Hematopathology / LUNG INVOLVEMENT BY LYMPHOMAS
Histologic Patterns of Lung Infiltration of B-Cell, T-Cell,
and Hodgkin Lymphomas
Maria B.G. Costa, MD, PhD,1 Sheila A.C. Siqueira, MD,2 Paulo H.N. Saldiva, MD, PhD,2
Klaus F. Rabe, MD, PhD,3 and Thais Mauad, MD, PhD2,3
Key Words: Lung infiltration; Autopsy; Hodgkin lymphoma; B-cell lymphoma; T-cell lymphoma
DOI: 10.1309/R4BRRADYF20K49HE
Abstract
Secondary lung infiltration by lymphomas occurs
frequently. To our knowledge, however, no recent
studies have attempted to discriminate histologic
patterns of lung infiltration in the lymphoma subtypes.
We retrospectively evaluated the frequency of lung
infiltration and the respective infiltration patterns by
lymphomas at autopsy, during an 11-year period.
Lymphomas were classified according to the 2001
World Health Organization Classification of
hematologic malignancies in B-cell, T-cell, and
Hodgkin lymphomas (HLs). In 21,157 autopsies, 414
reports with lymphoma diagnosis were reviewed
histologically, and 85 showed lung infiltration (20.5%).
We studied 14 HLs, 43 B-cell lymphomas, and 20 T-cell
lymphomas. Five infiltration patterns were identified:
peribronchial-perivascular, nodular, alveolar,
interstitial, and pleural. Approximately half of the
lymphomas had more than 1 infiltration pattern (mean,
1.7); peribronchial-perivascular and pleural were the
most frequent. The frequency of nodular infiltration was
larger in HL than in B-cell lymphomas. T-cell
lymphomas had a larger frequency of the interstitial
infiltration pattern compared with B-cell lymphomas.
Recognizing the frequency and patterns of lung
infiltration in the light of a more recent classification is
certainly useful for physicians dealing with lymphoma
diagnostic procedures, such as radiologists and
pathologists.
718
718
Am J Clin Pathol 2004;121:718-726
DOI: 10.1309/R4BRRADYF20K49HE
Primary pulmonary lymphomas are very uncommon,
whereas secondary infiltration of the lung by hematologic
malignancies is a frequent finding. Primary pulmonary
lymphomas represent 3% to 4% of the extranodal nonHodgkin lymphomas (NHLs) and only 0.5% of all primary
pulmonary malignant neoplasms.1 On the other hand, autopsy
studies revealed that the lung, after the liver, is the second
organ involved in the extranodal dissemination of these
disease processes.2 It has been shown that parenchymal lung
involvement will develop in 38% and 24% of the patients
with Hodgkin lymphomas (HL) and NHL, respectively,
during the disease.3
Secondary infiltration of the lungs can occur by
hematogenic dissemination or by contiguous invasion
from a hilar or mediastinal site of the lymphoma.1 It is
believed that the intrapulmonary distribution of the
lymphomas is determined by the location of the lymphoid
tissue in the lung. The lymphatic drainage of the lungs
occurs along peribronchial and perivascular axes, and
there is organized lymphoid tissue at the bifurcation of
bronchi and pulmonary vessels.4 Pleural effusions might
occur by one or a combination of the following 3 mechanisms: pleural infiltration by the lymphoma, obstruction
of the lymphatics draining the pleura by enlarged intramediastinal lymph nodes, and obstruction of the thoracic
duct.3 Pleural effusions are considered a bad prognostic
factor in NHL.3
Pulmonary patterns of lymphomatous infiltration have
been characterized radiologically as bronchovascularlymphangitic, alveolar, interstitial, and nodular.5-7 These
patterns were demonstrated to have a good correlation
with pathologic appearances in some studies.8 Although
© American Society for Clinical Pathology
Hematopathology / ORIGINAL ARTICLE
pulmonary involvement in NHL frequently occurs in
cases of disseminated disease,9 some authors believed that
the anatomic pattern of lung involvement could be a
biologic predictor of its behavior. 4,10 For instance,
Mentzer et al6 demonstrated that patients with a nodular
lesion have a 30% better prognosis than those with alveolar parenchymal disease. In these studies, no prognostic
relation could be established with the histologic type.
However, most of the studies regarding the secondary
infiltration of the lungs by lymphomas date to 30 years
ago. Some of these pathologic studies used now-historic
lymphoma classifications such as the Rappaport classification. During the last 30 years, major progresses in the
field of hematopathology occurred, culminating in the
World Health Organization (WHO) Classification of
hematologic malignancies in 2001.11 This classification
has delineated well-characterized clinical-pathologic entities and grouped the lymphomas in 3 major categories (Tcell, B-cell, and HL).
