DOI:10.1111/j.1365-2303.2009.00725.

Fine needle aspiration cytology and flow cytometry

immunophenotyping of non-Hodgkin lymphoma: can we do better?

P. Zeppa*, E. Vigliar*, I. Cozzolino*, G. Troncone*, M. Picardi, A. De Renzo,

F. Grimaldi, F. Pane, A. Vetrani* and Lucio Palombini*
*Dipartimenti di Scienze Biomorfologiche e Funzionali, Facolta di Medicina e Chirurgia, Universita di Napoli Federico II, Napoli,
Italia and e di Biochimica e Biotecnologie Mediche, Facolta di Medicina e Chirurgia, Universita di Napoli Federico II, Napoli, Italia

Accepted for publication 27 October 2009

P. Zeppa, E. Vigliar, I. Cozzolino, G. Troncone, M. Picardi, A. De Renzo, F. Grimaldi, F. Pane, A. Vetrani

and L. Palombini
Fine needle aspiration cytology and flow cytometry immunophenotyping of non-Hodgkin lymphoma: can
we do better?
Objective: To evaluate the diagnostic efficiency of fine needle aspiration cytology flow cytometry (FNAC FC)
in the diagnosis and classification of non-Hodgkin lymphoma (NHL) in a series of 446 cases and to compare the
results with those of previous experiences to evaluate whether there had been an improvement in FNAC FC
diagnostic accuracy.
Methods: FNAC FC was used to analyse 446 cases of benign reactive hyperplasia (BRH), NHL and NHL relapse
(rNHL) in 362 lymph nodes and 84 extranodal lesions. When a diagnosis of NHL was reached, a classification was
attempted combining FC data and cytological features. Sensitivity, specificity and positive and negative predictive
values (PPV and NPV) of FNAC FC in the diagnosis and classification of NHL were calculated and compared
with those available in the literature.
Results: FNAC FC provided a diagnosis of NHL and rNHL in 245 cases and of BRH in 188 cases. In nine cases, the
diagnosis was suggestive of NHL (sNHL) and in four cases was inadequate. Histology and clinical follow-up
confirmed 102 cases of NHL and detected one false positive. In 18 cases of BRH diagnosed by FNAC FC, histological
examination revealed 14 BRH and four NHL (false negatives). All nine cases diagnosed as sNHL were confirmed
by histology. Including sNHL cases as false negatives, statistical analysis showed 94.9% sensitivity, 99.4%
specificity, 99.6% PPV and 93.4% NPV in the diagnosis of NHL. A specific subtype was diagnosed in 125 cases and
confirmed in 67 of 70 cases that had histological biopsies. Statistical analysis did not demonstrate significant
improvements between the present series and previous studies either in diagnosis or in classification of NHL.
Conclusions: FNAC FC is a fundamental tool in the diagnosis and classification of NHL but the exiguity of
diagnostic material and other technical and clinical limitations will probably continue to limit further
improvement of the technique.

Keywords: fine needle aspiration cytology, flow cytometry, non-Hodgkin lymphoma, accuracy

(FNAC FC) has completely changed the approach to

Introduction
The combination of flow cytometry (FC) phenotyp-
ing and fine needle aspiration cytology (FNAC)
Correspondence:
Prof. P. Zeppa, MD, Dipartimento di Anatomia Patologica,
Facolta di Medicina e Chirurgia, Universita di Napoli
Federico II, Via Pansini n. 5, 80131 Napoli, Italia.
Tel.: +39 081 7 463 674; Fax: +39 081 7 463 679;
E-mail: zeppa@unina.it
the cytological diagnostic algorithm of lymph nodes
and lymphoid organs. In fact, following some pio-
neering studies,1,2 over recent years many laboratories
have adopted FNAC FC for the diagnosis of lymph
nodal and extranodal lymphoid disorders.321 The
assessment of clonality and the application of a
specific algorithm to classify most of the small cell
non-Hodgkin lymphomas (NHL) are the most appre-
ciated advantages generated by the combination of
the two techniques;9,11,12,18,19,21,22 moreover, these

