How to cite this article: Rollins-Raval MA, Fisher J, Craig FE. Monoclonal B lymphocytosis versus chronic lymphocytic leukemia: Factors affecting implementation of an absolute threshold. Cytometry Part B 2013; 84B: 149–156.
Monoclonal B lymphocytosis versus chronic lymphocytic leukemia: Factors affecting implementation of an absolute threshold†
Article first published online: 8 MAR 2013
Copyright © 2013 International Clinical Cytometry Society
Cytometry Part B: Clinical Cytometry
Volume 84B, Issue 3, pages 149–156, May 2013
How to Cite
Rollins-Raval, M. A., Fisher, J. and Craig, F. E. (2013), Monoclonal B lymphocytosis versus chronic lymphocytic leukemia: Factors affecting implementation of an absolute threshold. Cytometry, 84B: 149–156. doi: 10.1002/cyto.b.21077
- Issue published online: 19 APR 2013
- Article first published online: 8 MAR 2013
- Manuscript Accepted: 7 JAN 2013
- Manuscript Revised: 28 NOV 2012
- Manuscript Received: 24 JUL 2012
- monoclonal B lymphocytosis;
- chronic lymphocytic leukemia;
- flow cytometry;
Laboratories utilize diverse methods to determine the absolute count used to distinguish chronic lymphocytic leukemia (CLL) from monoclonal B lymphocytosis, with many using dual platform (DP) methods (flow cytometric percent of CLL-like B cells × hematology analyzer white blood cell count). However, flow cytometric tube-to-tube variation may make interpretation difficult, particularly when the CLL-like B-cell count straddles the recommended threshold. This study investigates the extent, and potential sources, of this variability.
Flow cytometric enumeration was performed on 20 samples with CLL-like B cells and 10 control specimens using single platform (SP) and two DP tubes A: kappa, lambda, CD5, CD10, CD19, CD38, CD45, CD20 and B: FMC-7, CD23, CD19, CD5, with and without racking to remove doublets. In addition, reproducibility studies were performed.
Three specimens showed discordant CLL-like B-cell counts relative to the threshold of 5 × 109/L. Bland–Altman analysis demonstrated a systematic bias with DP Tube B > A (5.9%) and SP > DP Tube A (9.5%). The doublet percent varied between specimens, but did not account for the bias between DP tubes. Variation was also seen in enumeration of other cell types, suggesting multiple potential sources of inconsistency.
Significant tube-to-tube variation may be seen in CLL-like B-cell counts. The precise cause of these differences is uncertain, but is likely multifactorial. If the clinical utility of an absolute threshold for the diagnosis of CLL can be confirmed, it will be important to establish recommendations for standardization, similar to those employed for CD4 and CD34 enumeration. © 2013 International Clinical Cytometry Society
Detection of a population of B cells with a phenotype characteristic of chronic lymphocytic leukemia (CLL) does not, by itself, predict clinical course. For some patients, this finding reflects overt CLL, whereas for others it is an incidental finding of uncertain clinical significance. However, the size of a population of CLL-like B cells does provide some additional prognostic information. Patients with overt CLL usually have marked lymphocytosis composed almost entirely of cells with a CLL-like phenotype. In contrast, those with fewer CLL-like cells have a more varied clinical course. The term monoclonal B lymphocytosis (MBL) was coined to emphasize the need to consider the significance of small populations of CLL-like B cells in the context of other clinical information. MBL was defined by the National Cancer Institute Working Group/International Workshop on CLL as monoclonal B lymphocytes fewer than 5 × 109/L, and this threshold has been adopted by the WHO classification of Tumors of the Hematopoietic and Lymphoid System (1). However, whether this is the most clinically useful threshold is controversial. CLL-like B cells identified during population screening are usually present at a very low level, often with a clonal B-cell count of <0.05 × 109/L and remain stable for many years. Patients