PDF
Citation
Print
INTRODUCTION
Flow cytometric immunophenotyping (FCI) plays a key role in the diagnosis, classification, and monitoring of hematolymphoid neoplasms (HLNs) [1-4]. According to the 2022 European Leukemia Net (ELN) guidelines, immunophenotyping using multiparameter flow cytometry (MFC) is required for identifying cell surface and intracellular markers during work-up for AML [5]. In addition, minimal residual disease (MRD) monitoring using MFC is recommended in AML, ALL, multiple myeloma (MM), and CLL [5-8]. According to the National Comprehensive Cancer Network guidelines for MM, next-generation flow cytometry (NGF) with a minimum sensitivity of 1 in 105 nucleated cells, corresponding to the sensitivity of molecular genetic tests such as PCR or next-generation sequencing (NGS) in MRD monitoring, is essential for assessing the therapeutic response in MM [9]. Therefore, accurate and precise flow cytometric measurements are required.
Although multicolor antibody (Ab) panels with new Abs and fluorophores have been introduced in clinical practice, numerous clinical laboratories still use laboratory-developed tests comprising in-house Ab panels and methodologies [1, 2]. FCI of HLNs (FCI-HLN) can be affected by various factors, including equipment, sample handling, reagents, instrument setup, procedures, and data analysis [1, 10, 11-13]. Staining patterns can markedly differ according to the types of Ab clones and conjugated fluorophores used, causing discrepancies between positive and negative cell populations [11]. As each laboratory has a different environment, standardization is strictly necessary to accurately analyze FCI-HLN results.
To the best of our knowledge, there have been no studies on the standardization of FCI-HLN in Korea. We aimed to gather basic data by surveying the current situation in Korea to establish a standardization plan for FCI-HLN, improve the quality of FCI-HLN, and provide highly objective and reliable results.
MATERIALS AND METHODS
Study population
The study protocol was approved by the Institutional Review Board (IRB) of Eunpyeong St. Mary’s Hospital, The Catholic University of Korea, Seoul, Korea (IRB No. PC22QASI0213). The questionnaire survey was designed and critically reviewed by laboratory medicine specialists from the Korean Society of Laboratory Hematology and Korean Society of Diagnostic Immunology. We conducted the survey by e-mailing the specialists at eight university hospitals in Korea actively performing FCI-HLN procedures.
Questionnaire survey
The questionnaire (Supplemental Data Table S1) consisted of 12 questions in three main categories to assess the current laboratory practice status of FCI-HLN: (1) basic characteristics of each hospital: hospital grade, number of licensed beds, annual flow cytometric test numbers, and type of flow cytometer used (manufacturer); (2) analytic phase (test phase) variables for each test panel: reagent Abs (RAs), fluorophores, amount of sample per tube, amount of RA per tube, and acquisition cell count; and (3) post-analytic phase (report phase) variables of each test panel: positive criterion (positive cut-off), negative criterion (negative control), and report contents. The test panels included acute HLN, chronic HLN, plasma cell neoplasm (PCN), and MRD panels. All eight laboratories completed the survey. We analyzed the survey results in detail.
RESULTS
The basic characteristics of the clinical laboratories in the eight university hospitals performing FCI-HLN are presented in Table 1. The hospitals used flow cytometers from Beckman Coulter (Navios series) or Becton Dickinson (FACS series).
RAs and fluorophores used
Acute HLN panels
The eight hospitals used acute HLN panels comprising 16–32 RAs (Table 2). In seven hospitals, the acute HLN panel comprised 18–32 RAs, regardless of the acute HLN cell type. In one hospital, the number of RAs in the acute HLN panel varied according to the cell type: 18 Abs for AML, 16 Abs for ALL, 22 Abs for mixed-phenotype acute leukemia (MPAL), and 18 Abs for T-lymphoblastic leukemia (T-ALL). Four hospitals added optional 2–8 RAs for cases requiring further confirmation after the primary test. The common RAs in the acute HLN panels used in the eight hospitals were cluster of differentiation (CD)2, CD3, CD5, CD7, CD10, CD13, CD14, CD19, CD20, CD33, CD34, CD41, CD45, CD56, CD64, CD117, cytoplasmic CD (cCD)3, cCD22, human leucocyte antigen DR (HLA-DR), myeloperoxidase (MPO), and terminal deoxynucleotidyl transferase (TdT).