300 Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd

 Fine needle aspiration cytology and flow cytometry in non-Hodgkin lymphoma
possibilities match well the latest World Health Orga-
nization (WHO) classification of tumours of haemat-
opoietic and lymphoid tissues.23 In fact, this
classification uses a multiparametric approach in
which the morphological, phenotypic and genetic
features concur with the identification of different
pathological entities. From this perspective, cytologi-
cal features and FC phenotyping represent funda-
mental tools in the diagnosis and classification of the
different entities and or the selection of other possible
ancillary techniques. Another advantage that the
FNAC FC combination has produced is the reproduc-
ibility of results in the diagnosis of benign reactive
hyperplasia (BRH), NHL and NHL relapse (rNHL), in
terms of sensitivity, specificity and positive and neg-
ative predictive values (PPV and NPV), with low
variation among different laboratories.1521,24
In a previous study performed using FNAC FC,19
we showed that this combination enhanced the
precision of cytological diagnosis in lymph nodal and
extranodal lymphoproliferative disorders and allowed
a correct classification in more than half of the cases;
moreover, we obtained results that, in terms of
sensitivity, specificity, PPV and NPV, were similar to
those obtained in the same period in other institu-
tions.1521,24 However, in all of these studies there
were still a number of inconclusive and false-negative
results and a correct classification of NHL was
obtained in a variable percentage of cases.1521,24
Since these studies represent the first extensive
clinical applications of FNAC FC, improvements
may since have been made on the basis of increased
experience. Therefore, on the basis of these studies,
we have extended our experience with an additional
series of 446 FNAC FC of lymph nodes and extran-
odal sites. The aim of this study was to evaluate the
diagnostic efficiency of FNAC FC in the diagnosis and
classification of NHL and to compare the obtained
results with those of our own and others previous
experiences. We also focused on the suspect cases,
false negatives and false positives of the present series
in order to determine whether clearing up these cases
might further improve the efficiency of the method.
Materials and methods
Patients and fine needle aspiration cytology
From the files of the cytopathology service of the
Federico II University of Naples Department of Pathol-
ogy, 1824 lymph nodal FNACs, performed during
Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd
301
a 5-year period between January 2004 and December
2008, were retrieved and reviewed retrospectively.
FNAC FC was used to analyse 446 cases of BRH,
suggestive of NHL (sNHL), NHL and rNHL. FNACs
were performed on 362 lymph nodes (81%) and on 84
extranodal lesions (19%). FNACs of 240 palpable
lesions were performed in the outpatient office of the
department; FNACs of 206 impalpable and or deeply
located lymph nodes and organs were performed under
ultrasound or computed tomography (CT) control in
the haematology and radiology departments. The
extra-nodal sites were kidney (one case), parotid (five
cases), breast (seven cases), orbit (eight cases), medi-
astinum (10 cases), thyroid (20 cases) and soft tissue
(33 cases); no bone marrow samples were examined.
The series includes patients with (133 cases) and
without (313 cases) a history of NHL. At the time of
the FNAC, the diagnostic procedure and its related risks
were first discussed with the patients and informed
consent was obtained; the FNAC procedure was then
performed as previously described.19,25 In all cases the
first pass was used to prepare two conventional smears;
the first was immediately Diff-Quik stained and
microscopically examined to evaluate the adequacy
of the sample and to select cases suitable for FC.
Inadequate cases were repeated and processed for
cytological and FC evaluation when a sufficient quan-
tity of cells was yielded. Four cases (1%) remained
inadequate after repeated passes and were requested
for surgical biopsy. The remaining material left in the
hub of the needle was carefully flushed out with
phosphate-buffered saline solution (PBS) and added to
an eventual second pass in cases of scanty cellularity.
Clinical data and microscopic features were also used to
lay out the panel of fluoresceinated antibodies to apply
in each case. Cases concerning unequivocal clinical
and microscopic reactive processes, Hodgkin lym-
phoma (HL) and metastases were not processed, and
in these cases material from additional FNACs was used
to prepare cytospin and cell blocks for conventional
immunocytochemical stains. In selected cases, material
from additional passes was used for storing cells
suitable for other, ancillary techniques.
Flow cytometry
Cell suspensions were processed within two hours;
they were washed twice by centrifugation for four
minutes at 2500 revolutions per minute (rpm),
removing the supernatant and adding 400 ll of PBS.
The final suspension was divided into four or more
302 P. Zeppa et al.
tubes when sufficient cells were available. One or two
tubes of cell suspensions were stored until the end of
the procedure in order to have further material
available in case of unsatisfactory results or the need
for additional tests. Samples were then incubated for
15 minutes in the dark with 10 ll of the following
basic combinations of phycoerythrin, perdin chloro-
phyll protein and fluorescein isothiocyanate antibod-
ies: CD3, CD4 8, CD2 3 7, CD5 10 19, CD19 j k, considered specific only for marginal zone mucosa-
FMC7 CD23 CD19, CD38 56 19, bcl-2 CD10 and
CD13 HLA-DR. All antibodies were purchased from
Becton Dickinson (San Jose, CA, USA), except bcl-2,
which was purchased from Pharmingen. After incu-
bation, red blood cells were lysed with ammonium
chloride lysing solution (diluted to 10%) for five
minutes and then washed twice. When small frag-
ments were still present, the suspension was filtered
through 50-lm filters; finally, an equal part of 1%
paraformaldehyde was added to each tube for cell
fixation. In cases of intracytoplasmic antigens, such as
bcl-2 or light chains, when the routine technique
failed to detect those on the surface, cells were
suspended in permeabilizing solution and incubated
for 30 minutes in the dark. FC was then performed
using a three-colour analysis technique on a Becton
Dickinson FACS scan as previously described.19,25 As
regards data evaluation, an antibody was considered
expressed when a minimum of 10% of the gated cells
were positive; for light chain evaluation, j:k ratios
greater than 4:1 or 1:2 were considered as definite
evidence of monoclonality.12,16,26 In cases of equivocal
results or technical difficulties, residual material in the
tubes was suitable for further analysis within 24 hours
and aliquot in stored tubes was available for further
phenotyping. A diagnosis of NHL was reached on the
basis of the cytological features and light chain
restriction or specific CD2 3 7, CD4 8, CD56 antigen
expression. Classification of NHL according to the
WHO system24 was attempted, combining the cyto-
logical features with differing expression of
CD2 CD3 CD7, CD5, CD10, CD19, CD23, FMC7,
bcl-2, CD38 and CD56 in different combined pheno-
types4,8,1114,18 to classify, when possible, the specific
subtype, and both were incorporated in the final
report. The following phenotypes, coupled with appro-
priate microscopic features, were considered specific to
the corresponding reported NHL subtypes:
CD5+ CD19+, CD23+ 19+ was considered specific to
small lymphocytic lymphoma chronic lymphatic leu-
kaemia (SLL CLL); CD38+ CD19+, CD10) CD19+
for lymphoplasmacytoid lymphoma (LpcL);
CD5+ CD19+, CD23) FMC7+ CD19+ for mantle cell
lymphoma (MCL); and CD5) 19+, CD23) 19+,
CD10+ CD19+ for follicular lymphoma (FL).
CD10+ 19+, even without light chain expression, in
an appropriate clinicalcytological setting, was consid-
ered specific for Burkitt lymphoma (BL). The pheno-
type CD5) CD19+, CD10) CD19+, CD23) CD19+
CD38) CD56) was frequently observed but was
associated lymphoid tissue (MZL MALT) lymphoma
in cases with a previous diagnosis or with specific
cytological features in extranodal sites such as the
parotid; in all other cases of NHL showing this
phenotype a diagnosis of NHL not otherwise specified
(NOS) was given. Phenotypes CD19), CD4+ CD8),
CD2+ CD3+ CD7) or CD2+ CD3) CD7+, in an
appropriate cytological setting, were considered spe-
cific for peripheral T-cell lymphoma (PTCL).27 Pheno-
type CD3+, CD56+, CD19), CD10), CD23) was
considered specific for natural killer NHL (NK).28 The
phenotype CD5) CD19+, CD10) CD19+,
CD23) CD19+, CD38) CD56), with or without light
chain expression and with appropriate cytological
features, was considered indicative of diffuse large B-
cell lymphoma (DLBCL). Aberrant phenotypes such as
CD10+ CD19+ and CD5+ CD19+, with appropriate
cytological features, were also considered specific for
DLBCL. The diagnostic FC algorithm for NHL classifi-
cation is summarized in Table 1. Cytological diagnoses
were checked by histology or clinical follow-up.
Primary diagnoses of NHL were checked histologically
except in some cases of SLL CLL, in deeply located
lesions and in old or problematic patients in whom
surgical biopsies were inadvisable for clinical reasons.
BRH diagnoses were checked histologically or by
clinical follow-up in cases in which clinical presenta-
tion did not match with the cytological diagnoses. Thus
130 FNAC FC diagnoses were checked histologically,
as follows: 103 patients with NHL, nine patients with
sNHL and 18 patients with BRH; the remaining 312
cases were checked clinically by follow-up. The cyto-
logical diagnoses of the clinically checked cases were
170 BRH, 110 rNHL and 32 NHL.
Statistical analysis
Sensitivity, specificity, PPV and NPV in the FNAC FC
diagnosis of BRH, NHL and rNHL and in their possible
classification according to the WHO system23 were
calculated. For this purpose, cases in which a definite
diagnosis of BRH or NHL with or without further
Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd
 Fine needle aspiration cytology and flow cytometry in non-Hodgkin lymphoma 303

