Can threshold for MPO by flow cytometry be reduced in classifying acute leukaemia? A comparison of flow cytometric and cytochemical myeloperoxidase using different flow cytometric cut-offs Prabhu Manivannan, Vandana Puri, Venkatesan Somasundaram, Abhishek Purohit, Rahul Kumar Sharma, Mandeep Dabas, Renu Saxena Department of Hematology, All India Institute of Medical Sciences, New Delhi, India Objectives: Myeloperoxidase (MPO) detection either by enzyme cytochemistry (cMPO) or flow cytometry (fMPO) plays a major role in acute leukaemia (AL) diagnosis as per World Health Organization (WHO) 2008 classification. Although 3% cMPO was recommended as positivity, no specific cut-off had been mentioned by WHO for fMPO. Various authors recommend different cut-offs ranging from 3 to 28% for fMPO. The aim of this study was to analyse fMPO cut-offs ranging from 3 to 10% in classifying AL and to assess whether a new cut-off could be suggested. Methods: Totally, 216 cases of AL were retrospectively analysed for fMPO ranging from 3 to 10% and compared with gold standard. Presence of cMPO (≥3%) and/or expression of two or more pan-myeloid markers (CD13, CD33, and CD117) in the absence of CD19 and CD3 were kept as gold standard for diagnosis of acute myeloid leukaemia (AML). Results: Sensitivities for classifying AL as AML/mixed phenotypic acute leukaemia (MPAL) at 3, 5.4, and 10% were 98.3, 98.3, and 96.6%, respectively, whereas specificities at this cut-off were 22.2, 91, and 71%, respectively. Discussion: Only few studies have been done in this aspect to define a consistent cut-off for fMPO for proper classification of acute leukaemias. This was one of the largest and few studies available till date in this regard. Conclusion: The newer cut-off for fMPO (5.4%) emerged out from our study with best sensitivity and specificity for accurately classifying AL cases into acute lymphoblastic leukaemia, AML, and MPAL. Keywords: MPO – myeloperoxidase, ALL – acute lymphoblastic leukaemia, AML – acute myeloid leukaemia, MPAL – mixed phenotypic acute leukaemia, FCM – flow cytometry

Introduction Acute leukaemias (ALs) are classified into three main categories namely acute myeloid leukaemia (AML), acute lymphoblastic leukaemia (ALL), and acute leukaemia of ambiguous lineage, which further includes mixed phenotype acute leukaemia (MPAL) and acute undifferentiated leukaemia (AUL).1 Myeloperoxidase (MPO) detection either by enzyme cytochemistry or flow cytometry (FCM) plays a major role for this sub-classification as per World Health Organization (WHO) 2008 classification on haematopoietic and lymphoid malignancies.2 The cut-off of 3% for cytochemical MPO (cMPO) as positive was recommended Correspondence to: Vandana Puri, Department of Hematology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi 110068, India. Email: [email protected]

© W. S. Maney & Son Ltd 2015 DOI 10.1179/1607845414Y.0000000223

by the French–American–British (FAB) classification and has been retained in WHO 2008 classification.3 The European Group for the Immunological Classification of Leukaemias (EGIL) had recommended 10% cut-off as positive for flow cytometric MPO (fMPO), but only few studies have validated this issue so far.4,5 Different cut-off for fMPO had been recommended by various authors ranging from 3 to 28%.6–10 However, this issue of fMPO cut-off was not dealt with in detail in WHO 2008 classification.2 The aim of this study was to compare different cut-offs such as 3, 5, and 10% for fMPO with ≥3% cut-off for cMPO and/or expression of ≥2 pan-myeloid markers in appropriately classifying AL and assess whether a new threshold could be suggested.

Hematology

2015

VOL.

20

NO.

8

455

Manivannan et al.