with CLL-like MBL with a count between 0.05 × 109/L and 2 × 109/L have a more varied clinical course, but often have stable counts over time, whereas those with counts higher than 2 × 109/L have a significantly increased risk of developing overt CLL, with a rate of progression of approximately 1% per year. Others have suggested that 10 × 109/L or 11 × 109/L is a more appropriate threshold between MBL and CLL, based on the clinical course and the need to treat (2, 3). Regardless of the threshold applied, introduction of the term MBL has led to increased emphasis on enumeration of cells with a CLL-like phenotype. However, one potential shortcoming of the proposed diagnostic criteria is lack of standardization (4, 5). A recent review of flow cytometric methods used to detect CLL-like B cells at centers actively studying MBL highlighted this lack of standardization, with sites differing in specimen preparation, antibodies and flow cytometry instrument used, number of events collected, and analysis strategy (4). Indeed, even within our own laboratory we have noted tube-to-tube variation in the number of CLL-like B cells identified from a single specimen. The main goal of this study was to investigate whether routine flow cytometric testing performed for clinical diagnosis is adequate to accurately determine an absolute value for the distinction between CLL and MBL. In addition, we evaluated some of the potential sources of variation that make it difficult to implement a threshold value of CLL-like B cells for the distinction between MBL and CLL.
MATERIALS AND METHODS
This study was performed as a flow cytometry laboratory quality improvement project with approval by the University of Pittsburgh Medical Center (UPMC, Pittsburgh, PA) Quality Improvement Committee.
Twenty whole-blood samples with a discrete population of cells with a phenotype characteristic of CLL/SLL and 10 normal control specimens were identified in the clinical flow cytometry laboratory. Thorough phenotypic characterization was performed to confirm the presence of a phenotype characteristic of CLL (CD19+, CD20+ (dim), CD5+, CD10−, CD23+, FMC7−, surface immunoglobulin light chain restricted (dim), or surface immunoglobulin negative) and distinguish from normal or reactive CD5+ B cells and other CD5+ B-cell neoplasms. Further flow cytometric enumeration was compared between tubes within a 24-h period, and within 48 h of collection, using the parameters CD19 and CD5 that were common to three different tubes: one using a single platform (SP) method and two using a dual platform (DP) method, as outlined below. A CBC was performed using the AcT Diff 2 (Beckman Coulter, Brea, CA). Basic clinical information, including recorded history of CLL and/or reason for flow cytometry, was collected when available from the electronic medical record. Statistical analysis was performed with GraphPad Prism, Version 5.00 (San Diego, CA). Wilcoxon matched-pairs signed rank test was used to compare the medians of the protocol groups. Spearman correlation and Bland–Altman analyses were used to further evaluate the relationships of these groups.
SP analysis was performed with BD Trucount™ (BD Biosciences, San Jose, CA) bead containing tubes and the combination of CD19 PE, CD5 APC, and CD45 PerCP (BD Biosciences, San Jose, CA). For the SP tube, 50 μL of whole blood and 10 μL of CD19 PE, 10 μL of CD45 PerCp, and10 μL of CD5 APC were added to the BD Trucount tube. The tube was then vortexed, incubated at room temperature for 15 min in the dark, 500 μL of FACS Lyse™ (BD Biosciences, San Jose, CA) was added, the tube vortexed again, incubated for 15 min at room temperature in the dark, and a maximum of 50,000 events were acquired using BD FACS Canto II instruments (BD Biosciences, San Jose, CA). Analysis was performed using DIVA software (BD Biosciences, San Jose, CA), and included assessment for doublets, using a plot of forward scatter (FSC) area versus FSC height (6). Dot plots from each case were reviewed by two of the authors (J.F and F.E.C).
