Quantitation of Plasma Cells in Bone Marrow Aspirates
by Flow Cytometric Analysis Compared With
Morphologic Assessment
Kristi J. Smock, MD; Sherrie L. Perkins, MD, PhD; David W. Bahler, MD, PhD
● Context.—Accurate quantitation of bone marrow plasma
cells is an important component in the diagnosis and posttreatment assessment of plasma cell dyscrasias. Although
flow cytometry is sometimes used for this purpose and can
rapidly evaluate many cells, the accuracy of flow-based
plasma cell quantitation compared with morphologic assessment (currently the gold standard) is uncertain as direct comparison studies have not been previously reported.
Objective.—To determine how percentages of plasma
cells in diagnostic aspirate smears quantitated by morphologic assessment relate to percentages of plasma cells
quantified by flow cytometry.
Design.—Thirty bone marrow cases with 10% or more
plasma cells and leukemia/lymphoma flow cytometry immunophenotyping studies were identified from our hema-
topathology database. The Wright-stained aspirate smears,
marrow biopsy sections, and flow cytometry histograms
were reviewed.
Results.—Morphologically determined plasma cell percentages from the diagnostic aspirate smears were consistently higher than those determined by flow cytometry.
Much of this difference appeared to be related to differences in sample quality. However, the cellular processing
involved in performing flow cytometry also appeared to
reduce plasma cell percentages in many cases.
Conclusions.—This study helps define the limitations of
flow cytometry for quantitating plasma cell loads in marrow aspirate specimens that may significantly affect the
diagnosis or assessment of treatment response.
(Arch Pathol Lab Med. 2007;131:951–955)
T
and typically resemble normal plasma cells, they may also
demonstrate a range of maturation and structures, including blastic, multinucleated, pleomorphic, or anaplastic
forms.1 Both aspirate smears and trephine biopsy sections
are often evaluated because neoplastic plasma cell infiltrates may be patchy and localized in some cases.3 For this
reason, immunohistochemical staining of biopsy sections
using antibodies against ␬ and ␭ light chains or CD138,
which are specific plasma cell markers, may be necessary
for accurate plasma cell quantitation and assessment of
plasma cell distribution or growth patterns.4,5
Flow cytometric immunophenotyping is commonly
used as an adjunct to morphologic analysis for evaluation
of plasma cell dyscrasias.6–8 Advantages of this method
include the ability to identify small numbers of plasma
cells, demonstration of monoclonality, identification of aberrant antigen expression or abnormal expression levels,
and rapid turnaround time. In addition, flow cytometry
can also often distinguish normal and neoplastic plasma
cells on the basis of the differential expression of CD45,
CD56, and other markers.8,9 Despite having certain advantages over morphologic analysis, the accuracy of flow cytometric immunophenotyping for quantitating plasma cell
percentages is unknown because comparison studies for
morphologic evaluation of diagnostic aspirate smears have
not been previously reported.
In this study, we compared plasma cell percentages determined by multiparameter flow cytometry and morphologic analysis in 30 bone marrow specimens from patients
with plasma cell dyscrasias to determine the accuracy of
he diagnosis and classification of plasma cell dyscrasias typically require quantitation of plasma cell numbers, along with correlation of clinical, radiographic, laboratory, and pathologic findings.1 For example, a diagnosis
of multiple myeloma involves application of both major
and minor criteria. The major criteria include greater than
30% bone marrow plasma cells, histologically confirmed
plasmacytoma, or identification of a monoclonal immunoglobulin component (M-spike) in the serum or urine
(IgG ⬎ 3.5 g/dL or IgA ⬎ 2.0 g/dL). Minor criteria include 10% to 30% marrow plasma cells the presence of an
M-spike at levels less than those of major criteria, lytic
bone lesions, or a decrease in normal serum immunoglobulin levels (⬍50% of normal). Quantitation of plasma cells
is also important in assessing treatment responses.
Morphologic evaluation of bone marrow specimens is
the most common method for quantitating plasma cells
for workup of plasma cell dyscrasias and is considered to
be the gold standard.2 Although neoplastic plasma cells
are usually straightforward to identify in aspirate smears
Accepted for publication November 29, 2006.
From the Department of Pathology, University of Utah, Salt Lake City.
The authors have no relevant financial interest in the products or
companies described in this article.
A poster (No. 1082) was presented at the 2006 United States and
Canadian Academy of Pathology meeting, Atlanta, Ga, February 14,
2006.