Chronic HLN panels
The eight hospitals used chronic HLN panels comprising 15–26 RAs (Table 2). Seven hospitals used common panels comprising 16–23 RAs, regardless of the chronic HLN cell type. In one hospital, the number of RAs in the chronic HLN panel varied: 15 Abs for B cells and 26 for T/natural killer cells. The common RAs in the chronic HLN panels used in the eight hospitals were CD2, CD3, CD4, CD5, CD7, CD8, CD10, CD19, CD20, CD23, CD45, CD56, Flinders Medical Centre (FMC)7, surface immunoglobulin kappa (sIgƙ), and sIg lambda (sIgƛ).
PCN panels
Only seven of the eight hospitals used PCN panels comprising 5–12 RAs (Table 2). The common RAs in the PCN panels used in the eight hospitals were CD19, CD38, CD45, CD56, CD138, cytoplasmic (c)Igƙ, and cIgƛ.
MRD panels
Seven hospitals used various MRD panels (Table 2). ALL-MRD and MM-MRD panels were used in six and four hospitals, respectively, and three hospitals used both ALL-MRD and MM-MRD panels. The ALL-MRD panels comprised 6–14 RAs, and the MM-MRD panels comprised 7–10 RAs. One hospital used CLL-MRD (eight RAs), mantle cell lymphoma (MCL)-MRD (10 RAs), and mature B cell neoplasm-MRD (five RAs) panels. In the B-lymphoblastic leukemia (B-ALL)-MRD panels, the commonly used RAs included CD10, CD19, CD20, CD34, CD38, CD45, and CD58. In the MM-MRD panels, the commonly used RAs included CD19, CD27, CD38, CD45, CD56, CD81, CD117, CD138, cIgƙ, and cIgƛ.
Fluorophores
The number of fluorophores in each test panel ranged from 4–10. The fluorophores used in each test panel are described in detail in Table 2.
Sample amount, antibody amount, and number of analyzed cells
Sample amount
For acute HLN, chronic HLN, or PCN panels, five to six hospitals used 50 µL or 100 µL of sample per tube, and two hospitals adjusted the sample amount to the desired cell count (1×106 cells). For the MRD panel, five hospitals adjusted the sample amount to a desired cell count (5×105–1×107 cells), and two hospitals used 50 µL or 100 µL of the sample (Table 2).
Antibody amount
The amount of RA per tube in all panels ranged from 2–20 µL. Four hospitals used 5 µL or 10 µL of RA, and four hospitals used 2 µL to 20 µL, depending on the fluorophore, antigen, and type of Ab used (Table 2).
Acquisition cell count
The acquisition cell count varied among the hospitals according to the test panel: 5×103–2×105 cells for acute and chronic HLN panels, 5×103–1×106 cells for the PCN panels, and 5×103–1×107 cells for the MRD panels. In one hospital (E), the acquisition cell count varied according to the Ab type used (1×104 cells for surface Abs [sAbs] and 5×103 cells for cytoplasmic Abs [cAbs]) in acute and chronic HLN panels (Table 2).
Post-analytic phase variables
Positive criterion (positive cut-off)
All hospitals except hospital D used ≥20% of target cells as a positive criterion for the acute and chronic HLN panels. Notably, for cytoplasmic or nuclear markers, including MPO and TdT, 10% was used as a positive criterion in three hospitals (E, F, and G) and 5% in one hospital (H). For the PCN panel, all hospitals used an abnormal phenotype as a positive criterion. For the MRD panel, four hospitals used the limit of detection (LoD) and lower limit of quantification (LLoQ), and the other three hospitals used an abnormal phenotype as a positive criterion (Table 3).
Negative criterion (negative control)
Five hospitals used an isotype control, and three used an internal control as a negative control (Table 3).