The following formula was used to calculate the absolute number of CD5+ B cells for the SP tube: Absolute CD5+ B-cell count = [(# Dual CD19+/CD5+ events)/# Total Bead events] × [(# of beads/tube)/50]/1,000
Staining for the flow cytometric portion of the DP method was performed using two different combinations of antibodies, Tubes A and B:
Tube A (8-color): anti-kappa fluorescein isothiocyanate (FITC), anti-lambda phycoerythrin (PE), CD5 peridinin chlorophyll protein–cyanin 5.5 (PerCP-Cy5.5), CD10 allophycocyanin (APC), CD19 PE-Cy7, CD38 APC-H7, CD45 V500, and CD20 V450.
Tube B (4-color): FMC-7 FITC, CD23 PE, CD19 PerCP-Cy5.5, and CD5 APC (BD Biosciences, San Jose, CA).
For the DP tubes, 100 μL of specimen (whole blood approximating a WBC of approximately 5 × 109/L with a dilution of the specimen made if necessary) was added to 100 μL of PBS/Azide/Fetal calf serum. Tube A, which included evaluation for kappa and lambda, underwent two prewashes with PBS/Azide using the Lyse Wash Assistant (BD Biosciences, San Jose, CA). Either 100μL of Tube A antibody cocktail or 20 μL of Tube B antibody cocktail was added. To further evaluate the effect of doublets, DP tubes were both racked (vigorously agitated against the test tube holders) to breakdown doublets or not racked. The racked samples were vortexed, racked, incubated on ice for 15 min in the dark, removed from the ice, racked, placed onto a BD Lyse/Wash Assistant (BD Biosciences, San Jose, CA) for lysing with ammonium chloride and fixing, racked, and then acquired. The same procedure was followed for the no-rack specimens, but without the racking step. All tubes were acquired using BD FACS Canto II instruments, with a maximum of 30,000 events, and analyzed with DIVA software (BD Biosciences, San Jose, CA). Dot plots from each case were reviewed by two of the authors (J.F and F.E.C). Further flow cytometric analysis was performed on all cases to evaluate for doublets, using a plot of FSC area versus FSC height (6).
The following formula was used to calculate the absolute number of CD5+ B cells for the DP tubes:
SP and DP results were compared and concordance was defined as results that are on the same side of the recommended threshold used to establish a diagnosis of MBL or CLL.
To determine the reproducibility of absolute CLL-like B-cell counts determined with each tube, cells from a CLL sample were spiked into aliquots of normal control peripheral blood sample with target CLL-like B-cell counts of 0.050 × 109/L, 5 × 109/L, and 10 × 109/L. These spiked samples were then assayed three times with each protocol: SP, DP Tube A, and DP Tube B.
Twenty whole-blood samples with a discrete population of cells with a phenotype characteristic of CLL were identified in the clinical flow cytometry laboratory (men, n = 13; women, n = 7; median age = 69). Nineteen were received in ethylenediaminetetraacetic acid, and one received in heparin. One ethylenediaminetetraacetic acid specimen was partially clotted, precluding an accurate WBC, and was excluded from further analysis.
Concordance of Absolute CLL-Like B-Cell Counts Relative to Threshold of 5 × 109/L
1. Fourteen samples had concordance CLL-like B-cell counts of >5 × 109/L for all SP and DP tubes evaluated (median SP count 29 × 109/L, range 7.64–0.33 × 109/L), including 11 samples with an absolute count of <10 × 109/L.