Reprints: David W. Bahler, MD, PhD, Department of Pathology, University of Utah Medical Center, Salt Lake City, UT 84132 (e-mail:
[email protected]).
Arch Pathol Lab Med—Vol 131, June 2007
Flow Cytometric Plasma Cell Quantitation—Smock et al 951
flow cytometry in identifying and quantifying neoplastic
plasma cells. In most cases, the percentages of plasma
cells quantitated by flow cytometry were markedly lower
than those determined by morphologic analysis.
MATERIALS AND METHODS
Case Selection and Demographic Characteristics
The ARUP Laboratories’ flow cytometry and hematopathology
databases were searched for recent bone marrow cases submitted
for both morphologic review and leukemia/lymphoma phenotyping by flow cytometry in which plasma cells accounted for
10% or more of the leukocytes by morphologic assessment of the
diagnostic aspirate smears. The 30 cases selected for this study
were obtained from 18 men and 12 women (mean age ⫾ SD, 64
⫾ 12; range, 39–83 years). The research use of these materials
was approved by the University of Utah institutional review
board (No. 18325)
Morphologic Analysis
All the original diagnostic material, including Wright-stained
bone marrow aspirate smears and paraffin-embedded, hematoxylin-eosin–stained sections of bone marrow cores, were reviewed.
The morphologically determined plasma cell values from the diagnostic aspirate smears were based on counts of 500 nucleated
cells. In 28 of 30 cases, the percentages of plasma cells listed on
the original surgical pathology reports appeared to be accurate
on the basis of inspection of the core biopsy specimens and aspirate smears and were used for this study (after correction to
exclude red blood precursors—see below). In 2 cases, however,
the plasma cell values appeared different from those listed on
the reports and were obtained from recounting 500 cells on the
aspirate smears by all 3 authors; an average value was used. Percentages of plasma cells from smears prepared from samples
sent to the flow cytometry laboratory were determined from leukocyte counts of 100 to 300 cells by one of the authors from direct
smears made before red cell lysis. However, in 2 cases with suboptimal flow cytometry smears, plasma cell counts were performed by all authors, with the average of the 3 values being
used. Because red blood cell precursors are not usually identified
by flow cytometry, all morphologically determined plasma cell
percentages were expressed as a percentage of total leukocytes
to facilitate comparison.
Flow Cytometric Analysis
Routine flow cytometric immunophenotyping studies were
performed on all cases at the time of diagnosis, as described elsewhere.10 More specifically, erythrocytes and erythrocyte precursors in the bone marrow specimens were lysed by incubating at
37⬚C for 10 minutes in a 10-fold excess of ammonium chloride
(0.15M with 0.01M sodium bicarbonate and 0.1M disodium
EDTA). Following centrifugation (for 5 minutes at 400g), liquid
was removed by decanting and blotting (procedure following all
centrifugation steps). The leukocytes were manually resuspended
by using a plastic bulbed pipette, washed with 15 mL Roswell
Park Memorial Institute (RPMI) medium 1% bovine serum albumin (BSA), and resuspended with a plastic bulbed pipette in
2 mL RPMI medium 1% BSA. For surface staining, 10 ␮L or 20
␮L of antibodies was added to approximately 1 million leukocytes in 100 ␮L of RPMI medium 1% BSA. All antibodies were
used as recommended by the supplier and were directly conjugated with fluorescein isothiocyanate (FITC), phycoerythrin (PE),
or PE-cyanine 5 (PC5). Following incubation for 15 minutes at
room temperature in the dark, cells were washed with 3 mL
phosphate-buffered saline (PBS) and resuspended in 0.5 mL 0.5%
paraformaldehyde before analysis by gently running the tubes
over an empty tube rack (procedure used for all subsequent resuspension steps). Cytoplasmic staining was performed by using
the IntraPrep permeabilization reagents according to the manufacturer’s directions (Beckman Coulter, Miami, Fla). More specifically, approximately 1 million erythrocyte-depleted leukocytes
952 Arch Pathol Lab Med—Vol 131, June 2007
were incubated in 100 ␮L of fixative solution at room temperature
for 15 minutes in the dark, washed with 3 mL of PBS, and resuspended in 100 ␮L of permeabilization solution. After 5 minutes, 10 ␮L or 20 ␮L of antibodies was added and washed with
3 mL of PBS following 15 minutes of incubation in the dark, and
resuspended in 0.5 mL of 0.5% paraformaldehyde before analysis. Data acquisition and analysis were performed with an EPICS
XL cytometer and EXPO 32 software (Beckman Coulter). Antibodies used to identify plasma cells included anti-CD38 PE (Leu17), anti-CD56 FITC (NCAM16.1), anti-␬ FITC (TB28-2), and
anti-␭ PE (1-155-2), purchased from BD Biosciences (San Jose, Calif), along with anti-CD45 PC5 (HI30), purchased from Caltag
Laboratories (Burlingame, Calif). Plasma cell quantitation was
performed by surface staining (using the combination of CD56
FITC/CD38 PE/CD45 PC5) and by cytoplasmic staining (using
the combination of ␬ FITC/␭ PE/CD45 PC5. With use of our
procedure, plasma cells displayed cytoplasmic light-chain levels
1 to 2 log higher than mature B cells. Flow cytometric plasma
cell percentages were determined by quantitating those cells with
dim to negative CD45 that had bright (fourth decade) surface
expression of CD38. Plasma cell percentages were also determined by quantitating cells with bright cytoplasmic light-chain
expression that were usually located in the CD45 dim to negative
gate. Ungated ␬␥ histograms were reviewed in all cases to ensure
that rare neoplastic plasma cells that may express CD45 could
also be detected. An average of these 2 flow cytometry–derived
plasma cell values was used, but in all cases these methods gave
values within 10% or less of one another.
