Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
487
CT in Differential Diagnosis Diffuse Pleural Disease
Ann Nestor Roberta
N. Leung1 L. MUller1 R. Miller2
of
The CT features of benign and malignant pleural diseases have been described. However, the accuracy of these features in the differential diagnosis of diffuse pleural disease has not been assessed before. Without knowledge of clinical or pathologic data, we reviewed the CT findings in 74 consecutive patients with proved diffuse pleural disease (39 malignant and 35 benign). The patients included 53 men and 21 women 2378 years old. Features that were helpful in distinguishing malignant from benign pleural disease were (1) circumferential pleural thickening, (2) nodular pleural thickening, (3) panetal pleural thickening greater than 1 cm, and (4) mediastinal pleural involvement. The specificities of these findings were 100%, 94%, 94%, and 88%, respectively. The sensitivities were 41%, 51%, 36%, and 56%, respectively. Twenty-eight of 39 malignant cases (sensitivity, 72%; specificity, 83%) were identified correctly by the presence of one or more of these criteria. Malignant mesothelioma (n = 1 1) could not be reliably differentiated from pleural metastases (n = 24). We conclude that CT is helpful in the differential diagnosis of diffuse pleural disease, particularly
in differentiation
AJR 154:487-492,
March
of malignant
from
benign
conditions.
1990
A number of benign and malignant diseases may cause diffuse pleural abnormalities. The most common causes are asbestos-related pleural fibrosis, fibrothorax, empyema, mesothelioma, and metastatic disease. The characteristic CT appearances of these and other diffuse pleural diseases have been described [110]. However, no study has analyzed the value of CT in the differential diagnosis of diffuse pleural disease. Because diffuse pleural abnormalities usually consist of various degrees of pleural thickening, calcification, and effusion, an overlap of the CT manifestations of the different benign and malignant pleural diseases would be expected. The aim of our study was to determine the CT features most helpful in the differential diagnosis of diffuse pleural disease.
Materials
Received August 31, 1989; accepted after revision October 16, 1989.
Department of Radiology, University of British Columbia, and Vancouver General Hospital, 855 W. 12th Ave., Vancouver, B.C., Canada V5Z 1M9. Address reprint requests to N. L. MUller. 2 Department of Pathology, University of British Columbia, and Vancouver General Hospital, 855 W. 12th Ave., Vancouver, B.C., Canada V5Z 1 M9. 0361 -803x/9o/1 © American
543-0487
Roentgen
Ray Society
and Methods
All patients with diffuse pleural disease March 1989 who had a definitive diagnosis
included
in this
retrospective
study.
seen in our institution between May 1 985 and and in whom CT scans had been obtained were
Seventy-four
patients
were
selected
by
reviewing
the
medical records and CT reports. Fifty-three were men and 21 were women, with a mean age of 60 years (range, 23-78 years). The hemithorax with the most involvement was the right in 41 cases and the left in 33. Definitive pathologic diagnosis based on the results of pleural biopsy was available in 38 of 39 cases of malignant pleural disease (11 mesothelioma, 22 metastatic pleural carcinoma, three non-Hodgkin lymphoma, one metastatic melanoma and one liposarcoma), five of eight cases of fibrothorax, and seven of 11 infectious processes. The diagnosis of metastatic adenocarcinoma in one patient was based on cytologic findings in a sample obtained by thoracentesis as well as the concomitant findings of multiple brain and lung metastases. The
LEUNG
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
488
ET AL.
AJR:154, March 1990
three patients with fibrothorax who did not have biopsy had known previous history of treated tuberculosis and no change in pleural
location
of the
fissural,
or
thickening
pleural
for
1 year
or more.
The 1 6 patients with asbestos-related pleural disease had documented exposure to asbestos and had been referred either for evaluation of questionable radiographic findings with associated abnormal pulmonary function tests or pleural effusion of unknown cause. All 16 patients had diffuse pleural thickening, defined as blunting of the costophrenic sulcus or localized areas of thickening extending for more than 5 cm in both craniocaudal and transverse diameters. Patients with pleural plaques only were excluded. Six months to 3 years of follow-up in these 1 6 patients has shown no evidence
of malignancy.
the
of positive
basis
performed
in three
tuberculous The
of four
empyemas,
CT
patients)
Infectious cultures
scans
or
(64
tuberculous obtained
patients)
disease
was
with
biopsy
fluid
empyemas,
and both cases
were
9800
pleural
of pleural
on
diagnosed
on
confirmation
two
of five
non-
of actinomycosis. a General
scanner
(GE
Electric
Medical
8800
(10
Systems,
Mil-
parietal
between
pleural
nodular. whether perimeter of
scans
were
thickening
ening
length
of
high-resolution
The
scans
were
presence, type, previous studies adenopathy;
thorax; diastinal
10 mm
assessed
for
the
presence
of
pleural
effusion;
and extent of pleural thickening as suggested by 1 -3]; presence or absence of extrapleural invasion;
calcified
nodes;
loss
of
volume
in the
involved
and coexisting pulmonary parenchymal abnormalities. nodes were considered enlarged if they were greater in long-axis
diameter
in the
transverse
hemi-
Methan
plane.