Table 1. Flow cytometry correlation between phenotypes and NHL subtypes

Phenotype LpcL SLL CLL MCL FL MZL MALT NK PTCL BL DLBCL

j:k >4:1or <1:2 + + + + + ) ) + ) + )

CD19 + + + + + ) ) + +
CD5 CD19 ) + + ) ) ) ) ) )
CD10 CD19 ) ) ) + ) ) + ) + ) + )
CD23 CD19 + ) + ) ) ) ) ) ) )
FMC7 CD19 + ) ) + + ) ) ) ) ) )
CD38 CD19 + ) ) ) ) ) ) ) )
bcl-2 + ) + ) + ) + + ) ) ) ) + )
bcl-2 CD10 ) ) ) + ) ) ) ) )
CD4+ 8) or CD4) 8+ ) ) ) ) ) ) + ) )
CD2+ 3) 7) ) ) ) ) ) + ) ) )
CD2+ 3+ 7), or CD2+ 7+ 3) ) ) ) ) ) ) + ) )
CD56 ) ) ) ) ) + ) ) )

LpcL, lymphoplasmacytoid lymphoma; SLL CLL, small lymphocytic lymphoma chronic lymphatic leukaemia; MCL, mantle
cell lymphoma; FL, follicular lymphoma; MZL MALT, marginal zone mucosa-associated lymphoid tissue B-cell lymphoma;
NK, natural killer non-Hodgkin lymphoma; PTCL, peripheral T-cell lymphoma; BL, Burkitt lymphoma; DLBCL, diffuse large B-
cell lymphoma.

classification was made were considered truly nega- Table 2. Correlation between fine needle aspiration cytol-
tive and truly positive; sNHL cases were considered ogy flow cytometry (FNAC FC) diagnoses and histol-
false negatives. The FNAC FC classification of positive ogy clinical follow-up
cases (NHL and rNHL), when carried out, was
Histology and
confirmed by previous or subsequent histology; sen- clinical follow-up
sitivity in classification was then calculated. For
statistical evaluation, cases in which a definite classi- FNAC FC No (%) NHL BRH
fication was achieved were considered truly positive NHL 135 (30) 134 1
and cases diagnosed as NHL or rNHL without further rNHL 110 (25) 110 0
classification were considered false negatives. The sNHL 9 (2) 9 0
obtained data were compared with those from our BRH 188 (42) 4 184
previous experience in the period 1999200319 and Inadequate 4 (1) 4 0
with those from other studies performed with the Total 446 (100) 261 185
same technical procedures 1521,24 to evaluate Sensitivity 94.9%, specificity 99.4%, PPV 99.6%, NPV
whether there had been further improvement in the 93.4%.
accuracy of FNAC FC diagnosis of NHL since then.
For this purpose, a 2 2 chi-square test was run NHL, non-Hodgkin lymphoma; rNHL, NHL relapse; sNHL,
between incorrect (FP + FN) and correct (TP + TN) suggestive of NHL; BRH, benign reactive hyperplasia; PPV,
positive predictive value; NPV, negative predictive value.
diagnoses of NHL and BRH from the present and the
previous series (19). For the classification of specific
subtypes, a 2 2 chi-square test was run between (Table 2). In nine cases (2%), the FNAC FC diagnosis
incorrectly (FP + FN) and correctly classified cases was suggestive of NHL: this diagnosis was given in
(TP + TN) in the present and previous series.19 eight cases in which FNAC was indicative of NHL but
Finally, all data obtained in these studies were corresponding FC did not confirm the cytological
compared with those from previous experiences diagnosis and in one case in which FNAC was not
reported in the literature (Table 4).1521,24 indicative of NHL but corresponding FC showed light
chain restriction (Table 5). As reported above, in four
cases (1%) FNC remained inadequate after repeated
Results
passes. FNAC FC diagnoses were checked histologi-
FNAC FC provided a definitive diagnosis of NHL in cally in 130 cases (103 NHL, 18 BRH and nine sNHL).
245 cases (55%) and of BRH in 188 cases (42%) The remaining 312 cases (110 rNHL, 32 NHL and 170

Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd

304 P. Zeppa et al.

BRH) were checked on the basis of clinical follow-up.
The histological examination confirmed the FNAC FC
diagnosis in 102 cases of NHL and detected one false
positive (Table 2). As for the 18 BRH histologically
controlled, the cytological diagnosis was confirmed in
14 of the 18 cases and the remaining four were shown
to be NHL (false negatives). These false negatives
proved to be FL (two cases), MZL (one case) and
DLBCL (one case). All nine cases of sNHL were
confirmed to be NHL (two FL, one MZL, one PTCL,
three DLBCL and two NHL NOS) (Table 5). Clinical
follow-up confirmed the diagnosis in all of the
remaining 312 cases (110 rNHL, 32 NHL and 170
BRH).
Statistical analysis showed the following results:
94.9% sensitivity, 99.4% specificity, 99.6% PPV and
93.4% NPV in NHL and BRH discrimination; these
data are summarized in Table 2. These data were
compared with the results of the previous series
(Table 4). The 2 2 chi-square test run between
incorrect diagnoses (FN + FP) and the total of correct
diagnoses (TP + TN) in the present and our previous
series19 did not demonstrate significant differences
between the two series (v2 = 0.77; P > 0.2).
In terms of NHL classification according to the WHO
system,23 the evaluation of different expression and
co-expression of the fluoresceinated antibodies, com-
bined with the cytological features, suggested a
specific subtype in 125 cases (51%). Histological
diagnosis was available in 70 of these cases, of which
the subtype diagnosed on FNAC FC was correct in 67
(Table 3). There were three misclassified cases: one
diagnosed as FL, which was shown on histological
examination to be BL, and two cases diagnosed as
SLL CLL, of which one was shown histologically to be
MCL, and the other to be FL. All other cases in which
FC did not help in classification and or unspecific
cytological features were present were diagnosed only
as NHL or rNHL NOS; 33 of those had histological
biopsies and were taken as false negatives in this
context as explained above. Therefore a sensitivity of
67% was obtained.
The 2 2 chi-square test did not demonstrate
statistically significant differences in FNAC FC classi-
fication between the two series19 (v2 = 0.28; P > 0.5).
In order to verify the possible causes of error we
reviewed the 13 false negative (four truly false
negatives and nine suspicious) and the false positive
case. As regards the four truly false negative cases, in
two cases FNAC was polymorphous, reactive-like and
FC not contributory; in these cases we trusted FNAC
only because of the unsuspicious clinical context (case
1) (Figure 1a,b) or because FC showed light chain
polyclonality (case 2). In one case FNAC was mono-
morphous and suggestive for NHL but FC showed light
chain polyclonality (case 3). In one case (case 4),
FNAC was polymorphous, reactive-like (Figure 2), FC
showed light chain polyclonality but the histology
revealed partial involvement of the lymph node
(Figure 3a,b,c,d). As far as the nine suspicious cases
are concerned, they were determined by FNAC-
suggestive smears and FC non-contributory for tech-
nical reasons [scanty cellularity (cases 5, 6, 11), partial
involvement of the lymph node (case 12), light chain

Table 3. Correlation of fine needle aspiration cytology flow cytometry (FNAC FC) and histological classification of NHL
subtypes

Histology

FNAC FC No. SLL CLL MCL FL MZL MALT PTCL BL DLBCL NK

SLL CLL 4 2 1 1
MCL 7 7
FL 33 32 1
MZL MALT 3 3
PTCL 2 2
BL 4 4
DLBCL 14 14
NK 3 3
TOTAL 70 2 8 33 3 2 5 14 3

SLL CLL, small lymphocytic lymphoma chronic lymphatic leukaemia; MCL, mantle cell lymphoma; FL, follicular lymphoma;
MZL MALT, marginal malt B-cell lymphoma; PTCL, peripheral T-cell lymphoma; BL, Burkitt lymphoma; DLBCL, diffuse large
B-cell lymphoma; NK, natural killer non-Hodgkin lymphoma.

Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd

 Fine needle aspiration cytology and flow cytometry in non-Hodgkin lymphoma
(a)
(b)
Figure 1. Case 1. (a) Fine needle aspiration cytology smear
shows a dispersed polymorphous cell population represented
by follicle centre cells and small lymphocytes (460).
(b) Histological control shows a large lymphomatous follicle
with a very thin marginal zone (haematoxylin and eosin 270).
polyclonality (cases 7, 9, 10) and a T-cell NHL in
which FC for T lineage was not performed (case 8)]. In
one case FNAC was scantily cellular but FC was
indicative for NHL (case 13).
The false positive case concerned the right cervical
lymph node of a 40-year-old woman, human immu-
nodeficiency virus (HIV)-positive, stage C, for four
years. Two years previously the patient had been
diagnosed with follicular B-cell NHL; she had been
treated with four cycles of multi-agent chemotherapy
plus rituximab, the last cycle being administered
10 months previously. Fluorodeoxyglucose-positron
emission tomography with CT scan showed positivity
in the corresponding cervical area. Ultrasound-guided
FNAC showed an atypical lymphoid cell proliferation
(Figure 4a). FC phenotype showed CD10 19
co-expression and k light chain restriction (Figure 4b).
Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd
305
Figure 2. Case 4. Fine needle aspiration cytology smear
shows a dispersed polymorphous cell population represented
by medium-sized follicle centre cells and small lymphocytes
(270). Flow cytometry demonstrated polyclonality of this
cell population.
Since low lactate dehydrogenase values and a reduced
lymph node size were observed, the lymph node was
therefore excised; the histology revealed a reactive
hyperplastic lymph node, bcl-2, with florid follicular
pattern (Figure 4c,d). The patient is currently undergo-
ing strict clinical and instrumental follow-up and after
18 months is alive without recurrence of lymphoma.
Discussion
FC immunophenotyping combined with FNAC has
been demonstrated to be useful in the cytological
diagnosis and classification of NHL.121,24 Specifically,
FC is mainly effective in discriminating between BRH
and small cell NHL, and in their classification;
conversely, traditional FNAC contributes more to the
diagnosis of large cell NHL. The combination of the
two techniques therefore enhances their diagnostic
possibilities. In fact, when FNAC is limited, as in the
differentiation between small cell, low-grade NHL and
BRH, FC is effective; conversely, when FC is limited,
which is mainly in evaluating large cell NHL, FNAC
contributes more.11,13,18,21 Apart from improving
diagnostic efficiency, the combination of the two
techniques has generated fewer discordant results and
consequently more homogeneous data in different
laboratories (Table 4). This is a remarkable result
considering that the non-reproducibility of data from
different laboratories has been one of the historical
limitations of the application of FNAC to the diagnosis
of lymph nodal disorders.29 In fact, as reported above,
306 P. Zeppa et al.

(a) (c)

(b) (d)

Figure 3. Case 4. (a) Partial involvement of the lymph node by a large cell non-Hodgkin lymphoma: on the left a deeper located
lymphomatous area and on the right the reactive, subcapsular, non-involved area (haematoxylin and eosin 270).
(b) CD20 immunostain of the same area showing a sharp demarcation between the lymphomatous on the right and the
reactive, non-involved area on the left (CD20 immunostain 270). (c) Higher magnification of the deeper located,
lymphomatous area showing a monomorphous population of large, atypical nucleolated cells (haematoxylin and eosin 430).
(d) Higher magnification of the subcapsular reactive, non-involved area showing small lymphocytes lacking nuclear atypia
probably sampled at the time of the fine needle aspiration cytology (haematoxylin and eosin 430).

diagnostic sensitivity and specificity in the diagnosis of
NHL and rNHL obtained in different laboratories in the
period 19972007, including the results of our previ-
ous study, were quite similar (Table 4).1521,24 This
homogeneity of results is also remarkable considering
that, to the best of our knowledge, these studies
represent the first systematic and extensive clinical
applications of FNAC FC in the diagnosis of lympho-
proliferative processes in the laboratories concerned.
However, the rate of inadequate and suspicious cases
still ranges from 4% to 20% in the literature,11,18 and
it is possible that consolidated and prolonged experi-
ence has led to an improvement. Therefore, in this
study we tested FNAC FC in an additional series to
investigate whether our own and others previous
experience had produced further improvements in
diagnostic accuracy.
As reported above, in the present study we reported
94.9% sensitivity, 99.4% specificity, 99.6% PPV and
93.4% NPV in NHL and BRH discrimination. These
results are similar to those obtained in our previous
experience19 (Table 4). Moreover, statistical analysis
of the data did not demonstrate a significant difference
between the two series, confirming the homogeneity
of the results (v2 = 0.77; P > 0.2).
Further diagnostic improvement has been limited,
despite increased experience, by the presence in the
present series of nine cases of sNHL, four false nega-
tives and one false positive (Table 5), much as hap-
pened in our previous study.19 There was no
prevalence of a specific pathological entity among
these incorrect diagnoses, which were shown to consist
of FL (four cases), MZL (two cases), PTCL (one case),
DLBCL (four cases) and NHL NOS (two cases) by the
histological check. In seven cases FC did not demon-
strate light chain restriction, in three cases cytological
features were equivocal and an alternative diagnosis of
possible BRH was considered and finally, in three cases
neither FC nor cytological features were contributory.
Therefore, summarizing, defects of sampling, loss of
cells during processing, scanty cellularity and lack of
detection clonality were the most frequent causes of a
diagnosis of sNHL and false negative results, rather
than a specific NHL subtype.
Clonal B-cell populations in non-lymphomatous
processes have been rarely reported in reactive lymph

Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd

 Fine needle aspiration cytology and flow cytometry in non-Hodgkin lymphoma 307

(a) (c) (d)

(b)

Figure 4. Case 14. (a) Fine needle aspiration cytology smear shows a dispersed polymorphous cell population with a prevalence
of large cells with dispersed chromatin pattern and two or three marginal nucleoli (Diff Quik stain 460). (b) Flow cytometry
analysis showing CD10 CD19 co-expression in the upper right quadrant on the right and k light-chain restriction in the
upper left quadrant. (c) Histological control shows a florid follicular hyperplasia (haematoxylin and eosin 430). (d)
Immunohistochemical stain for bcl-2 showing negativity of a follicle centre (430).