RESULTS
Percentages of plasma cells were determined morphologically and by flow cytometry for the 30 cases (Figures
1 and 2). The flow cytometry–determined plasma cell percentages were typically markedly lower that those determined morphologically from the diagnostic aspirate specimens (mean difference ⫾ SD, 78% ⫾ 20%; mean foldreduction, 10 ⫾ 10), with only 2 cases (cases 17 and 18)
showing similar, but still slightly lower, values by flow
cytometry (Table; Figure 3, A). Morphologic analysis indicated that the specimens submitted for flow cytometry
usually had markedly reduced plasma cell percentages
relative to the diagnostic aspirate specimens (mean difference ⫾ SD, 69% ⫾ 28%; mean fold-reduction, 9 ⫾ 14), with
only 3 cases having similar values (cases 17, 18, and 24).
Plasma cell percentages determined by flow cytometry
correlated better with morphologically determined plasma
cell percentages in the flow cytometry specimens (Figure
3, B), although many of the flow cytometry values were
low. Of the 11 cases with 10% or more morphologically
determined plasma cells in the flow cytometry specimens,
5 had values similar to the flow percentages, whereas 6
had markedly lower plasma cell percentages by flow cytometry (mean difference ⫾ SD, 48% ⫾ 16%; mean foldreduction, 2 ⫾ 1).
COMMENT
It is commonly assumed that the accuracy of flow cytometry in quantitating plasma cell numbers in bone marrow aspirates may be problematic because of hemodilution
or increased plasma cell fragility relative to other leukocytes that are typically present.9 However, to our knowledge, studies that specifically address the accuracy of routine flow cytometric immunophenotyping in quantitating
marrow aspirate plasma cell values have not yet been published. In this study, we document the frequency and
magnitude by which plasma cell percentages determined
by flow cytometry typically underestimate percentages
determined morphologically from diagnostic aspirate
Flow Cytometric Plasma Cell Quantitation—Smock et al
Figure 1. Analysis of plasma cells by 3-color flow cytometry. Most cells with dim to negative expression of CD45 (A) are plasma cells that express
bright monoclonal cytoplasmic light chains (B) and bright surface CD38 (C) (this case also expressed surface CD56).
Figure 2. Morphologic quantitation of plasma cells. A, The core biopsy specimen contained sheets of plasma cells (hematoxylin-eosin, original
magnification ⫻400). B, The diagnostic aspirate smear contained 77% plasma cells (hematoxylin-eosin, original magnification ⫻400). C, The flow
cytometry aspirate contained 11% morphologically determined plasma cells (hematoxylin-eosin, original magnification ⫻400).
smears. In addition, our data also support the hypothesis
that the process of performing flow cytometric immunophenotyping may also reduce plasma cell numbers in
some cases.