Comparative study of flow cytometric and cytochemical myeloperoxidase

Materials and methods This was a retrospective observational study done in the Department of Hematology at a tertiary care centre in North India. It included review of all AL cases retrieved between January 2013 and December 2013. In all cases, peripheral smear, bone marrow aspirate (BMA) smears, bone marrow biopsies, as well as immunophenotyping data from FCM were retrieved from our records.

out any residual myeloid cells falling in the blast window. Myeloid lineage was considered when two or more of the pan-myeloid markers show positivity (CD13, CD33, and CD117) in the absence of other lineage specific markers (CD19 and CD3). fMPO was not considered as myeloid identification criteria, because it was assessed in the study.

Discordant cases Enzyme cytochemistry The special cytochemical stains such as myeloperoxidase (cMPO), Sudan black B (SBB), non-specific esterase (NSE), and periodic acid Schiff (PAS), in addition to routine Jenner–Giemsa stain were performed on BMA smears in all cases. In all the cases, smears were benzidine-stained (cMPO) and a minimum of 100 blasts were screened for the presence of cytoplasmic granularity.11 The myeloid lineage was considered when ≥3% of the blasts were showing positivity for cMPO.

Flow cytometry Ethylene diamine tetra acetic acid (EDTA) anti-coagulated BMA samples were processed by direct Stain-Lyse-Wash method. All processed samples were acquired by BD FACS canto (six colour two LASER) and analysed by FACS-Diva software. In all AL cases, minimum 10 000 events were acquired. In all these cases, samples were processed and acquired within 24 hours of collection. Anti-MPO antibody labelled with fluorescein isothiocyanate (FITC) (clone 5B8, Ms IgG1) (BD-Biosciences), panmyeloid markers (CD13-PE, CD33-APC, and CD117-APC), immaturity marker (CD34-PE-Cy7), cell-lineage-specific antibodies (CD19 and CD3), and other markers, such as HLA-DR, CD11c, and CD64, were performed and analysed. These markers were considered positive when more than 20% of the blasts were stained, except for fMPO, which was variable under consideration. Each tube was preceded by a control tube containing unstained patient cells for evaluating auto-fluorescence of monoclonal antibody reagents used in individual patient specimens. The blast population was defined by low or intermediate side scatter (SSC) and low to negative CD45 expression (CD45low/ SSClow) and residual lymphocytes were gated on a CD45 versus SSC dot-plot as CD45high/SSClow. For internal negative control to fMPO, we assessed unstained tube as well as residual lymphocytes. For internal positive control to fMPO, residual granulocytes were used. fMPO positivity was determined on the basis of quadrant placed in unstained tube to delineate all possible auto-fluorescence level, and back gating was also done on MPO positive cells to rule

456

Hematology

2015

VOL.

20

NO.

8

In this study, three positive cut-off thresholds for fMPO were evaluated: 10% or more of the blasts as per EGIL, 5% or more, and 3% or more of the blast. We kept cMPO (≥3%) and/or expression of two or more pan-myeloid markers (CD13, CD33, and CD117) in the absence of CD19 and CD3 as gold standard for diagnosis of AML. Using these different thresholds for both methods, all the cases, that were positive by fMPO and negative by cMPO, or vice versa, were considered as discordant. If there was a discrepancy between cMPO and fMPO, stains were repeated and reviewed both on cytochemistry and FCM. Finally, all these AL cases were sub-categorised as per WHO 2008/FAB classification into ALL, AML, and MPAL/AUL.

Treatment details Cases diagnosed as ALL were given therapy based on BFM-90 protocol as per the risk stratification, while AML and MPAL were treated with 3 + 7 induction followed by three cycles of consolidation chemotherapy with high dose cytarabine. A single case of CML with myeloid blast crisis was treated like AML along with tyrosine kinase inhibitors (imatinib).

Statistical analysis Receiver operator curve (ROC) analysis was done to identify the best cut-off to differentiate AML, ALL, and MPAL using SPSS software.