To our knowledge, there are no studies analyzing lung
involvement by different types of lymphoma based on this
recent classification. In this study, our aim was to discriminate whether WHO Classification entities also could behave
differently concerning lung infiltration. Recognizing the
patterns of lung infiltration of the different lymphomas in the
light of the new classification might be useful for diagnostic
(eg, recognition of radiologic patterns) and prognostic
purposes. Therefore, we retrospectively analyzed the
frequency and the histologic patterns of lung infiltration by
B- and T-cell lymphomas and HL at autopsy, comprising a
period of 11 years.
A
Materials and Methods
Autopsies
We retrieved from the files of the Department of
Pathology, São Paulo University, São Paulo, Brazil, all
autopsy reports with a diagnosis of HL or NHL comprising
the period between January 1988 and July 1999. We further
discriminated the presence of AIDS as an underlying disease
and the occurrence of primary pulmonary lymphomas and
lymphomatoid granulomatosis (LYG). Patients’ ages
(expressed as median) and sex were obtained. For some
cases, we obtained clinical data related to the presence of
dyspnea, cough, and abnormal pulmonary auscultation
during the last admission. The São Paulo University Medical
School Institutional Ethical Committee approved the study.
Tissue Analysis
We reviewed H&E-stained slides of all hematologic
autopsies and scored for the presence of lung infiltration.
There were 1 to 3 slides containing lung tissue per autopsy.
Exclusion criteria were associated infections. Regarding the
positive cases, further subclassification according to the
following histologic infiltration patterns was done: (1)
nodular, unique or multiple, confluent ❚Image 1❚; (2) peribronchial-perivascular, when the lymphoma cells were
present along bronchial and vascular bundles ❚Image 2❚; (3)
alveolar, when lymphoma cells filled the alveolar spaces in a
pneumonic pattern ❚Image 3❚ ; (4) interstitial, when
lymphoma cells infiltrated predominantly along the alveolar
septa ❚Image 4❚; and (5) pleural, when lymphoma cells were
present in the pleura ❚Image 5❚.
B
❚Image 1❚ Hodgkin lymphoma, mixed cellularity subtype. A, Nodular pattern of lung infiltration. Tumor infiltration consists of a
solid mass, forming a nodule (H&E, ×1.6). B, Detail of Image 1A. In the tumor tissue there are large atypical mononucleated or
binucleated cells with prominent nucleoli, characterizing Hodgkin cells (arrows) (H&E, ×400).
Am J Clin Pathol 2004;121:718-726
© American Society for Clinical Pathology
719
DOI: 10.1309/R4BRRADYF20K49HE
719
719
Costa et al / LUNG INVOLVEMENT BY LYMPHOMAS
Lymphoma classification for each case was obtained
from the original biopsy or excision material, by consulting
the surgical pathology files of the Department of Pathology.
All cases were reviewed by a hematopathologist (S.A.C.S.)
who classified the archived material according to the WHO
classification, using new diagnostic immunohistochemical
analysis panels ❚Table 1❚ and histologic sections when indicated. Lymphomas were grouped as HL, B-cell, and T-cell
neoplasms.
As for the immunohistochemical technique, standard
procedures were used. Briefly, sections were deparaffinized,
A
and a 0.5% peroxidase in methanol solution was applied for
10 minutes to inhibit endogenous peroxidase activity. Antibody binding was detected with a streptavidin-biotin-peroxidase immunohistochemical system (LSAB Kit, DAKO Duet
System, K0675, DAKO, Carpinteria, CA) and revealed with
diaminobenzidine. Positive and negative control slides were
used in each staining run. Slides were counterstained with
Harris hematoxylin.
We were able to retrieve the results of the oncocytologic
analyses of pleural effusions studied by means of cytomorphologic and cytochemical examination in 16 cases. Pleural
B
B
❚Image 2❚ A, Diffuse large B-cell lymphoma. Peribronchial-vascular pattern of lung infiltration. Lymphoma cells are present along
the bronchovascular bundles (arrows) (H&E, ×1.6). B, Follicular B-cell lymphoma. Peribronchial pattern of lung infiltration.
Lymphoma cells infiltrate all layers of the bronchiole (B) (H&E, ×100).
A
B
A
AS
❚Image 3❚ Diffuse large B-cell lymphoma. A, Alveolar pattern of lung infiltration. Lymphoma cells fill the alveolar spaces,
conferring a pneumonic aspect to the tissue (H&E, ×100). B, Detail of Image 3A. Note the large lymphocytes filling the alveolar
spaces (A) (H&E, ×400). AS, Alveolar septa.
720
720
Am J Clin Pathol 2004;121:718-726
DOI: 10.1309/R4BRRADYF20K49HE
© American Society for Clinical Pathology
Hematopathology / ORIGINAL ARTICLE
effusion results were compared with the presence or absence
of pleural infiltration. We identified the interval in days
between the procedure and the autopsy.
Statistical Analysis
Patterns of lung infiltration were expressed as percentages. The χ2 or the Fisher exact test was used to compare the
frequency of histologic type and patterns of lung infiltration.
P values smaller than .05 were considered significant. SPSS,
version 11.0 (SPSS, Chicago, IL) was used for statistical
analysis.