Pleural thickening was classified as either focal plaques or diffuse thickening. Pleural plaques were defined as areas of pleural thickening less than 5 cm in either transverse or craniocaudal extent. The
TABLE
1: Characteristics
>1 cm
Mediastinal
pleural
involvement
Calcification Invasion Unilateral
involvement
Extension through entire hemithorax Lung base involvement Visceral pleural involvement Fissural involvement Effusion Plaques Hilar or mediastinal adenopathy Calcified nodes Loss of volume Note-One category.
as
involve-
thickening
The contour
smooth,
of
irregular,
or
in the
lower
extension,
third
of the
and
(defined
hemithorax).
presence
as pleural
Pleural
or
thick-
calcification
was
All
pathologic
specimens
assessment
basis of standard theliomas
one
one desmoplastic
and cytologic
low-grade
In cases
could
in
comparison
formed
a pathologist. determined
the
mesothelioma,
usual
differential
The on the
The 1 1 meso-
epithelial
and nine more which
epithelial diagnosis
and was
and epithelial mesothelioma, the diagnosis made if neutral mucin and/or carcinoem-
be shown
[11
The clinical data were obtained Statistical
by was
features.
papillary
mesothelioma,
patterns.
reviewed
vs malignant
architectural
included
antigen
were
of benign
between
by using the chi-square
12]. from hospital ,
selected
patient
and office charts. groups
was
per-
test.
Results Seventy-one patients had pleural thickening on CT. In three cases, unilateral pleural effusion was the sole CT manifestation of diffuse pleural malignancy (Table 1). In patients with pleural thickening, the features most suggestive of a malignant cause were pleural rind (a finding seen only in patients with malignant pleural disease), with a specificity of 1 00% and a sensitivity of 41%; nodular pleural thickening, with a specificity of 94% and a sensitivity of 51%; parietal pleural thickening greater than 1 cm, with a specificity
of Pleural Disease
Ch aracteristic Nodularity Rind Thickness
as
was
classified as involving plaques or diffuse thickening. In the cases for which high-resolution CT was available, these additional scans were judged to be of benefit when they enabled a classification change in any of the previously noted characteristics.
bryonic
algorithm,
pleural
pleural
of pleural effusion.
of the lung bases
a high-spatial-resolution
(bone
mediastinal
characterized
visceral, defined
and no attempt
and parietal
craniocaudal
of involvement
between adenocarcinoma of adenocarcinoma was
algorithm
visceral
visceral
was
obtained at selected levels specific to each case. These thin-section images were reconstructed by using a field of view of 20-25 cm and CT). Except for the asbestos-exposed group, contrast material was administered routinely. The CT scans were done as part of the initial assessment of these patients at our institution. The CT scans and accompanying chest radiographs were reviewed by two observers who were unaware of the pathologic diagnosis; a conclusion was reached by consensus.
was
Parietal pleural involvement was further characterized by it was circumferential (defined as involvement of entire of hemithorax including mediastinum, or pleural rind), width
thickening,
absence
as parietal,
involvement
from
was made only in the presence the
classified
the mediastinum,
The distinction
mixed
collimation
bordering
to differentiate
mediastinum.
1 .5-mm
thickening
was
Mediastinal
ment.
histologic
additional
thickening
made
waukee, WI). Contiguous 1 -cm collimation scans were obtained from lung apices to the level of the adrenals. All scans were obtained at end-inspiratory lung volumes by using the standard algorithm and were photographed at windows appropriate for lung parenchyma and In 31 cases,
pleural
mediastinal.
case of pleural liposarcoma
Mesothelioma
Metastases
Lymphoma
(1 1 )
(24)
(3)
7 8 6
12 8 5
1 0 2
Asbestos
Empyema
Fibrothorax
Actinomycosis
(1 6)
(9)
(8)
(2)
0 0 1
0 0 0
2 0 1
8
13
1
1
2
1
0
3
1 9
2 14
0 0
11 0
1 0
4 0
0 0 0 0 0 0
2
0
8
4
2
6 8 7 4 11 3 5 0 7
8 21 11 3 21 3 10 2 13
0 1 0 0 2 0 2 0 1
2 15 6 5 6 10 4 1 7
3 9 9 2 9 0 2 1 7
2 8 5 2 5 1 4 3 3
0 2 0 0 0 0 0 0 0
was not included in this table but was included
in the series.