Report contents
Most hospitals reported positive/negative assignments, percentages, or general opinions for all panels. For the acute HLN, chronic HLN, and PCN panels, four hospitals reported the fluorescence intensity. One hospital (H) reported the acquisition cell count. For the MRD panel, four hospitals reported LoD, LLoQ, debris percentage, or previous results (Table 3).
DISCUSSION
This study demonstrated the variable pre-analytical, analytical, and post-analytical status of FCI-HLN in Korea. Several studies on the standardization of FCI-HLN have been conducted [1, 2, 4, 11, 14-25]. The EuroFlow Consortium developed a fully standardized approach for FCI, encompassing instrument setup, software tools, and data analysis [1, 16].
The WHO classification defines lineage assignment criteria for MPAL, and all acute leukemia panels include MPO for the myeloid lineage, CD10 and CD19 for the B lineage, and CD3 (cytoplasmic or surface) for the T lineage. Other Abs for monocytic differentiation (CD11c, CD14, and CD64) and the B lineage (CD22 and CD79a) are selectively included [26]. The EuroFlow Consortium suggests the use of a screening panel prior to comprehensive FCI using extended Ab panels for the diagnosis and classification of hematological malignancies, and two types of markers are combined in each 8-color tube: (i) backbone markers to identify distinct cell populations in a sample, and (ii) markers for the characterization of specific cell populations [22]. The CLSI H43-A2 guidelines recommend that sufficient Abs should be used to distinguish between abnormal and normal cell populations in FCI-HLN [3]. Extensive Ab panels with 20–24 Abs are required for both acute and chronic HLN [3]. Our data showed that each hospital used 16–32 Abs for acute HLN and 15–26 Abs for chronic HLN. For panel design for FCI-HLN, the following approaches are generally used [3]. The first approach is to use various Abs in a single step for the complete characterization of both abnormal and normal cell populations, regardless of the HLN subtype [3]. The second approach is to use a small number of Abs for an initial screening, followed by additional Abs selected based on the initial results [3]. Our data showed that most hospitals used the first approach for both acute and chronic HLN panels. Only a few hospitals used the second approach; two hospitals used this approach for acute HLN and one hospital for chronic HLN.
Typically, the larger the number of Abs used, the higher the sensitivity and specificity for the detection and characterization of abnormal cell populations in FCI-HLN [3]. Acute and chronic HLN panels proposed by the French Groupe d’Étude Immunologique des Leucémies (GEIL) include approximately 30 and 20–30 Abs, respectively [2]. However, most hospitals in Korea use panels with approximately 18 Abs for acute and chronic HLN, in compliance with the number of Abs permitted by the national health insurance standards. The current standards for cell marker analysis for national health insurance coverage in Korea are <18 Abs at initial diagnosis and <5 Abs (eight Abs for ALL since 2020) at MRD monitoring. At the time of initial HLN diagnosis, accurate and precise analysis of target cells is essential, and >18 Abs are required for FCI-HLN. Therefore, hospitals cannot claim health insurance fees from the Health Insurance Review and Assessment Service for Abs beyond the national health insurance coverage. As suggested by the EuroFlow Consortium, backbone markers must be used for accurate and objective FCI-HLN analysis [22]. However, national health insurance in Korea does not cover overlapping use of the same Ab. Therefore, multiple Abs (≥18) and overlapping use of Abs required in clinical practice must be considered when determining the national health insurance fee for FCI-HLN in Korea. In addition, diagnostic test kits and software for standardized FCI-HLN analysis have been introduced, and precise analytic methods are used in clinical practice [27-28]. Therefore, the national health insurance fee schedule in Korea must be urgently improved based on the number of Abs used in FCI-HLN.
In chronic HLN panels, 15 Abs are commonly used in Korea, including CD2, CD3, CD4, CD5, CD7, CD8, CD10, CD19, CD20, CD23, CD45, CD56, FMC7, sIgƙ, and sIgƛ. CLL has a characteristic immunophenotype, expressing CD5, CD19, dim CD20, dim CD22, CD23, bright CD43, dim CD45, dim to negative CD79b, dim CD81, CD200, and dim monoclonal sIg [8, 29, 30]. The prevalence of CLL has recently been increasing in Korea [31]. As FCI is beneficial for diagnosing CLL, the number of Abs must be expanded to facilitate additional confirmatory tests following screening [8, 22, 29]. The GEIL proposed 8-color and 10-color acute and chronic HLN panels [2]. In our study, 8-color Ab panels were the most common, followed by 4-color panels. Only one clinical laboratory (H) used 10-color Ab panels for B-ALL-MRD and MCL-MRD.