A
Results
We analyzed the reports of 21,157 autopsies performed
between January 1988 and July 1999, and 414 cases
(1.957%) in which the diagnosis of primary disease was
NHL or HL: 82 autopsies had HL and 332 had NHL as the
diagnosis. Of the 414 cases, 85 (20.5%) of them showed
lung infiltration. The frequency of infiltration was 17%
(14/82) for HL cases and 21.4% (71/332) for the NHL
cases. We were not able to study 8 of 71 NHLs, mostly
because the original biopsy specimens were not available in
B
A
AS
❚Image 4❚ Peripheral T-cell lymphoma. A, Interstitial pattern of lung infiltration. Lymphoma cells infiltrate along the alveolar septa
that become thickened (H&E, ×100). B, Detail of Image 4A. The alveolar septa (AS) become thickened owing to the infiltration
of lymphoma cells (H&E, ×200). A, alveolar space.
A
B
P
❚Image 5❚ Diffuse large B-cell lymphoma. A, Pleural infiltration pattern. Lymphoma cells are present in the pleural compartment
(P), which is visibly thickened (H&E, ×50). B, Detail of Image 5A. Large lymphoid cells are infiltrating the pleura (H&E, ×400).
Am J Clin Pathol 2004;121:718-726
© American Society for Clinical Pathology
721
DOI: 10.1309/R4BRRADYF20K49HE
721
721
Costa et al / LUNG INVOLVEMENT BY LYMPHOMAS
❚Table 1❚
Antibodies Used in the Study
Antigen
Source
CD3
CD5
CD10
CD15
CD20
CD30
CD43
CD45RO
bcl-2
Ki-67
Epithelial membrane antigen
Leukocyte common antigen
AE1/AE3
Anaplastic lymphoma kinase-1
Cyclin D1
Clone
DAKO A/S, Glostrup, Denmark
Novocastra, Newcastle upon Tyne, England
Novocastra
DAKO A/S
DAKO, Carpinteria, CA
DAKO A/S
Novocastra
DAKO A/S
DAKO A/S
DAKO A/S
DAKO A/S
DAKO
DAKO
DAKO A/S
DAKO A/S
the Department of Pathology or material was not appropriate for adequate diagnosis. Therefore, among the NHLs,
there were 43 B-cell and 20 T-cell lymphomas. In 2 HLs,
the primary diagnosis was made at the autopsy.
❚Table 2❚ shows patient age and sex and the infiltration patterns of all B- and T-cell lymphomas and HLs. Of
the lymphomas, 38 (49%) of 77 showed 2 or more infiltration patterns; the average number of infiltrations per
case was 1.7 ❚Image 6❚. Peribronchial-vascular involvement was the most frequent and alveolar the least frequent
infiltration pattern (Table 2). HLs had a higher frequency
of nodular involvement than did B- and T-cell
lymphomas. The χ2 test revealed a trend of .07 for the
nodular pattern when all 3 subtypes were compared and
statistical significance when HL was compared with Band T-cell lymphoma combined (P = .023, χ2; P = .033,
Fisher exact test). There was a larger frequency of the
nodular infiltration pattern in the HLs than in the B-cell
lymphomas (P = .036, χ2; P = .059, Fisher exact test) and
a similar trend between HL and T-cell lymphomas (P =
.048, χ2; P = .08, Fisher exact test). There was a larger
frequency of interstitial infiltration in the T-cell lymphomas
than in the B-cell lymphomas (P = .056, χ2; P = .085,
Dilutions
Polyclonal
4c7
56 C6
C3 D -1
L26
Ber-H2
DFT 1
UCHL-1
124
KiS5
E29
2 B11 + P07/26
AE1/AE3
ALK-1
DCS-6
1:3,000
1:200
1:400
1:50
1:800
1:400
1:1,400
1:1,000
1:500
1:400
1:800
1:800
1:200
1:20
1:200
Fisher exact test). The pleural infiltration pattern was
common in the 3 groups, and pleural infiltration alone
was present in 5 (7%) of 75 cases.
❚Table 3❚ shows the patient age and sex and infiltration
patterns for the B-cell subtypes. Diffuse large B-cell
lymphoma (DLBCL) was the most prevalent type, followed
by small lymphocytic lymphoma/chronic lymphocytic
leukemia (SLL/CLL) and Burkitt lymphoma (4 patients).
SLL/CLL had a lower frequency of nodular involvement than
the other types, except follicular lymphoma. There was statistical significance when nodular involvement was compared in
the 2 major subgroups SLL/CLL and DLBCL (P = .003, χ2;
P = .04, Fisher exact test). There was no pleural infiltration in
any of the 4 Burkitt lymphoma cases in this series. There was
a lower frequency of the pleural infiltration pattern in the
Burkitt lymphomas compared with DLBCL and SLL/CLL
cases (P = .001 and P = .011, respectively, χ2; P = .002 and P
= .023, respectively, Fisher exact test).