Based on morphologic analysis of the flow cytometry
specimens, most of the underestimation of plasma cell
percentages by flow cytometry appears to be due to hemodilution or poorer quality of the flow cytometry specimens compared with the aspirate smears used for diagnosis. At our institution and most others, the aspirate material used to make smears for diagnostic morphologic
evaluation is obtained before any others and is often designated as first pull. Aspirate material obtained for ancillary studies, such as flow cytometry, is typically obtained
from later aspirations or pulls and can often be of poor
quality and nonrepresentative because fewer marrow cells
were aspirated and/or substantial peripheral blood contamination was present. An interesting finding from our
study was the large magnitude of this factor (mean percentage difference, 69% ⫾ 28%), with some flow cytometry specimens having almost no morphologically detectable plasma cells and diagnostic smears having abundant
plasma cells above 30%. In only 10% of cases were the
morphologically determined plasma cell percentages in
the flow cytometry specimens similar to the diagnostic
aspirate smears.
Flow cytometrically determined plasma cell percentages
correlated better with the morphologically determined
plasma cell percentages from the specimens submitted for
Arch Pathol Lab Med—Vol 131, June 2007
flow cytometric analysis. However, in the 11 cases containing more than 10% plasma cells by morphologic analysis of the flow cytometry specimens, which represent
more reliable values, slightly more than half (6/11) had
markedly lower plasma cell values by flow cytometry
(mean difference, 48% ⫾ 16%; range, 36%–76%). These
data support the common assumption that the process of
performing routine flow cytometric immunophenotyping
also contributes to reduction of reported plasma cell values in many cases. The mechanism whereby plasma cells
are sometimes depleted by routine flow cytometric immunophenotyping is unclear but in our cases did not appear to be related to differences in specimen viability or
specimen age. Perhaps the malignant plasma cells in some
cases are more susceptible to ammonium chloride lysis
than others or are less able to survive resuspension after
centrifugation. Others have reported that ammonium chloride leaves the greatest number of plasma cells intact relative to other processing agents or methods,11 and this
method is also used for sensitive nonroutine flow cytometric tests that reportedly can detect 1 plasma cell in
10 000 leukocytes.8 However, few if any data have been
published concerning loss of plasma cells related to specimen processing, and ammonium chloride may decrease
the expression levels of some antigens, such as CD138, that
are sometimes used to identify plasma cells by flow cytometry.12 It appears, therefore, that additional study of
this issue is warranted.
Our data suggest that there are several reasons why
Flow Cytometric Plasma Cell Quantitation—Smock et al 953
Figure 3. Comparisons of morphologic and
flow cytometry–determined plasma cell percentages. A, Diagnostic aspirate smear versus
flow cytometric analysis. B, Flow cytometry
specimen versus flow cytometric analysis.
routine flow cytometric immunophenotyping studies are
unlikely to generate accurate plasma cell percentage values in most cases, including differences in plasma cell content between samples submitted for morphologic review
and samples submitted for flow cytometry, and plasma
cell loss during flow cytometry sample processing steps.
Although we used only 3-color analysis to identify plasma
cells and did not use antibodies against CD138, it is unlikely that substantial numbers of plasma cells were
missed because of this technical approach. Our method of
Comparison of Different Plasma Cell Percentages*
Morphologic Analysis, %
Case No.
Diagnostic
Aspirate
Smear
Flow
Cytometry
Specimen
1
77
11
2
23
1
3
70
1
4
39
1
5
26
3
6
45
21
7
26
19
8
16
10
9
31
6
10
29
3
11
15
2
12
25
2
13
15
3
14
48
18
15
28
2
16
15
3
17
16
15
18
25
23
19
12
5
20
13
4
21
13
5
22
45
12
23
51
6
24
74
70
25
32
4
26
45
9
27
18
1
28
29
3
29
35
13
30
75
45
Mean ⫾ SD
34 ⫾ 20
11 ⫾ 13
(range)
(12–77)
(1–70)
* Values are the percentages of leukocytes.
954 Arch Pathol Lab Med—Vol 131, June 2007
Flow
Cytometric
Analysis, %
7
2
2
1
2
5
12
5
8
2
3
1
1
20
0
2
11
21
2
4
1
11
4
33
2
8
2
4
14
29
7⫾8
(0–33)
identifying plasma cells as those cells with dim to negative expression of CD45 that also expressed bright CD38
is used by other laboratories,6 and also gave excellent
agreement with the percentages of plasma cells independently determined by quantitating cells that expressed
bright monoclonal cytoplasmic light chains. Moreover, our
analysis of cytoplasmic light-chain expression can also
easily identify plasma cells that may express CD45, although all the cases in this study were CD45 negative. In
addition, variable expression of CD138 on plasma cells has
been reported,13,14 and most participants in an international flow cytometry consensus meeting felt that the use of
CD138 was not essential in identifying plasma cells,
whereas the markers we evaluated (cytoplasmic light
chains, CD38, CD45) were the most important.15
In conclusion, this study provides considerable evidence
that routine flow cytometric immunophenotyping is not a
good method for generating accurate bone marrow plasma cell load values. Moreover, the large discrepancies between morphologically determined plasma cell values of
the diagnostic aspirates and the flow cytometry–determined values suggest that flow cytometry may also easily
miss plasma cell neoplasms in many cases unless antibodies specific for plasma cell identification are requested or
relevant history is provided. Although flow cytometry
typically underestimates the bone marrow plasma cell
load, it appears to provide a lower limit estimate and is
also useful in distinguishing between normal and malignant plasma cells, which can be very important in evaluating posttreatment follow-up specimens for residual disease.