Table 4. Fine needle aspiration cytology flow cytometry of lymph node and extranodal data. Comparison of different
publications

Sensitivity Specificity No of No NHL (% NHL precisely

Ref. Author year % % cases primary recurrent) classified, %

15 Dunphy and Ramos, 1997 80 100 73 60 (23 77) 71

16 Young et al., 1998 80 100 100 72 (40 60) 95
17 Jeffers et al., 1998 86 100 46 21 (76 24) 81
18 Meda et al., 2000 95 85 100* 290 199 (54 46) 70
11 Dong et al., 2001 76 Not performed 139 105 (84 55) 77
19 Zeppa et al., 2004 93 100 307 147 (48 52) 48
20 Bangerter et al., 2007 85 100 131 115 (unknown) Not performed
21 Swart et al., 2007 97 87 124 64 (unknown) Not performed
Present series 94.9 99.4 446 245 (55 45) 51

*Specificity was 100% when only definitive diagnoses of lymphoma were considered.

nodes4,7,3034 and have been mainly observed in
patients suffering from autoimmune or immunodefi-
ciency syndromes.3034 This phenomenon is consid-
ered a non-malignant, antigen-driven proliferation of
B lymphocytes determined by an abnormal response
to bacterial or viral antigen stimulation. Non-lympho-
matous clonal expansions are characterized by light
chain restriction with or without CD10 CD19
co-expression but should lack bcl-2 overexpression
and t1418 translocation, which allow this phenome-

Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd

308 P. Zeppa et al.

Table 5. Cases suggestive of NHL, false negatives and false positives in the present series

FNAC FC
Precedent suggestive indicative FNAC FC
Case Sex NHL of NHL of NHL diagnosis Histology Possible source of error

1 F ) ) ) BRH FL Cytological polymorphism, FC not

non- contributory (few gated cells)
2 F ) ) ) BRH MZL FNAC monomorphous with small cells,
FC: light chains polyclonal
3 M ) ) + ) BRH FL FNAC monomorphous, small cell. FC:
light chains polyclonal, CD10 19)
4 M ) ) ) BRH DLBCL Partial involvement; FNAC sampling of
reactive area
5 M ) + ) sNHL DLBCL Few diagnostic cells not detected by FC
6 M + + ) sNHL DLBCL Few diagnostic cells not detected by FC
7 M + ) sNHL FL Cytology suggestive, FC not
non-contributory
8 M ) + ) sNHL PTCL FNAC: malignant, FC not performed for
T-cell lineage
9 M + + ) sNHL MZL FNAC monomorphous, small cell. FC:
light chains polyclonal, CD10 19)
10 M ) + ) sNHL FL Cytology suggestive, FC not
non-contributory
11 F ) + ) sNHL DLBCL Few diagnostic cells not detected by FC
12 M ) + ) sNHL NHL NOS Cytology suggestive, FC not
non- contributory (few gated cells)
13 F ) ) + sNHL NHL NOS Smears scantly cellular, FC: light chains
restriction
14 + + + rNHL BHR Clonal expansion

BRH, benign reactive hyperplasia; sNHL, suggestive of NHL; FL, follicular lymphoma; MZL, marginal B-cell lymphoma; DLBCL,
diffuse large B-cell lymphoma; PTCL, peripheral T-cell lymphoma; NHL NOS, NHL not otherwise specified; rNHL, non Hodgkin
lymphoma relapse.

non to be distinguished from the so-called lymphoma
in situ.34 The present case occurred in a HIV patient
and was clinically complex because of the patients
previous history of FL. Therefore this case was finally
considered a non-lymphomatous clonal expansion
and suggests that careful attention needs to be paid
while evaluating radiological, cytological and FC data
along with clinical correlation in patients suffering
from autoimmune or immunodeficiency syndromes.
As regards FNAC FC classification of NHL, as
reported above we were able to classify 125 out of
245 cases (51%); the remaining 120 cases were
diagnosed only as NHL NOS. FNAC FC classification
was controlled by the previous or following histolog-
ical control of NHL and rNHL, respectively. Comparing
the incidence of the different phenotypes in the
present series with that referred to by the WHO,23
we observed a general agreement with the latter for
the proportion of all the single entities with the
exception of MZL MALT, which were underdiag-
nosed in the present series. This relative underesti-
mation of MZL MALT NHL could be explained by the
relatively non-specific cytological features and the
mute phenotype (CD5), CD10), CD23), FMC7))
that characterize these entities. Therefore, with the
exception of some extranodal MALT, some of the
MZL MALT NHL of the present series have probably
been diagnosed as NHL NOS on the basis of the FNAC
features and light chain restriction only. Statistical
analysis calculated on the basis of histological exam-
inations showed 67% sensitivity in the classification
of NHL. Comparing this result with our previous
series,19 we observed a slight improvement, mainly
determined by the correct classification of 14 cases of
DLBCL in the present series. In fact, CD19 positivity
and or aberrant phenotypes combined with specific
cytological features were considered specific for
DLBCL in the present series, while these cases were
considered as NHL NOS in our previous study,19 and
hence not classified.

Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd

 Fine needle aspiration cytology and flow cytometry in non-Hodgkin lymphoma
However, this improvement was not statistically
significant (v2 = 0.28; P > 0.5) because of the pres-
ence of three misclassified cases. As reported above,
these consisted of one FNAC FC diagnosis of FL,
which histological examination showed to be BL, and
two FNAC FC diagnoses of SLL CLL, shown to be one
case of MCL and one of FL (Table 3). In these cases,
loss of expression of specific antigens, equivocal
cytological features or partial lymph nodal involve-
ment seemed to be the cause of misclassification.
These values of accuracy in diagnosis and classification
probably represent the best results we are able to
obtain using FNAC FC. Additional ancillary tech-
niques such as polymerase chain reaction for clonality
determination or fluorescence in situ hybridization for
the identification of specific translocations might
further improve the sensitivity of cytology,35 either
in diagnosis or in classification, whereas we have to be
aware that the exiguity of samples that is character-
istic of FNAC, and lymph node partial involvement,
necrosis or fibrosis,36 will probably continue to limit
further improvements in FNAC FC diagnostic effi-
ciency.
Nonetheless, apart from the value of FNAC as a
basic tool for the lymph nodal diagnosis of clearly
reactive processes, HL or metastases that is tradition-
ally recognized, the combination of FNAC FC has
conferred further credibility on the whole procedure.
In fact a simple cytological diagnosis of NHL, without
any phenotypic or molecular integration, would only
be indicative and histological control would be man-
datory in all the cases. We believe that the pivotal role
of FNAC FC in the diagnosis of NHL lies in the
possibility of accurate diagnosis and classification in a
high percentage of cases. FNAC FC may simplify the
whole diagnostic procedure, avoiding surgical biopsies
and speeding up therapeutic procedures in rNHL and,
in selected cases, even in primary NHL. At the same
time, surgical excision and histological control are still
the second step and the gold standard for primary
diagnoses as well as for all cases in which FNAC FC
does not succeed. In a wider perspective, the results of
this study, while not showing significant improve-
ment in diagnostic sensitivity, overlap the data
reported in literature, thus demonstrating that the
method is highly reproducible and possibly strength-
ens the confidence of clinicians.
In conclusion, FNAC FC is effective in the cytolog-
ical diagnosis and classification of NHL and generates
reproducible results. At the same time, technical and
clinicopathological limitations represent a definite
Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd
309
obstacle to further improvements. FNAC FC should
be used as a basic tool in the diagnosis and classifi-
cation of NHL, in the choice of possible additional
ancillary techniques and to indicate the most conve-
nient target when a surgical biopsy is necessary.
Acknowledgments
The authors thank Dr Antonino Iaccarino, Dr Carmela
Frangella and Dr Maria Russo for their technical
assistance.
References
1. Johnson A, Akerman M, Cavallin-Stahl E. Flow cytometric
detection of B-clonal excess in fine needle aspirates for
enhanced diagnostic accuracy in non-Hodgkins lym-
phoma in adults. Histopathology 1987;11:58190.
2. Dunphy CH, Katz RL, Fanning CV, Dalton WT Jr.
Leukemic lymphadenopathy: diagnosis by fine needle
aspiration. Hematol Pathol 1989;3:3544.
3. Chernoff WG, Lampe HB, Cramer H et al. The potential
clinical impact of the fine needle aspiration flow cyto-
metric diagnosis of malignant lymphoma. J Otolaryngol
1992;21:115.
4. Robins DB, Katz RL, Swan F Jr et al. Immunotyping of
lymphoma by fine-needle aspiration. A comparative
study of cytospin preparations and flow cytometry. Am
J Clin Pathol 1994;101:56976.
5. Cha I, Goates JJ. Fine-needle aspiration of lymph nodes:
use of flow cytometry immunophenotyping. Pathology
1996;4:33764.
6. Jenning CD, Foon KA. Recent advances in flow cytom-
etry: application to the diagnosis of hematologic malig-
nancy. Blood 1997;90:286392.
7. Zardawi IM, Jain S, Bennet G. Flow-cytometric algo-
rithm on fine-needle aspirates for the clinical workup of
patients with lymphoadenopathy. Diagn Cytopathol
1998;19:2748.
8. Young NA, Al-Saleem T. Diagnosis of lymphoma by fine
needle aspiration cytology using the revised European-
American classification of lymphoid neoplasms. Cancer
(Cancer Cytopathol) 1999;87:32545.
9. Nicol TL, Silberman M, Rosenthal DL, Borowitz MJ. The
accuracy of combined cytopathologic and flow cytomet-
ric analysis of fine-needle aspirates of lymph nodes. Am J
Clin Pathol 2000;114:1828.
10. Siebert JD, Weeks MLT, List LW et al. Utility of flow
cytometry immunophenotyping for the diagnosis and
classification of lymphoma in community hospital clin-
ical needle aspiration biopsies. Arch Pathol Lab Med
2000;124:17929.
11. Dong HY, Harris NL, Preffer FI, Pitman MB. Fine-needle