Among the T-cell neoplasms, the most common type
infiltrating the lungs was peripheral T-cell lymphoma, not
otherwise specified, followed by adult T-cell
lymphoma/leukemia and anaplastic cell lymphoma. ❚Table 4❚
shows patient age and sex and infiltration patterns in the lung
❚Table 2❚
Number of B-Cell, T-Cell, and Hodgkin Lymphomas That Showed Lung Infiltration at Autopsy (N = 77)*
Infiltration Pattern
Histologic
Lymphoma Type
No. (%)
of Cases
Median
Age (y)
Sex Ratio
(M/F)
PeribronchialVascular
Nodular
Alveolar
Interstitial
Pleural
B-cell
T-cell
Hodgkin
43 (56)
20 (26)
14 (18)
57
38
27
33:10
12:8
11:3
30 (70)
15 (75)
11 (79)
14 (32)
6 (30)
9 (64)
6 (14)
6 (30)
4 (29)
11 (25)
10 (50)
5 (36)
31 (76)†
11 (55)
9 (64)
*
†
Data are given as number (percentage) unless otherwise indicated.
Pleural tissue was not represented in tissue sections of 2 B-cell lymphomas, so the percentage is based on 41 cases.
722
722
Am J Clin Pathol 2004;121:718-726
DOI: 10.1309/R4BRRADYF20K49HE
© American Society for Clinical Pathology
Hematopathology / ORIGINAL ARTICLE
for the T-cell lymphoma subtypes. The adult T-cell lymphoma
group had a higher frequency of interstitial involvement than
the other types. However, the Fisher exact test performed
among the 3 major T-cell lymphoma groups yielded no significant differences, probably because of the small sample.
Of 82 HLs, 14 showed lung infiltration. The most
common subtype infiltrating the lungs was the mixed cellularity subtype, followed by the lymphocyte depletion subtype.
❚Table 5❚ shows patient age and sex and infiltration patterns
for the classic HL subtypes. The distribution of the infiltration
patterns was similar between the 2 principal subtypes.
We identified 1 case of primary pulmonary lymphoma in
this series, a mature peripheral T-cell lymphoma, from the
subtype adult T-cell lymphoma/leukemia, human Tlymphotropic virus (HTLV)-1+. Three patients had AIDSassociated lymphoma, 2 with the DLBCL subtype and 1 with
Burkitt lymphoma, with plasmacytoid differentiation. All
patients with AIDS had disseminated lymphoma at autopsy.
We identified 4 patients with clinically diagnosed LYG of
different primary locations: skin, lungs, brains, and nose. At
autopsy, 3 had infectious or congestive abnormalities of the
B
P
❚Image 6❚ Follicular B-cell lymphoma. Mixed pattern of lung
infiltration, panoramic view. Lymphoma infiltration occurred
along the bronchovascular bundles and in the pleura, which
is thickened (H&E, ×1.6). B, bronchiole; P, pleura.
❚Table 3❚
Histologic Subtypes of B-Cell Lymphomas That Showed Lung Infiltration at Autopsy (n = 43)*
Infiltration Pattern
Histologic Subtype
DLBCL†
SLL/CLL
Burkitt lymphoma/leukemia§
Follicular lymphoma
Mantle cell lymphoma
No. (%) of
Cases
Median
Age (y)
Sex Ratio
(M/F)
PeribronchialVascular
Nodular
Alveolar
Interstitial
18 (42)
15 (35)
4 (9)
3 (7)
3 (7)
57
64
30
61
54
13:5
11:4
3:1
3:0
3:0
14 (78)
10 (67)
2 (50)
3 (100)
1 (33)
10 (56)
1 (7)
2 (50)
0 (0)
1 (33)
3 (17)
2 (13)
0 (0)
0 (0)
1 (33)
4 (22)
6 (40)
0 (0)
0 (0)
1 (33)
Pleural
16 (89)
10 (71)‡
0 (0)
3 (100)
2 (100)‡
DLBCL, diffuse large B-cell lymphoma; SLL/CLL, small lymphocytic lymphoma/chronic lymphocytic leukemia.
* Data are given as number (percentage) unless otherwise indicated.
† Two patients had AIDS.
‡ Pleural tissue was not represented in 1 patient, so n = 14 for SLL/CLL and n = 2 for mantle cell lymphoma.
§ One patient had AIDS, with the plasmacytoid variant.