2. Two samples had concordant absolute CLL-like B-cell counts of <5 × 109/L (SP count, 4.17 × 109/L and 4.84 × 109/L).
3. Three samples showed discordant values with some tubes showing CLL-like B cells more than 5 × 109/L and other tubes showing less (Table 1). Although one of these cases (Case 1) had significant doublets that were decreased with racking, this did not appear to contribute to the discordant results. Review of the medical records for these three cases revealed the following information:
|Tube A||Tube B|
|1 (Red)||5.66||3.22 (1.6%)||3.14 (4.8%)||3.63 (2.7)||3.63 (7.0%)||MBL versus CLL|
|2 (Purple)||5.69||4.59 (0.1%)||4.74 (0.9%)||5.64 (2.9%)||5.42 (0.8%)||Prior diagnosis CLL|
|3 (Green)||4.36||5.67 (0.8%)||6.45 (1.4%)||5.71 (0.6%)||5.96 (1.7%)||Subsequent diagnosis CLL|
Case 1: The patient is an 82-year-old male with a history of lung carcinoma who developed shortness of breath and cough raising concern for recurrence of carcinoma. During this evaluation and follow–up, the patient was noted to have a persistent mild leukocytosis for 6 months with mild lymphocytosis and mild neutrophilia, as well as mild anemia. He had no obvious infections, inflammatory disease, or other obvious malignancies. There was no reported previous history of CLL. The differential diagnosis for this patient is CLL versus MBL.
Case 2: The patient is a 67-year-old female with a history of CLL/small lymphocytic lymphoma (SLL), Stage II, established 1 month earlier. At the time of the current peripheral blood flow cytometric evaluation, she presented with lymphadenopathy, and a lymph node biopsy performed was diagnostic of CLL/SLL.
Case 3: The patient is an 84-year-old male with no previous history. At the time of this peripheral blood evaluation, the differential diagnosis was CLL versus MBL. However, a subsequent bone marrow biopsy evaluation was diagnostic of CLL/SLL.
It is worth noting that if a threshold of 10 × 109/L was selected, instead of 5 × 109/L, four specimens demonstrated discordant results, that is absolute values that straddle the recommended threshold.
Tube-to-Tube Variation in Absolute CLL-Like B-Cell Counts
To further investigate tube-to-tube variation, CLL-like B-cell counts obtained from different tubes were compared.
- 1DP Tube A versus Tube B. There was excellent correlation between the absolute number of CLL-like B cells obtained using different DP tubes (Spearman r > 0.95 for all combinations) (Fig. 1A). However, a statistically significant difference was identified between Tube B (FMC-7/CD23/CD19/CD5) and Tube A (kappa/lambda/CD5/CD10/CD19/CD38/CD45/CD20) (P = 0.0021) (Table 2). In addition to random error, Bland–Altman analysis (Fig. 1B) demonstrated a constant, systematic, bias of −5.9% for racked Tube A versus Tube B, that is higher results for Tube B.
- 2SP versus DP. There was good correlation between the SP and the DP methods (SP versus DP-R Tube A, r =0.96; SP versus DP-R Tube B, r = 0.95) (Fig. 2A). However, there was a statistically significant difference between DP-racked Tube A and SP tube (P = 0.0230). Bland–Altman analysis demonstrated a constant, systematic bias of 9.5%, indicating a trend for Tube A to give lower results than the SP tube, in addition to random bias (Fig. 2B). There was no significant difference between the Tube B and the SP method (Table 2). If the SP tube is considered to be the gold standard, these results suggest that Tube A is giving falsely low values CLL-like B-cell counts. One potential source of a low CLL-like B-cell count would be an increased number of doublets and/or coincident events, which would be falsely counted as single events.
- 3Role of doublets/coincident events. To further investigate the potential role of doublets in tube-to-tube variation, the percentage of doublets and/or coincident events identified with a plot of FSC-Area versus FSC-Height was compared. The percentage of doublets and/or coincident events varied between specimens, but there was no significant difference in the number of doublet events between the DP Tubes A and B (Tube A: median 2.5% of total events; range, 0.5–7.0%; Tube B: median 2.5%; range, 0.8–7.9). Following racking, the number of doublets decreased slightly for both tubes (Tube A: median 1.35; range, 0.1–5.2; Tube B: 1.8%; range, 0.5–6.2%) and for Tube A was significantly lower than for the unracked tube (P = 0.017) (Fig. 3). However, there was no significant difference between the absolute CLL-like B-cell count for Tube A racked versus nonracked or Tube B racked versus nonracked (Table 2) (Fig. 4). Therefore, for these study patients, doublets and/or coincident events did not appear to contribute significantly to the tube-to-tube differences in absolute CLL-like B-cell counts.
