14, Table 2b) were positive for SBB (>3%), but negative for cMPO (<10%). In only a minority of patients (n = 6), we found negative SBB (≤3%) and positive cMPO (≥10%) (UPN3 and UPN 15–19, Table 2c, 3% of the cohort). These discrepancies did not affect the overall classification. Two examples of SBB and cMPO discordance are shown in Supporting Information Figure 1.
Table 2b. Marker Expression of Acute Leukemia with Positive SBB and Negative cMPO (<10%)
|B cell markers|| || || || || || || || |
| CD79a||20||0||5||0||8||2|| ||2|
|T cell markers|| || || || || || || || |
|Myeloid markers|| || || || || || || || |
|Cytohistochemical|| || || || || || || || |
| Histomorphological|| || || || || || || || |
|Conclusion|| || || || || || || || |
|WHO2008 10% cMPO||AL*t (AL with (9;22) (b2a2))||Acute Myelo- monocytic Leukemia||Acute monoblastic and monocytic leukemia||Acute monoblastic and monocytic leukemia||AML with minimal differentiation||AML with differentiation||Acute erythroid leukemia||Acute monoblastic and monocytic leukemia|
|WHO2008 20% cMPO||AL*||Acute Myelo- monocytic Leukemia||Acute monoblastic and monocytic leukemia||Acute monoblastic and monocytic leukemia||AML with minimal differentiation||AML with differentiation||Acute erythroid leukemia||Acute monoblastic and monocytic leukemia|
Table 2c. Marker Expression of Acute Leukemia with Negative SBB and Positive cMPO (≥20%)
|B cell markers|| || || || || |
|T cell markers|| || || || || |
| cCD3||39||48|| ||2||0|
|Myeloid markers|| || || || || |
| MFI cMPO||12.6||4.7||17.4||3||15.7|
|Histomorphological|| || || || || |
|Conclusion|| || || || || |
|WHO2008 10% cMPO||MPAL T/myeloid, NOS||MPAL T/myeloid, NOS||MPAL with t(9;22)e1a2||MPAL B/myeloid, NOS||AML with MDS-related changes|
|WHO2008 20% cMPO||MPAL T/myeloid, NOS||MPAL T/myeloid, NOS||MPAL with t(9;22)e1a2||MPAL B/myeloid, NOS||AML with MDS-related changes|
Receiver operating curve was used to compare cMPO results with SBB positivity or negativity (either >3% or below); When choosing a 20% cut-off a sensitivity of 87% with a specificity of 92% was reached. However, when choosing a cut-off of 10% a slightly more favorable ratio is found: a sensitivity of 95% and specificity of 87% (data not shown); the area under the curve was 0.959 (P < 0.001).
Within the current WHO2008 classification of hematopoietic neoplasms, cMPO is considered the hallmark protein for defining the myeloid lineage. In this regard, we compared the effect of a cut-off of 10 and 20% for cMPO in diagnosing AL and compared the results with SBB percentages. By applying a cut-off of 10% for cMPO two patients would have been reclassified from ALL to MPAL based on the expression of solely cMPO. Interestingly, both patients scored positive for SBB and were thereby classified as AML by morphology.
It might be more appropriate to use comparable cut-offs for cMPO and SBB. This cut-off should be high enough to ensure that the subpopulation is of malignant origin, since normal myeloid progenitor cells (CD34+ and cMPO+) might reside in the same CD45dim blast compartment. In CD34− AML it has been shown that up to 1.0% non-malignant CD34+ progenitors can be found (8). For other AL, the percentage of normal myeloid progenitor cells has to be revealed by further studies and should direct to the SBB and cMPO cut-off of choice. Overall, our data support to a cut-off of 10% for cMPO by flow cytometric analysis, independently of results found for SBB staining. Furthermore, by applying a cut-off of 10% a sensitivity of 95% and specificity of 87% is reached, compared with 87 and 92% when using a 20% cut-off.
Rarely in cases of ALL, SBB can stain intracellular lipoid inclusions not containing cMPO (9). In our cohort, eight patients were positive for SBB and negative for cMPO (<10%). Of note, these AL were considered to be of myeloid origin according to morphology and their immunophenotypic profile. Thus, discussion on cMPO positivity would not have had diagnostic consequences in these cases.
In 3% of patients discrepancies caused by negative SBB (3≤%) and positive cMPO (≥10%) were seen. However, cMPO was expressed at a very low intensity. It is known that in case of AML without and AML with minimal differentiation SBB can be below 3% by light microscopy while cMPO may be positive by flow cytometry (1,10). Yet, four out of five of these particular cases were classified as MPAL based on their immunophenotypic profile, one patient (UPN 19) was classified as AML with MDS-related changes.
In summary, a cut-off of 10% for cMPO expression would have changed diagnostic classification in only a minority of cases (1%); both were SBB positive by morphology. Therefore, we conclude that a 10% cut-off is a secure lower limit for cMPO expression and can be used independently from SBB expression. Clinical consequences have to be prospectively investigated.
The authors thank Guus Westra for expert technical assistance.