❚Table 4❚
Histologic Subtypes of T-Cell Lymphomas That Showed Lung Infiltration at Autopsy (n = 20)*
Infiltration Pattern
Histologic Subtype
Peripheral T-cell lymphoma, NOS
Anaplastic lymphoma, T-cell/null type
Adult T-cell lymphoma/leukemia (HTLV-1+)†
Angioimmunoblastic lymphoma
Precursor T lymphoblastic lymphoma
Enteropathy type T-cell lymphoma
Mycosis fungoides/Sézary syndrome
No. (%)
of Cases
Median
Age (y)
6 (30)
4 (20)
4 (20)
2 (10)
2 (10)
1 (5)
1 (5)
32
58
29
45
29
23
67
Sex Ratio Peribronchial(M/F)
Vascular
Nodular
1:5
3:1
3:1
1:1
2:0
1:0
1:0
5 (83)
4 (100)
1 (25)
2 (100)
2 (100)
1 (100)
0 (0)
4 (67)
0 (0)
1 (25)
0 (0)
1 (50)
0 (0)
0 (0)
Alveolar
Interstitial
3 (50)
2 (50)
0 (0)
0 (0)
1 (50)
0 (0)
0 (0)
2 (33)
2 (50)
4 (100)
1 (50)
1 (50)
0 (0)
0 (0)
Pleural
3 (50)
2 (50)
1 (25)
1 (50)
2 (100)
1 (100)
1 (100)
HTLV, human T-lymphotropic virus; NOS, not otherwise specified.
* Data are given as number (percentage) unless otherwise indicated.
† Includes 1 case of primary pulmonary lymphoma.
Am J Clin Pathol 2004;121:718-726
© American Society for Clinical Pathology
723
DOI: 10.1309/R4BRRADYF20K49HE
723
723
Costa et al / LUNG INVOLVEMENT BY LYMPHOMAS
❚Table 5❚
Histologic Subtypes of Classic Hodgkin Lymphomas That Showed Lung Infiltration at Autopsy (n = 14)*
Infiltration Pattern
Histologic Subtype
Mixed cellularity
Lymphocyte depletion
Nodular sclerosis
*
No. (%)
of Cases
Median
Age (y)
Sex Ratio
(M/F)
PeribronchialVascular
Nodular
Alveolar
Interstitial
Pleural
7 (50)
6 (43)
1 (7)
22
35
16
7:0
4:2
0:1
5 (71)
5 (83)
1 (100)
4 (57)
5 (83)
0 (0)
3 (43)
1 (17)
0 (0)
3 (43)
1 (17)
1 (100)
5 (71)
4 (67)
1 (100)
Data are given as number (percentage) unless otherwise indicated.
lungs. One patient had extensive angiocentric necrotic lesions
of the lungs with secondary bacterial infection, but no
remnants of an atypical lymphoid infiltrate were observed.
In 49 cases, we were able to obtain clinical data related
to pulmonary abnormalities during the last admission: 28
(57%) had dyspnea, 18 (37%) had a cough, and 35 (71%)
had altered pulmonary auscultation findings. There was no
significant association of lymphoma subtypes and patterns of
lung infiltration with these clinical parameters.
In 5 of 16 cases in which cytologic examination of
pleural effusions was done, there was concordance between
pleural effusion and pleural findings (31%). In 2 cases, oncocytologic analysis was positive, but no pleural infiltration
could be detected at autopsy. Nine patients had pleural infiltration at autopsy but negative pleural fluids. The median
number of days between the procedure and death was not
different among positive (8 days) and negative (8 days)
pleural effusions (P = .280, χ2; P = .528, Fisher exact test).
Discussion
In this large, 11-year, retrospective autopsy study, we
found that the lung was infiltrated in 20.5% of patients with a
previous diagnosis of HL and NHL. To our knowledge, we are
the first to attempt to verify the pattern of lung infiltration in
different histologic types according to the recent WHO Classification. Our results show that, at autopsy, lung infiltration is
predominantly peribronchial-perivascular and pleural in all 3
subtypes examined. Our results confirm previous radiologic
studies5,12 showing that frequently more than 1 infiltration
pattern is present in the lungs infiltrated by lymphomas. We
further observed that, in comparison with B-cell lymphomas,
HLs have a large frequency of the nodular infiltration pattern.
T-cell lymphomas, in turn, have a larger frequency of the interstitial pattern than do the B-cell lymphomas.
Similar studies date back to the 1970s and 1980s. Most
of them analyzed patterns of lung infiltration through radiologic methods and used histologic examination to confirm
the lymphoma diagnosis. 6 Furthermore, most of these
studies used historic nomenclatures such as the Rappaport
724
724
Am J Clin Pathol 2004;121:718-726
DOI: 10.1309/R4BRRADYF20K49HE
classification, making comparisons with present data difficult.6 Mentzer et al6 reviewed the medical records of 651
patients with lymphoma and identified lung involvement in
8.4% of the surgical specimens. The lung was involved
secondarily in 61% of the cases, and diffuse NHL was the
predominant lymphoma type. These authors verified that the
histologic diagnosis did not predict the gross anatomic
pattern of lung involvement, analyzed by means of radiologic techniques.6 In our study, we detected similar percentages of lung involvement by NHL and HL (21.4% [71/332]
and 17% [14/82], respectively). Although some authors
assert that parenchymal lung involvement is more common
in HL than in NHL,3 our study and other autopsy studies6
have shown similar frequencies of infiltration.