We thank Ms Jackie McCowen-Rose for her help in producing
Figure 1.
References
1. Jaffe ES, Harris NL, Stein H, Vardiman JW. Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC Press; 2001.
World Health Organization Classification of Tumours.
2. Wei A, Juneja S. Bone marrow immunohistology of plasma cell neoplasms.
J Clin Pathol. 2003;56:406–411.
3. Sukpanichnant S, Cousar JB, Leelasiri A, Graber SE, Greer JP, Collins RD.
Diagnostic criteria and histologic grading in multiple myeloma: histologic and
immunohistologic analysis of 176 cases with clinical correlation. Hum Pathol.
1994;25:308–318.
4. Chilosi M, Adami F, Lestani M, et al. CD138/syndecan-1: a useful immunohistochemical marker of normal and neoplastic plasma cells on routine trephine bone marrow biopsies. Mod Pathol. 1999;12:1101–1106.
5. O’Connell FP, Pinkus JL, Pinkus GS. CD138 (syndecan-1), a plasma cell
marker immunohistochemical profile in hematopoietic and nonhematopoietic
neoplasms. Am J Clin Pathol. 2004;121:254–263.
Flow Cytometric Plasma Cell Quantitation—Smock et al
6. Nowakowski GS, Witzig TE, Dingli D, et al. Circulating plasma cells detected by flow cytometry as a predictor of survival in 302 patients with newly
diagnosed multiple myeloma. Blood. 2005;106:2276–2279.
7. van Zaanen HC, Vet RJ, de Jong CM, von dem Borne AE, van Oers MH. A
simple and sensitive method for determining plasma cell isotype and monoclonality in bone marrow using flowcytometry. Br J Haematol. 1995;91:55–59.
8. Rawstron AC, Davies FE, DasGupta R, et al. Flow cytometric disease monitoring in multiple myeloma: the relationship between normal and neoplastic
plasma cells predicts outcome after transplantation. Blood. 2002;100:3095–3100.
9. Braylan RC. Impact of flow cytometry on the diagnosis and characterization
of lymphomas, chronic lymphoproliferative disorders and plasma cell neoplasias.
Cytometry A. 2004;58:57–61.
10. Newell JO, Cessna MH, Greenwood J, Hartung L, Bahler DW. Importance
of CD117 in the evaluation of acute leukemias by flow cytometry. Cytometry.
2003;52B:40–43.
Arch Pathol Lab Med—Vol 131, June 2007
11. Lin P, Owens R, Tricot G, Wilson CS. Flow cytometric immunophenotypic
analysis of 306 cases of multiple myeloma. Am J Clin Pathol. 2004;121:482–488.
12. Witzig TE, Kimlinger T, Stenson M, Therneau T. Syndecan-1 expression on
malignant cells from the blood and marrow of patients with plasma cell proliferative disorders and B-cell chronic lymphocytic leukemia. Leuk Lymphoma.
1998;31:167–175.
13. Bayer-Garner IB, Sanderson RD, Dhodapkar MV, Owens RB, Wilson CS.
Syndecan-1 (CD138) immunoreactivity in bone marrow biopsies of multiple myeloma: shed syndecan-1 accumulates in fibrotic regions. Mod Pathol. 2001;14:
1052–1058.
14. Rawstron AC, Owen RG, Davies FE, et al. Circulating plasma cells in multiple myeloma: characterization and correlation with disease stage. Br J Haematol. 1997;97:46–55.
15. Braylan RC, Orfao A, Borowitz MJ, Davis BH. Optimal number of reagents
required to evaluate hematolymphoid neoplasias: results of an international consensus meeting. Cytometry. 2001;46:23–27.
Flow Cytometric Plasma Cell Quantitation—Smock et al 955