aspiration biopsy in the diagnosis and classification of
310 P. Zeppa et al.
primary and recurrent lymphoma: a retrospective anal-
ysis of the utility of cytomorphology and flow cytometry.
Mod Pathol 2001;14:47281.
12. Kaleem Z, White G, Vollmer RT. Critical analysis and
diagnostic usefulness of limited immunophenotyping of
B-cell non-Hodgkin lymphomas and flow cytometry. Am
J Clin Pathol 2001;115:13642.
13. Liu K, Stern RC, Rogers RT, Dodd LG, Mann KP.
Diagnosis of hematopoietic processes by fine-needle
aspiration in conjunction with flow cytometry: a review
of 127 cases. Diagn Cytopathol 2001;24:110.
14. Saboorian MH, Ashfaq R. The use of fine needle
aspiration biopsy in the evaluation of lymphadenopathy.
Semin Diagn Pathol 2001;18:11023.
15. Dunphy CH, Ramos R. Combining fine-needle aspiration
and flow cytometric immunophenotyping in evaluation
of nodal and extranodal sites for possible lymphoma: a
retrospective review. Diagn Cytopathol 1997;16:2006.
16. Young NA, Al-Saleem TI, Ehya H, Smith MR. Utilization of
fine-needle aspiration cytology and flow cytometry in the
diagnosis and sub-classification of primary and recurrent
lymphoma. Cancer (Cancer Cytopathol) 1998;84:25261.
17. Jeffers MD, Milton J, Herriot R, McKean M. Fine needle
aspiration cytology in the investigation of non-Hodgkins
lymphoma. J Clin Pathol 1998;51:18996.
18. Meda BA, Buss DH, Woodruff RD et al. Diagnosis and
sub-classification of primary and recurrent lymphoma.
The usefulness and limitations of combined fine-needle
aspiration cytomorphology and flow cytometry. Am J
Clin Pathol 2000;113:68899.
19. Zeppa P, Marino G, Troncone G et al. Fine-needle
cytology and flow cytometry immunophenotyping and
subclassification of non-Hodgkin lymphoma. A critical
review of 307 cases with technical suggestions. Cancer
(Cancer Cytopathol) 2004;102:5565.
20. Bangerter M, Brudler O, Heinrich B, Griesshamnuer M.
Fine needle aspiration cytology and flow cytometry in
the diagnosis and subclassification of non-Hodgkins
lymphoma based on the World Health Organization
classification. Acta Cytol 2007;51:3908.
21. Swart GJ, Wright C, Brundyn K et al. Fine needle
aspiration biopsy and flow cytometry in the diagnosis
of lymphoma. Transfus Apher Sci 2007;37:719.
22. Ravinsky E, Morales C, Kutryk E, Chrobak A, Paraskevas
F. Cytodiagnosis of lymphoid proliferations by fine
needle aspiration biopsy. Adjunctive value of flow
cytometry. Acta Cytol 1999;43:10708.
23. Jaffe ES, Harris NL, Stein H et al. World Health Organiza-
tion Classification of Tumours: Pathology and Genetics of
Cytopathology 2010, 21, 300310 2010 Blackwell Publishing Ltd
Tumours of Haematopoietic and Lymphoid Tissues. Lyon:
IARC Press; 2001.
24. Wakely PE Jr. Fine-needle aspiration cytopathology in
diagnosis and classification of malignant lymphoma:
accurate and reliable? Diagn Cytopathol 2000;22:1205.
25. Zeppa P, Picardi M, Marino G et al. Fine-needle aspira-
tion biopsy and flow cytometry immunophenotyping of
lymphoid and myeloproliferative disorders of the spleen.
Cancer (Cancer Cytopathol) 2003;99:11827.
26. Dey P. Role of ancillary techniques in diagnosing and
subclassifying non-Hodgkins lymphomas on fine needle
aspiration cytology. Cytopathology 2006;17:27587.
27. Yao JL, Cangiarella JF, Cohen JM, Chhieng DC. Fine-
needle aspiration biopsy of peripheral T-cell lymphomas.
A cytologic and immunophenotypic study of 33 cases.
Cancer 2001;93:1519.
28. Loo CK, Quach HT, Gallo J. Diagnosis of natural killer cell
lymphoma by cytology and flow cytometric immuno-
phenotyping. A case report. Acta Cytol 2002;46:87782.
29. Hehn ST, Grogan TM, Miller TP. Utility of fine-needle
aspiration as a diagnostic technique in lymphoma. J Clin
Oncol 2004;22:304652.
30. Cartagena N Jr, Katz RL, Hirsch-Ginsberg C et al. Accu-
racy of diagnosis of malignant lymphoma by combining
fine-needle aspiration cytomorphology with immunocy-
tochemistry and in selected cases, Southern blotting of
aspirated cells: a tissue-controlled study of 86 patients.
Diagn Cytopathol 1992;8:45664.
31. Kussick SJ, Kalnoski M, Braziel RM, Wood BL. Promi-
nent clonal B-cell populations identified by flow cytom-
etry in histologically reactive lymphoid proliferations.
Am J Clin Pathol 2004;121:46472.
32. Nam-Cha SH, San-Millan B, Mollejo M et al. Light-chain-
restricted germinal centres in reactive lymphadenitis:
report of eight cases. Histopathology 2008;52:43644.
33. Bhargava P, Parker JA, Dezube BJ. Pitfalls of diagnosis
based on abnormal flow cytometry and [(18)F] fluoro-
deoxyglucose positron emission tomography. Clin Lym-
phoma Myeloma 2008;8:117120.
34. Cong P, Raffeld M, Teruya-Feldstein J et al. In situ
localization of follicular lymphoma: description and
analysis by laser capture microdissection. Blood
2002;99:337682.
35. Kocjan G. Best Practice No. 185. Cytological and molec-
ular diagnosis of lymphoma. J Clin Pathol 2005;58:5617.
36. Jorgensen JL. State of the Art Symposium: flow cytom-
etry in the diagnosis of lymphoproliferative disorders by
fine-needle aspiration. Cancer 2005;105:44351.