The majority of the studies dealing specifically with
lung infiltration by different B- and T-cell lymphomas are
case reports.10,13,14 In our series, we were able to analyze 43
B-cell and 19 T-cell lymphomas that had lung infiltration at
autopsy. DLBCLs and SLL/CLLs accounted for the majority
of the B-cell lymphomas in our series. When comparing
SLL/CLL with DLBCL (Table 3), a distinguishing pattern
was the low frequency of the nodular infiltration pattern in
SLL/CLL. None of the 4 Burkitt lymphomas studied in this
series showed pleural infiltration. Nzeh15 stated that pleural
involvement also was uncommon in Nigerian children with
Burkitt lymphoma, while it seems to be more frequent in
white patients.16 In Brazil, Burkitt lymphoma manifests with
characteristics intermediate between the epidemic and
sporadic types of this lymphoma.17
Peripheral T-cell lymphomas seem to involve the lung
frequently, 20% at diagnosis and a further 20% during the
course of the disease.18 In our study, T-cell lymphomas represented 26% of the lymphomas that had lung infiltration
(20/77). This percentage might reflect an increased incidence
of T-cell lymphomas in Brazil, although no official data are
available. In fact, a single-center study published in a Brazilian
journal,19 in which 145 NHLs were classified into subtypes,
revealed that 22.7% of the cases were T-cell lymphomas,
contrasting with 12% worldwide.20 HTLV-1 infection is
endemic in some regions of Brazil, and this fact could partially
explain the high incidence of T-cell disorders in our country.21
© American Society for Clinical Pathology
Hematopathology / ORIGINAL ARTICLE
T-cell lymphomas represent several entities that have clinicoradiologic features different from those of lymphomas with
the B-cell phenotype. A distinguishing feature of the T-cell
lymphomas in this study was an increased frequency of the
interstitial pattern compared with the B-cell lymphomas. Interestingly, there were 4 cases of anaplastic T-cell lymphoma
infiltrating the lungs. In a large series of 146 patients with
anaplastic large cell lymphoma, lung infiltration without mediastinopathy was one of the peculiarities of this type of
neoplasia compared with other cases of nonanaplastic large Bcell lymphoma.22 All the adult T-cell lymphomas/leukemias
had an interstitial pattern of involvement, which was more
frequent in the leukemic lung infiltrates.23
HL frequently infiltrates the lungs, in up to 20% at diagnosis and to 40% during the clinical course of the disease,
and several radiologic studies have examined the patterns of
intrathoracic infiltration in this lymphoma.7 In an autopsy
study, Colby et al24 demonstrated that 39% of patients with
HL had lung involvement. Diederich et al7 demonstrated that
radiologically, the most common infiltration pattern was
nodules (88% of the computed tomography scans). In the
present series, the lungs were infiltrated in 17% of the
patients with HL (14/82), a percentage considerably less than
in the study by Colby et al24 but similar to that found by
Mentzer et al6 (20%). We confirmed that the nodular pattern
is a characteristic of lung infiltration in HL, but no differences could be detected in the subtypes. HL infiltrates
commonly are associated with a fibrohistiocytic reaction, a
fact that possibly can explain the frequent nodular infiltration
pattern observed in the present study.
Some studies have reported predominance of the
nodular sclerosis HL subtype infiltrating the lungs,7,24 while
our study indicated predominance of the mixed cellularity
subtype. Such a difference might be explained by the distribution of the different HL subtypes in Brazil, ie, a predominance of the mixed cellularity subtype rather than a special
predilection of this subtype to infiltrate the lungs.25
Primary pulmonary lymphomas are uncommon. They
are low-grade B-cell lymphomas in 80% of the cases, especially the bronchus-associated lymphoid tissue–associated
lymphomas, and have a low mortality rate.26 We were able to
identify 1 case of primary pulmonary lymphoma in this
series, confirming its rarity and low mortality rates, and it
was a T-cell lymphoma, HTLV-1+, infiltrating primarily the
lungs, an entity that has been described rarely.27
In our series, the lymphomas infiltrating the lungs in
AIDS were, as expected, high-grade B-cell lymphomas,
including 1 case of the new WHO subcategory Burkitt
lymphoma with plasmacytoid differentiation. Lung infiltration
occurred as a result of extensive disease dissemination in all
patients. The introduction of highly active antiretroviral
therapy in AIDS seems to be related to an increase in
pulmonary involvement by lymphomas,28 and pathologists
and radiologists should be prepared to recognize these entities.
LYG was described classically as an angiocentric and
angiodestructive lymphoreticular proliferation frequently
involving the lungs.29 It is now accepted that, in most cases,
LYG is an Epstein-Barr virus–related B-cell lymphoma with
a prominent T-cell response. In the WHO classification, LYG
is considered a variant of DLBCL,11 although T-cell LYG
also has been described.30 In our study, 1 patient with LYG
had extensive necrotic lesions of the lung. Although we
could not identify remnants of the atypical infiltrate at
autopsy, the histologic pattern of angiocentric necrotic
lesions is characteristic of this entity.
