Distribution of subtypes and immunophenotypic characterization of 1379 cases of paediatric acute leukaemia

Objectives: Acute leukaemia is the most common and highly curable childhood malignancy; subtyping and identification of antigens via immunophenotyping helps in treatment plan as well as minimal residual disease monitoring. Methods: This retrospective study was conducted at the Haematology section of the clinical laboratories of Ziauddin University Hospital and The Indus Hospital, Karachi conducted at January 1st, 2012 to December 31st, 2017. The study included 1379 cases of de novo acute leukemia from 2012 to 2017. Among these, 80% were diagnosed by using four color flowcytometry (FACS Calibur), 9% and 11% via immunohistochemistry on bone marrow trephine biopsy samples and morphological examination respectively. Results: The mean age of patients was 7.4 ± 4.3 years while male to female ratio was 1.75:1. Lymphoblastic leukaemia accounted for 77.2% and myeloid leukaemia 21.2%. Amongst lymphoblastic lineage, B-ALL was 80.4% while T-ALL was 19.6%. Among the phenotypic expression of B-ALL, CD79a (99.8%) had the highest positivity. In B-ALL, CD13 (29.8%) was the most common aberrant myeloid marker. Aberrant expression of CD79a observed in 11.1% of T-ALL cases. In non APL AML, aberrant expression of CD79a and CD19 was observed in 6.6% and 5.5% of cases respectively. Conclusion: Overall immunophenotypic profile, expression of aberrant phenotypes and subtype distribution in our patients was similar to international literature except for a relatively high frequency of T-ALL which was discordant from the western data.


INTRODUCTION
delineation. 3,4 Detailed understanding of these phenotypic patterns of differentiation allows for more precise classification of acute leukemia than does morphology alone. 5 Moreover, identification of aberrant antigen expression is important in characterizing neoplastic population among non-neoplastic counterparts. Aberrant phenotypes aid in subsequent testing of minimal residual disease during the course of treatment.
A recently published study demonstrates incidence, prognostic factors and outcomes of childhood acute myeloid leukemia (AML). 6 Unfortunately, no large study is published from Pakistan which demonstrates the distribution of childhood acute leukaemia. It is important to know the trends in our population and identify differences from other regional and western populations with respect to acute leukaemia subtypes and immunophenotypic profile. This study, the largest in the country, is aimed to report the subtypes of childhood acute leukaemia, antigen profile and aberrant antigen expression according to our demographics. Sequential data analysis was carried out on all consecutive samples received from paediatric hematology oncology unit of The Indus Hospital, Karachi Pakistan, in order to identify immunophenotypic patterns in our paediatric population. No patient was excluded from the study resulting in a very large sample size, which enhanced the accuracy and precision of the results. Cytochemical myeloperoxidase (cytoMPO) staining was also performed on all peripheral blood samples and/or bone marrow aspirates; samples with ≥3% positive blasts were considered positive for cytoMPO and classified based on morphological classification of AML as per WHO classification. Immunophenotyping by 4-colour flowcytometry using FACS Calibur flowcytometer (Becton Dickinson, Biosciences) and/or immunohistochemistry on trephine biopsies was performed on all cytoMPO negative samples as well as on samples with less than 3% cyto MPO positivity (Fig.1). The antibodies used for flowcytometry (BD Biosciences) were as follows:
Flowcytometric analysis was carried out using Paint-A-Gate software. Samples with low WBC, hemodiluted/dry tap bone marrow aspirates were immunophenotyped by immunohistochemistry (IHC) performed on bone marrow trephine biopsies. The panel of IHC included TdT, CD34, CD117, CD79a, CD20, CD3, cytoMPO, CD64, CD163, Glycophrin A, CD61, CD4, CD8, CD5, CD7 and CD10. Statistical analysis was performed using SPSS version 24.0. Numbers with percentages were used for discrete variables and mean ± standard deviation was used for normally distributed continuous variables. Chi-square test was used to assess an association between disease entity and pre-defined age groups. Significance was assessed at P ≤ 0.05.

Patient's characteristics:
A total of 1379 paediatric acute leukaemia cases were included in the study. Distribution of modalities used for diagnosis is shown in Fig.1. Male to female ratio was 1.75:1 with male preponderance in all acute leukaemia cases. In B-ALL, T-ALL and AML male to female ratio was 1.5:1, 3.8:1 and 1.5:1 respectively. Age distribution of B-ALL and T-ALL showed that 75.7% and 62.7% were between 1 and 10 years respectively, while AML occurred in 63.4% in this age group (Table-I). Statistically significant difference was observed between ALL and AML (p=0.003) as well as in B-ALL and T-ALL (p=0.00) in the 3 age groups. Lymphoblastic leukemia: Among the B-ALL, CD79a and CD19 were expressed in 99.8% and 99% cases, respectively whereas expression of CD20 was observed in 59.4% cases. There were 5.3% of cases negative for CD10 (Table-II) and 2.8% were negative for both TdT and CD34 (Table-III).