The International Clinical Cytometry Society and European Society for Clinical Cell Analysis suggested that the use of eight Abs (CD38, CD138, CD45, CD19, CD56, CD27, CD81, and CD117) is the best practice for MM-MRD monitoring, and routine evaluation of intracellular light chains is not recommended [23]. In our study, three hospitals used nine or 10 Abs for MM-MRD divided into two tubes following EuroFlow standard operating procedures (SOPs), including eight recommended Abs and two intracellular Ig light chains (cIgƙ and cIgƛ) [23]. In addition, CD19, CD27, CD38, CD45, CD56, and CD138 were used as backbone markers. In MFC-MM-MRD monitoring, the assessment of 5×106 cells and a minimum sensitivity of 10–5 are required [7]. NGF with improved sensitivity, applicability, and reproducibility than that of MFC has been developed and is used in clinical practice [7, 9, 10, 31]. Over 10×106 cells can be evaluated using NGF, with a sensitivity of 10–6, which is comparable to the performance of molecular methods (e.g., NGS) [10, 31]. In our study, two hospitals (B and H) achieved this sensitivity in MM-MRD.
In FCI-HLN, the minimum acquisition cell count is 1×104–5×104 cells per tube, and at least 200–1,000 target cells should be assessed [3]. Our data showed that 5×103–2×105 cells were used for FCI of acute and chronic HLN. In MFC-MRD monitoring, obtaining a high number of events is recommended to detect rare cell populations [10, 12, 17, 23, 25, 32]. The ELN recommends >5×105 CD45+ cells and at least 100 viable cells in the blast region for MFC-AML-MRD monitoring [33]. Studies have indicated that millions of cells (for instance, 3×106 cells for AML and CLL, 4×106 cells for B- and T-ALL, and 5–20×106 cells for MM) are required for MRD panels [10, 17]. A white blood cell count of 5×109/L suggests that 10 mL of whole blood likely contains approximately 5×107 cells [32]. To obtain a desired CV of ≤5%, a minimum of 4×103 cells at a frequency of 10% and a maximum of 1011 cells at a frequency of 0.00001% are required [32].
According to the EuroFlow SOPs for 8-color MFC, 300 µL of sample and an appropriate volume of Abs are needed for using the EuroFlow Ab panel with sAbs [16]. In our study, most hospitals used 50–100 µL of sample and <20 µL of Abs, which is lower than the quantities recommended by EuroFlow. This variance is attributed to the common practice of using laboratory-developed tests for FCL-HLN in Korea. Each clinical laboratory should establish the optimal sample amount for FCI-HLN to ensure the minimum number of target cells [34]. In addition, the minimal Ab amount should be determined for optimal resolution and positive-staining intensity [3]. Using Ab titration, the Ab amount and concentration can be determined to achieve the highest signal and lowest noise levels [34]. Two hospitals adjusted their sample amounts to the cell count per tube (1×106 cells). This is in line with the CLSI H43-A2 guidelines. As most manufacturers recommend Ab amounts to be set in the normal range of target cells (0.2–2×106 cells), each clinical laboratory should adjust the target cell range either to 0.2–2×106 cells or according to the manufacturer’s recommendations [3].