The presence of a pleural effusion in a patient with
lymphoma is associated with a poor overall prognosis.3
There is some controversy in the literature whether pleural
effusion in lymphomas is caused by direct tumor invasion of
the pleura or by lymphatic obstruction.31,32 In our series, the
pleural infiltration pattern was a common finding in all Band T-cell lymphomas and HLs (31/41 [76%], 11/20 [55%],
and 9/14 [64%], respectively; Table 2) and was accompanied
by pulmonary parenchymal involvement in the large
majority of cases.
One autopsy study showed that 50% of patients with HL
and 72% of patients with NHL with pleural effusion had
malignant pleural infiltration.33 Obviously, these populations
most likely reflect cases of extreme disease dissemination and
cannot be compared with cases of limited disease. The results
suggest, however, that pleural effusion could be the result of
tumor invasion in the majority of the cases. The fact that
pleural involvement alone without parenchymal involvement
occurred in a few cases (5/75 [7%]) confirms the fact that
pleural involvement alone is an unusual finding in
lymphomas.31 In our series, there was concordance between
positive pleural effusions and pleural infiltration at autopsy in
5 (31%) of 16 cases. The analysis of lymphomatous pleural
effusions might pose diagnostic difficulties with reactive
conditions, and some studies have shown that the use of ancillary techniques such as immunohistochemical analysis and
flow cytometry increase the sensitivity of this procedure.34
Indeed, 2 cases had positive results in oncocytologic analyses,
but no lymphoma infiltration was detected at autopsy.
A limitation of this study lies in the fact that it reflects
cases of extreme disease dissemination, and some of the
results have to be interpreted in this context, such as the
multiple patterns of lung infiltration in approximately half of
the cases. Also, detailed knowledge about clinical and radiologic data would have enriched our data. Our findings might
be used as a basis for prospective radiologic studies.
In this autopsy study, we determined the frequency of lung
infiltration and described the histologic infiltration patterns of
lymphomas, classified according to the WHO Classification.
Am J Clin Pathol 2004;121:718-726
© American Society for Clinical Pathology
725
DOI: 10.1309/R4BRRADYF20K49HE
725
725
Costa et al / LUNG INVOLVEMENT BY LYMPHOMAS
Increasing this kind of knowledge clearly seems desirable for
the physicians dealing with the diagnosis of pulmonary infiltration in lymphomas, such as radiologists and pathologists.
From the Departments of 1Pathology, Division of Clinical
Laboratory and 2Pathology, São Paulo University, São Paulo,
Brazil; and 3Pulmonology, Leiden University Medical Center,
Leiden, the Netherlands.
Supported by Fundação de Amparo à Pesquisa do Estado de
São Paulo and Laboratório de Investigação Médica 03 do
Hospital das Clinicas da FMUSP.
Address reprint requests to Dr Mauad: Dept of Pathology,
Medical School, São Paulo University, Avenida Dr Arnaldo, 455
1st Floor, 01246-903, São Paulo, SP, Brazil, or [email protected].
References
1. Cadranel J, Wislez M, Antoine M. Primary pulmonary
lymphoma. Eur Respir J. 2002;20:750-762.
2. Risdall R, Hoppe RT, Warnke R. Non-Hodgkin’s lymphoma:
a study of the evolution of the disease based upon 92
autopsied cases. Cancer. 1979;44:529-542.
3. Berkman N, Breuer R, Mordechai RK, et al. Pulmonary
involvement in lymphoma. Leuk Lymphoma. 1996;20:229-237.
4. Jenkins PF, Ward MJ, Davies P, et al. Non-Hodgkin’s
lymphoma, chronic lymphatic leukaemia and the lung. Br J
Dis Chest. 1981;75:22-30.
5. Lewis ER, Caskey CI, Fishman EK. Lymphoma of the lung:
CT findings in 31 patients. AJR Am J Roentgenol.
1991;156:711-714.
6. Mentzer SJ, Reilly JJ, Skarin AT, et al. Patterns of lung involvement by malignant lymphoma. Surgery. 1993;113:507-514.
7. Diederich S, Link TM, Zuhlsdorf H, et al. Pulmonary
manifestations of Hodgkin’s disease: radiographic and CT
findings. Eur Radiol. 2001;11:2295-2305.
8. Tanaka N, Matsumoto T, Miura G, et al. CT findings of
leukemic pulmonary infiltration with pathologic correlation.
Eur Radiol. 2002;12:166-174.
9. Manoharan A, Pitney WR, Schonell ME, et al. Intrathoracic
manifestation in non-Hodgkin’s lymphoma. Thorax.
1979;34:29-32.
10. Dear AE, Goldstein D, Hayman JA. Malignant pulmonary
lymphoid disease: case reports illustrating anatomical pattern
of disease as a prognostic marker. Pathology. 1996;28:20-24.
11. Jaffe ES, Harris NL, Stein H, et al, eds. Pathology and Genetics
of Tumours of Haematopoietic and Lymphoid Tissues. Lyon,
France: IARC Press; 2001. World Health Organization
Classification of Tumours.
12. Cobby M, Whipp E, Bullimore J, et al. CT appearances of
relapse of lymphoma in the lung. Clin Radiol. 1990;41:232-238.