DISCUSSION
Despite the fact that childhood acute leukaemia is one of the most common 7,8 and curable 9 malignancies, data published from Pakistan is limited describing subtype distribution and detailed immunophenotypic characterization.
To the best of our knowledge, the present data from Pakistan is one of the largest studies (1379 cases in six years) in which all consecutive cases of acute leukaemia presenting to the paediatric haematology-oncology department were categorized by extensive immunophenotyping. Despite high incidence of leukemia, currently no registry exist in our country to demonstrate actual data. This study represents the true demographics and distribution of the subtypes of acute leukaemia in the paediatric population of Pakistan. Most of the cases were diagnosed by advanced flowcytometry technique utilizing extensive panel of markers. These panels not only comprised of lineage specific markers as well as other large number of markers that can be utilized in subsequent minimal residual disease (MRD) testing. This immunophenotypic profile will definitely aid to design panels for MRD, which play an important role in risk stratification of these patients.
In the present study, patients were categorized into three age groups and results showed that highest number of acute leukaemia cases was found in age group 1-10 years with male gender predominance observed in all subtypes. Published literature from Pakistan as well as other regional studies stated similar findings. 1,10 A study conducted by Amna et al from Pakistan reported 83.3% ALL, 11 two studies from India reported 76.9 and 77.84% 12,13 while the present study has reported 77.2% cases of ALL. Similar Pakistani study reported 14.7% AML cases 11 however the incidence of AML was slightly higher in Indian studies 12,13 and was similar to present study. In terms of MPAL Indian studies reported 1.1-2.3% while our study had slightly low incidence i.e. 0.9%. However, reports from the western countries showed contradictory results from as low as 0.3% to as high as 4.2%. 14 In our study, low incidence of MPAL may be due to institutional policy i.e. all cytoMPO negative acute leukaemia cases were analyzed by flowcytometry, while only selective cytoMPO positive cases were tested by flowcytometry.
Our study reported incidence (80.4%) of B-ALL which was similar to other Pakistani studies (78.5-87%) 10,15 as well as an Indian study by Madhumathi et al. 16 However studies from West 10,14,17 reported between 72.9 and 91% frequency. Among lymphoblastic leukaemia, 13-23% of T-ALL cases were reported in other studies from Pakistan in last 10 years 10,11,15,18,19 while similar distribution was reported from India 12,13 with concordant distribution reported in the present study. Studies from West reported a relatively lower frequency. Lack of expression of immature markers in this study and comparison to other studies is shown in Table-III. B-ALL cases showed expression of CD79a (99.8%) and CD19 (99%), which is consistent with published literature 17 and WHO also assigned CD19 with co-expression of CD79a, CD22 and CD10 for B-lineage. 2 Studies reported CD19 as the most sensitive marker in diagnosing B-ALL, 20,21 whereas Tong et al. reported CD79a as the most often expressed antigen 17 in concordance to our study. This however lacks specificity as aberrant expression of B-cell antigen CD79a was observed in 11.1% cases of T-ALL in this study, which is consistent to the reported literature. 22 Negativity for CD20, a B-cell specific marker was observed in 61.8% of B-ALL cases reported by Tong et al. 17 which is slightly lower in the current study (297/731, 40.6%) while another study by Seegmiller et al. reported 33.8% CD20 negative cases. 23 Negativity for CD45 also called leukocyte common antigen was observed in 13.2% B-ALL cohort and was comparable to Seegmiller et al. 23 Absence of CD10 expression was observed in 5-18.8% reported in eastern and western studies 17,23 and similar frequency (5.3%) was reported in current study.
In our study all T-ALL cases were cyCD3 positive. It is well known fact that cyCD3 is the best marker for T-ALL cases. In our study, CD4/ CD8 co-expression was seen in 50.2% cases, CD4/ CD8 double negativity in 17.2%, CD4 pos /CD8 neg in 34.3% and CD8 pos /CD4 neg in 3.4% T-ALL cases. Gupta at al. reported CD4/CD8 co-expression in 39.3% cases, CD4/CD8 double negativity in 32.8%, CD4 pos /CD8 neg in 21.3% and CD8 pos /CD4 neg in 6.6% cases. 13 A wide range of aberrant expression of myeloid antigens including CD13 and CD33 in B-ALL and T-ALL cases reported in both eastern and western literature. 13,17,22,23 While in present study CD13 (29.8%) was the most common aberrant myeloid marker followed by CD15 (20.3%), and CD 33 (17.2%) in B-ALL. T-cell antigen (CD4) aberrant expression was reported by Seegmiller et al. in 2.2% B-ALL children, 23 which was 1.8% in the present study.
Among the subtypes of AML, the most frequent was AML with maturation similar to the reported literature. 19 Most published data indicated predominance of acute myeloid leukaemia with and without maturation similar to our study. TdT is a precursor lymphocyte marker, aberrantly expressed in AML. In this study, TdT was expressed in 24.5% of AML cases. B-cell antigens CD79a and CD19 were also expressed in 6.6% and 5.5% cases of AML. This was less frequent than published literature. 24 Limitations of the study: Cytogenetics is the major limitation of this study, as we do not have established karyotyping and mutation analysis which is an important variable according to current WHO Classification. Moreover, due to constraint resources, we could not perform flowcytometric testing on some MPO positive cases that is another deficiency in this study.

CONCLUSION
On the basis of this large cohort, we can conclude that distribution of acute leukaemia subtypes in our population is close to published literature from Pakistan and India. Aberrant expression of antigens is a well-documented phenomenon and have no therapeutic implications however utilized in minimal residual disease analysis. Prognostic significance of these aberrancies should be evaluated through further studies by correlating with cytogenetics and clinical outcome.