Clinical laboratories traditionally use 20% as the positive cut-off value for sAbs and 10% for cAbs (MPO, cCD3, cCD79a, and TdT) [35]. Internal (negative) or isotype controls, which are used to measure background fluorescence, can be used as negative controls to distinguish between positive and negative populations [3, 18, 34, 36]. For sample QC, both negative and positive controls are required [3]. Several materials can be used as positive controls to confirm the methods for target cells, RAs, and the optimal staining procedures, including human white blood cells, normal bone marrow cells, or cryopreserved human HLN cells (or cell lines) [3]. Our data showed that a conventional cut-off value (20%, except for hospital D, which used 10%) for internal or isotype controls was used to determine positive results in all panels. Previous studies have suggested cut-off values ranging from 5.4%–28% for MPO positivity in FCI-HLN [37-39]. As a negative criterion, isotype controls were used more than internal controls in this study. As the positive cut-off value is crucial for determining the lineage in HLN diagnosis, the positive cut-off value for FCI-HLN requires standardization. It can be defined by using an internal control or by comparing the degree of antigen expression in the patient sample with that in a normal cell sample under identical instrument settings [3]. For example, the normal T cell population can be defined based on the expression of CD3, CD4, and CD8 [3]. The cut-off values for MFC-MRD positivity were 0.01% (10–4) for ALL, 0.01% (10–4) for CLL, and 0.001% (10–5) for MM [6-10]. When 5×106 cells are analyzed for MFC-MM-MRD, the LoD and LLoQ are 0.0006% and 0.001%, respectively [25].
In reporting quantitative or qualitative data, the following variables should be included: numeric results (such as percentage and cell counts), reference range (if applicable), abnormal phenotype, percentage of the population of interest and gated populations (such as gated CD45+ cells), summarized findings, the level of antigen expression (negative, positive, partially expressed, normal-positive, dim-positive, bright-positive, and heterogeneous or variably intense staining), list of Abs, and interpretational comments or recommendations [34]. When reporting MM-MRD panel results, including the LoD and LLoQ is recommended [25]. However, determining the LoD and LLoQ is challenging in clinical practice. As MRD panel design and evaluation are complicated, standardized guidelines reflecting the real hospital setting are required. In our study, most clinical laboratories reported positive/negative assignments, percentages, or general opinions. Some reported fluorescence intensity and acquisition cell count, and a few reported the LoD and LLoQ when reporting MRD panel results.
Regarding HLN diagnosis and MRD monitoring, variability in FCI may directly affect medical decisions on treatment and prognosis. Therefore, standardization of FCI is critical to ensure reliable results [14]. A standardized protocol including pre-analytical, analytical, and post-analytical steps would lead to harmonized and reproducible FCI results in clinical practice [15]. Multicolor flow cytometry has become a routine method of FCI-HLN, and clinical cytometers can measure up to 10 fluorescent parameters. To leverage a large number of channels in multicolor experiments and well-designed Ab panels, proper compensation for spectral spillovers of the dyes is essential [40].
We observed remarkable differences among the hospitals in terms of the RA combinations used and the acquisition cell count. This variance can be attributed to the lack of guidelines on standardized Ab use and the fact that the criteria are established solely for the number of Abs permitted by national health insurance in Korea. Hospitals should consider these differences as they may affect FCI-HLN results. Concordantly, these variations are reflected in the analysis of external quality assessment results in Korea.
We surveyed eight clinical laboratories actively conducting FCI-HLN, focusing on the most common challenges in panel design and data analysis and interpretation when implementing FCI-HLN panels in clinical practice. Detailed steps for QC were not within the study scope. There are various control materials or reagents for QC: setup controls, unstained controls, compensation controls, viability dyes for excluding dead cells, isotype controls, Fc receptor blockers (including commercial blocking solutions, unlabeled Ig, and animal serum), and fluorescence-minus-one controls [3, 16, 18, 21, 34, 36]. We surveyed only the use of isotype controls commonly used as negative controls in clinical practice. Further investigation is required to obtain practically useful information on QC of flow cytometry.
In conclusion, this was the first extensive evaluation of the current status of FCI-HLN in Korea. Our findings revealed a high heterogeneity and complexity of FCI-HLN practice. Standardization of FCI-HLN is urgently required, and the study data provide a reference for establishing standard FCI-HLN guidelines that are clinically useful. In addition, the national health insurance fee for FCI-HLN should be increased to facilitate the use of an appropriate number of Abs in accordance with international diagnostic standards.
SUPPLEMENTARY MATERIALS
Supplementary materials can be found via https://doi.org/10.3343/alm.2023.0298
XML Download