13. Chernilo S, Trujillo S, Meneses M, et al. Pulmonary
involvement due to disseminated non Hodgkin lymphoma in
one patient with AIDS. Rev Med Chil. 2001;129:1315-1319.
14. Saeki S, Soejima Y, Shimoda T, et al. A case of pulmonary
involvement of malignant lymphoma with diffuse groundglass opacity in chest CT. Nihon Kokyuki Gakkai Zasshi.
2003;41:44-47.
15. Nzeh DA. Atypical intrathoracic manifestation of Burkitt’s
lymphoma. Pediatric Radiol. 1988;18:411-412.
726
726
Am J Clin Pathol 2004;121:718-726
DOI: 10.1309/R4BRRADYF20K49HE
16. Dunnick NR, Reaman GH, Head GL, et al. Radiographic
manifestations of Burkitt’s lymphoma in American patients.
AJR Am J Roentgenol. 1979;132:1-6.
17. Sandlund JT, Fonseca T, Leimig T, et al. Predominance and
characteristics of Burkitt lymphoma among children with
non-Hodgkin lymphoma in northeastern Brazil. Leukemia.
1997;11:743-746.
18. Weis JW, Winter MW, Phyliky RL, et al. Peripheral T-cell
lymphomas: histologic, immunohistologic, and clinical
characterization. Mayo Clin Proc. 1986;61:411-426.
19. Milito CB, Morais JC, Nucci M, et al. Classification of nonHodgkin’s lymphoma: morphological and immunological
study of 145 cases. Jornal Brasileiro de Patologia e Medicina
Laboratorial. 2002;38:315-324.
20. A clinical evaluation of the International Lymphoma Study
Group classification of non-Hodgkin’s lymphoma: the NonHodgkin’s Lymphoma Classification Project. Blood.
1997;89:3909-3918.
21. Carvalho SMF, Pombo de Oliveira MS, Bezerra ACS, et al.
HTLV-I/II in haematological and cancer patients in Rio de
Janeiro, Brazil [abstract]. Blood. 1993;82(suppl):624a.
22. Tilly H, Gaulard P, Lepage E, et al. Primary anaplastic largecell lymphoma in adults: clinical presentation,
immunophenotype, and outcome. Blood. 1997;90:3727-3734.
23. Kikuchi M, Jaffe ES, Ralfkiaer E. Adult T-cell
leukaemia/lymphoma. In: Jaffe ES, Harris NL, Stein H, et al,
eds. Pathology and Genetics of Tumours of Haematopoietic and
Lymphoid Tissues. Lyon, France: IARC Press; 2001:200-203.
World Health Organization Classification of Tumours.
24. Colby TV, Hoppe RT, Warnke RA. Hodgkin’s disease at
autopsy: 1972-1977. Cancer. 1981;47:1852-1862.
25. Elgui de Oliveira D, Bacchi MM, Abreu ES, et al. Hodgkin
disease in adult and juvenile groups from two different
geographic regions in Brazil: characterization of
clinicopathologic aspects and relationship with Epstein-Barr
virus infection. Am J Clin Pathol. 2002;118:25-30.
26. Zinzani PL, Tani M, Gabriele A, et al. Extranodal marginal
zone B-cell lymphoma of MALT-type of the lung; single-center
experience with 12 patients. Leuk Lymphoma. 2003;44:821-824.
27. Fujihara T, Mori F, Kawano K, et al. A case of T-cell type
malignant lymphoma of the lung [in Japanese]. Nippon Kyobu
Geka Gakkai Zasshi. 1993;41:258-261.
28. Wolff AJ, O’Donnell AE. Pulmonary manifestations of HIV
infection in the era of highly active antiretroviral therapy.
Chest. 2001;120:1888-1893.
29. Liebow AA, Carrington CR, Friedman PJ. Lymphomatoid
granulomatosis. Hum Pathol. 1972;3:457-558.
30. Morice WG, Kurtin PJ, Myers JL. Expression of cytolytic
lymphocyte-associated antigens in pulmonary lymphomatoid
granulomatosis. Am J Clin Pathol. 2002;118:391-398.
31. Celikoglu F, Teirstein A, Krellenstein DJ, et al. Pleural effusion
in non-Hodgkin’s lymphoma. Chest. 1992;101:1357-1360.
32. Aquino LS, Chen MYM, Kuo WT, et al. The CT appearance
of pleural and extrapleural disease in lymphoma. Clin Radiol.
1999;54:647-650.
33. Billingham ME, Rawlinson DG, Berry PF, et al. The
cytodiagnosis of malignant lymphomas and Hodgkin’s disease
in cerebrospinal, pleural and ascitic fluids. Acta Cytol.
1975;19:547-556.
34. Bangerter M, Hildebrand A, Griesshammer M. Combined
cytomorphologic and immunophenotypic analysis in the
diagnostic workup of lymphomatous effusions. Acta Cytol.
2001;45:307-312.
© American Society for Clinical Pathology