|DP Tube A||Rack||9.01 (3.22–69.61)||0.023 versus SPb|
|0.0021 versus Tube Bc|
|No-rack||8.750 (3.1–68.55)||NS versus Racka|
|DP Tube B||Rack||10.00 (3.63–70.75)||NS versus SPb|
|No-rack||10.17 (3.630–66.59)||NS versus Racka|
Tube-to-Tube Variation in the Proportion of Non-CLL-Like Cells
The counts obtained for different, non-CLL-like cell populations were compared to determine if the differences observed above were restricted to the CLL-like cells.
- 1Samples with CLL-like B cells. No significant difference was identified between the percent of T cells obtained using different DP analysis tubes (Tube A racked median 8.25; range, 1.6–17.0; Tube B racked median 8.15; range, 2.2–17.7). However, the proportion of lymphoid cells identified with a plot of FSC versus SSC was significantly lower for Tube A (median, 74.80; range, 36.2–97.1) than Tube B (median, 75.9; range, 40.3–97.9) (P = 0.0019). This could either reflect a decrease in lymphoid cells in Tube A or a decrease in nonlymphoid cells, such as monocytes and granulocytes in Tube B.
- 2Normal control specimens. The relative proportion of B cells and T cells was further investigated using peripheral blood samples from normal control specimens. No significant difference was identified between the percent of B cell identified using Tube A and Tube B. The percent of T cells was significantly lower for Tube A (median, 19.25; range, 3.2–32.40) compared to Tube B (median, 22.15; range, 4.1–31.30) (P = 0.0371), and demonstrated a systematic bias (Bland–Altman % difference Tube A–Tube B versus mean Tube A and B = −9.59). There was also a statistically significant difference between the percent of lymphoid cells obtained with Tube A (median, 27.15; range, 5.3–43.20) and Tube B (median, 30.05; range, 7.7–44.70) (P = 0.0078).
These results suggest that, in addition to random bias, there is a systematic bias between the tubes, leading to differences in the proportion of lymphoid cells. This difference appears to contribute to the differences in absolute CLL-like B-cell count, but is not restricted to CLL-like B cells. It is uncertain if this difference reflects a decrease in the proportion of lymphoid cells in Tube A, or a decrease in nonlymphoid cells in Tube B. The SP results show better agreement with Tube B and would therefore favor the former. However, the possibility of additional sources of bias between the SP and the DP tubes cannot be excluded. In addition, the source of this bias, and other random error, remains uncertain.
Reproducibility of Absolute CLL-Like B-Cell Counts for Each Tube
|Absolute target value (0.050 × 109/L)||Absolute target value (5 × 109/L)||Absolute target value (10 × 109/L)|
|Protocol||CLL-like B-cell count||Difference from mean||% Difference||CLL-like B-cell count||Difference from mean||% Difference||CLL-like B-cell count||Difference from mean||% Difference|
|DP Tube A||0.043||0.002||5.38||5.19||0.054||1.03||10.15||−0.07||−0.69|
|DP Tube B||0.049||−0.004||−7.82||5.14||0.098||1.87||10.32||−0.25||−2.46|
This study investigated variation in CLL-like B-cell counts performed for the distinction between CLL and MBL. In our clinical practice, we have noticed tube-to-tube variation in the percent of CLL-like B cells identified within a single specimen by routine flow cytometric immunophenotyping. In addition to this variation in flow cytometric results, there is also the potential for error in the CBC results used to calculate an absolute count. Within our institution, the CBC studies obtained for clinical purposes are not performed within the flow cytometry laboratory and, particularly for references cases, may not be available to the interpreter of the flow cytometric data. Although CBC assays are performed in the flow cytometry laboratory, the focus is to calculate the appropriate dilution required to standardize the number of cells stained, and therefore the requirements for accuracy and precision are perhaps less stringent than those required for diagnostic testing. The presence of potential variability in both flow cytometric and hematology analyzer determinations makes it difficult to interpret the significance of identifying CLL-like B cells by flow cytometry, particularly when the estimated absolute count straddles the diagnostic threshold of 5 × 109/L. We sought to document the degree of variation in CLL-like B-cell counts identified by three different flow cytometric tubes, two routine DP (flow cytometry determined percent CLL-like B cells and hematology analyzer determined WBC count) and one SP flow cytometry method, and investigate potential sources of variability.