Results A total of 216 AL cases were retrieved in the abovementioned period. Median age was 28 years (range: 2.5–81 years) and male to female ratio was 1.8:1. The various subtypes of AL were highlighted in Fig. 1. The FAB classification was done for AML in 135 cases. In one AML case, BMA smears were diluted, so that cMPO could not be performed, and this case was excluded from the study for further statistical analysis. Three cases had features of AML with myelodysplasia-related changes. In 58 and 14 cases, diagnosis of B-ALL and T-ALL was made based on CD19 and cytoplasmic CD3 (cCD3) positivity respectively on FCM. In four cases, diagnosis of MPAL was made as per WHO 2008 classification. Preliminary comparison of cMPO and fMPO was

Manivannan et al.

Comparative study of flow cytometric and cytochemical myeloperoxidase

Figure 1 Distribution of various subtypes of acute leukaemia. Footnote: AML – acute myeloid leukaemia, MDS – myelodysplastic syndrome, ALL – acute lymphoblastic leukaemia, CML – chronic myeloid leukaemia, and MPAL – mixed phenotypic acute leukaemia.

depicted in Table 1. Subsequently, ROC analysis of all these cases with 3, 5, and 10% cut-off for fMPO was done and compared against the gold standard as shown in Fig. 2.

A 3% cut-off for MPO staining by FCM Totally 27 cases were discrepant between cMPO with a cut-off of 3% and/or expression of ≥2 pan-myeloid markers as positivity and fMPO with a cut-off of 3% as positivity, as given in Table 2. Seven ALL cases would be diagnosed as MPAL, which would be unacceptable from the clinical perspective, two cases (AML-M1 and AML-M5) had cMPO positivity and fMPO negativity, five AML-M0 cases all of which had >3% fMPO would be labelled as AML-M1, remaining 13 cases had the same diagnosis. Sensitivity and specificity for classifying AL as AML/ MPAL were 98.3 and 22.2%, respectively (Fig. 2).

and AML-M4) had 5.5 and 7.3% on fMPO becoming positive (non-discrepant), which were previously considered as negative, and both these cases also had cMPO positivity. This 5% cut-off helped us in accurately diagnosing both these cases. Nevertheless, even at this lower cut-off, one case each of AML-M1 and AML-M5 had fMPO value of only 2%, but still be diagnosed because of cMPO positivity. However, vice-versa situation occurred in AML-M1 (three cases), one case each of AML-M4 and AML-M5. None of the seven ALL cases were diagnosed as MPAL. All three MPAL cases were diagnosed as such. Sensitivity and specificity for classifying AL as AML/MPAL at 5.4% fMPO were 98.3 and 91%,

A 5% cut-off for MPO staining by FCM Totally 15 cases were discrepant when a cut-off of 5% is taken as positive for fMPO, two cases (AML-M1 Table 1 Preliminary comparison of flow cytometric and cytochemical stain for myeloperoxidase (MPO)

Different cut-offs for fMPO (number of cases)

cMPO (3%) and/ or expression of ≥2 pan-myeloid markers

0–2.9% (76) 3–4.9% (12) 5–9.9% (09) 10% or more (118) Total (215)

74 12 7 6 99

2 0 2 112 116

cMPO – cytochemical myeloperoxidase, fMPO – flow cytometry MPO. Total discrepant cases at 3, 5, and 10% (EGIL criteria) cut-off for fMPO are 27, 15, and 11 cases, respectively.

Figure 2 ROC curve: the best cut-off with sensitivity of 98.4% and specificity of 91% was achieved when comparing fMPO of 5.4% against the gold standard.

Hematology

2015

VOL.

20

NO.

8

457

Manivannan et al.

Table 2 Sl no.

Comparative study of flow cytometric and cytochemical myeloperoxidase

Detailed analyses of 27 discrepant cases

Age

Gender

fMPO (%)

cMPO

Other cytochemical stains

1

66

F

12.5

Can threshold for MPO by flow cytometry be reduced in classifying acute leukaemia? A comparison of flow cytometric and cytochemical myeloperoxidase using different flow cytometric cut-offs.

Objectives Myeloperoxidase (MPO) detection either by enzyme cytochemistry (cMPO) or flow cytometry (fMPO) plays a major role in acute leukaemia (AL) d...
196KB Sizes 0 Downloads 7 Views