Although the flow cytometric tubes tested demonstrated good reproducibility and the CLL-like B-cell counts obtained from the three different flow cytometric tubes correlated well, there was some variation in the results obtained for each specimen with three straddling the absolute threshold of 5 × 109/L (Table 1). Although for two patients, clinical information helped to resolve the significance of this flow cytometric finding, it remained uncertain for one patient. Further investigation of the results obtained from the different DP and SP tubes, demonstrated both random error and systematic bias. There was a statistically significant difference between the results obtained for the two tubes used for the DP method, with a bias toward lower results for Tube A (kappa FITC/lambda-PE /CD5 Per-CP-Cy5.5/CD10 APC/CD19 PE-Cy7/CD38 APC-H7/CD45 V500/CD20 V450) than Tube B (FMC-7 FITC/CD23 PE/CD19 Per-CP-cy5,5/CD5 APC). The DP results from Tube B were in good agreement with the SP bead method (CD19 PE/CD5 APC/CD45 PerCP), but there was a bias toward lower results for Tube A.
To further investigate potential sources of variability, we assessed the impact of doublets and/or coincident events. We, and others, have noticed that CLL-like B cells seem to be more prone to the formation of doublets or coincident events, which theoretically could be a source of variation (7). Although doublets may be identified and excluded from the analysis (8), some cell types might be more prone to form doublets than others, and therefore lead to an erroneous determination of the relative proportion of cells present. We hypothesized that this source of variation could be decreased by manual disruption through racking (dragging vigorously across a test-tube rack several times), as proposed for the disruption of red blood cells aggregates in the flow cytometric evaluation for Paroxysmal Nocturnal Hemoglobinuria (9). Doublets and/or coincident events were detected in all tubes, likely contributed to the random error, and did decrease with racking. Therefore, minimizing doublets and/or coincident events is probably of value, but this variable did not appear to contribute significantly to the bias between tubes seen in this study. In addition to strategies used to reduce doublets, it is also important to dilute high cellularity samples to minimize coincident events and use an optimal rate or acquisition (10).
What can we learn about possible sources of variability from the implementation, and standardization, of other quantitative flow cytometric assays? As early as 1992, Robinson et al. (11) showed that leukocyte counts from a hematology analyzer could be a great source of bias in enumerating CD4 counts. Since that time, several studies have shown that eliminating this factor with the use of SP methods can improve accuracy and decreased interlaboratory variability in CD4 counts (12–14). However, recently Hultin et al. (15) found no decrease in interlaboratory variation between SP and DP methods for T-cell enumeration, and suggested that other factors, such as processing with lyse-and-wash versus lyse-no-wash methods and different gating strategies, might be more responsible for variability between labs. Similar sources of inconsistency have been encountered for the flow cytometric enumeration of CD34+ progenitor cells. Although SP methods have been associated with decreased interlaboratory variability in CD34 counts, optimal standardization has also required implementation of recommended procedures for specimen processing, CD34 antibody class and fluorochrome, and gating strategy (16–18).
There is limited literature comparing SP and DP methods for enumerating CLL-like B cells. The study of Mekouar et al. (19) reported a tendency toward slightly lower absolute counts using an SP bead method and in contrast Hanson et al. (5) illustrated slightly higher results for an SP method for many of the cases illustrated in their letter to the editor. However, neither of these previous reports includes detailed information about the methods employed. In this study, differences were found between the results obtained using the SP method and at least one of the DP tubes. Although it is tempting to assume that the SP method is correct because it removes any error associated with performing a hematology analyzer count, there are several other potential sources of variability. The SP method in this study used a commercial lyse reagent, whereas the DP tubes used an in-house ammonium chloride lysing procedure. However, in a recent report by Mekouar et al. (19), no significant difference was identified in absolute CD19+ lymphocyte counts performed after lysis with different solutions, including FACS lyse and ammonium chloride. Another variable in this study was use of a lyse-and-wash procedure for the DP tubes and a lyse-no-wash method for the SP method. For CD4 enumeration, several studies have documented less variation with lyse-no-wash procedures (13, 15), possibly because of the elimination of the centrifugation step that may cause cells loss and formation of cellular aggregates or doublets. Also, different antibody combinations and fluorochromes were used in the three study tubes and affected the ability of these tubes to separate CD5+ B cells from other B cells and T cells. In addition, the possibility could be considered that some of the antibodies used might be toxic to different cells in the sample, leading to either increased or decreased CLL-like B-cell counts. The difference in antibodies also resulted in different gating strategies, with initial gating on CD45 versus SSC used for the SP tube and one DP tube, whereas initial gating on FSC versus SSC was performed for the other DP tube because of the lack of a CD45 antibody. CD45 gating has been shown to lead to decreased interlaboratory variation for CD4 counting (15). However, given the decreased expression of CD45 on CLL cells (20), it is uncertain if the information can be effectively applied to the current assay. In addition to the differences in the antibody specificity and fluorochrome used, there were differences in antibody class, with inclusion of an IgM class FMC-7 antibody in DP Tube B. In theory, inclusion of an IgM antibody might interfere with the results because of the reported overexpression of the Fc receptor for IgM in CLL (21) and the crosslinking ability of IgM. However, limited studies performed with and without the FMC-7 antibody (data not shown) did not identify a difference in the number of CLL-like B cells or percent of doublets.
In conclusion, this study highlights some of the factors that can lead to tube-to-tube variation in CLL-like B-cell counts and have an impact on implementing an absolute threshold value. Although it is difficult to identify the precise cause of the differences between the tubes used in this study, and impossible to say which result is correct, it is important to acknowledge that tube-to-tube and interlaboratory variation may be seen when enumerating CLL-like B cells by flow cytometry. Given this variability, we recommend the following strategy. When a population with a CLL-like phenotype is identified, the approximate white blood cell count should be obtained and an absolute CLL-like B-cell count determined. Reporting of CLL or MBL is relatively straightforward when the count is either very high or very low, respectively. For specimens with a CLL-like B-cell count close to the threshold of 5 × 109/L, perhaps 3–6 × 109/L, the possibility of CLL and MBL should be raised and the flow cytometric findings considered in the context of clinical and other information. Moving forward, if the clinical utility of an absolute threshold for the diagnosis of CLL is confirmed, it will be important to establish recommendations for standardization of flow cytometric testing including specimen processing, antibody combinations, number of events acquired, gating strategy, and interpretation. Given the potential need for serial determinations and the possible added cost of bead based SP methods, it might be prudent to eliminate potential sources of variability in DP flow cytometric testing first, and then assess interlaboratory variability, before recommending implementation of SP methods.
- 7Practical Flow Cytometry, 4th ed. Hoboken: John Wiley & Sons, Inc; 2003.,
- 10Clinical Flow Cytometric Analysis of Neoplastic Hematolymphoid Cells; Approved Guideline—Second Edition, Vol. 27. Wayne, Pennsylvania: Clinical Laboratory Standards Institute; 2007. h43–A2., , , , , , , , , , et al.