Numbers
in parentheses
are number
of patients
in each
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
AJR:154,
March
1990
CT
OF
DIFFUSE
of 94% and a sensitivity of 36%; and mediastinal pleural involvement, with a specificity of 88% and a sensitivity of 56% (Table 1 Figs. 1 and 2). All these features were significantly more common in malignant than in benign pleural thickening (p < .01, chi-square test). Twenty-eight of 39 cases of malignant pleural disease had one or more of these features as compared with only six of 35 cases of benign pleural disease, representing a sensitivity of 72% and a specificity of 83% for malignancy (Table 1, Fig. 3). The presence of pleural calcification was suggestive of benign cause, with a sensitivity of 46% and a specificity of 92% (Table 1). In most patients with mesothelioma, the CT findings were identical to those of metastatic pleural disease, including the presence of nodular pleural thickening, pleural rind, mediastinal or chest wall invasion, and loss of volume (Table 1, Fig. 4). Pleural effusion as the only manifestation of pleural malig,
PLEURAL
489
DISEASE
nancy was a feature seen in pleural metastases and not in mesothelioma. Pleural lymphoma was indistinguishable from nant
pleural
involvement.
or mediastinal)
was
In this group,
a more
consistent
(two
of 24)
other
malig-
adenopathy
finding
(hilar
and/
(two
of three); circumferential pleural rind was not seen. The only case of pleural liposarcoma had a virtually pathognomonic appearance: an inhomogeneous pleurally based mass of tissue density intermixed with several areas of fat. Parietal pleural thickening was greater than 1 cm and irregular in contour; invasion into both mediastinum and chest wall was present. Features distinguishing mesothelioma from benign asbestos-related pleural thickening were essentially the same as those differentiating neoplastic from benign disease (Table 1). The two most useful features in differentiating asbestosrelated pleural disease from other benign conditions were the presence of pleural plaques, with a specificity of 95% (p < .01) and bilateral pleural involvement, with a specificity of 74% (p < .01) for asbestos-related pleural disease (Fig. 5). All nine cases of empyema were seen on CT as a round or lenticular fluid collection separating slightly thickened visceral and parietal pleural surfaces (Fig. 6). The thickened pleura was usually smooth although sometimes irregular; neither nodularity nor thickening greater than 1 cm was seen (Table 1). High-resolution CT was helpful in 19 of 31 cases. In 10 cases, better definition of asbestos-related parenchymal changes was obtained; in the remainder, the higher resolution scans confirmed equivocal mediastinal or fissural involvement and nodular pleural thickening (Fig. 2).
Discussion Fig. 1.-i-cm collimation CT scan at level of left atrium in a patient with metastatic melanoma shows right-sided nodular pleural thickening >1 cm. Nodular thickening also involves mediastinal pleura (arrow) and major fissure.
Pleural to three thickening,
response to a variety of diseases is limited largely radiologically detectable manifestations: effusion, and calcification. .‘ I;
Fig. 2.-CT scans at level of tracheal carina in a 59-year-old man with desmoplastic mesothelioma. A, i-cm collimation scan at time of presentation shows circumferential right-sided pleural thickening and subtle mediastinal involvement (arrow). B, 1.5-mm collimation scan at time of presentation shows improved definition of the circumferential pleural changes and nodular pleural thickening (arrow). C, 1-cm collimation scan 4 months after presentation shows an increase in width of pleural thickening.
LEUNG
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
490
Fig. 3-1.5-mm
collimation
tient with fibrothorax
CT scan
caused
in a pa-
by Mycobacterium
avium-intracellulare infection shows nodular pleural thickening with involvement of right major fissure and mediastinal pleura. Anterior surface of parietal pleura was normal. Nodular pleural thickening and mediastinal pleural thickening are usually indicative of malignant pleural disease; they are seen in 14% or less of patients with benign pleural thickening.
ET AL.
AJR:154, March 1990
Fig. 4.-Pleural rind. A, 1-cm collimation CT scan in a 75-year-old man with mesothelioma shows circumferential pleural thickening with mediastinal involvement (pleural rind). Pleural thickening is >1 cm in maximal width (arrow). B, 1-cm collimation CT scan in a 45-year-old woman with metastatic adenocarcinoma shows a pleural rind with focal areas of nodularity (arrows). CT appearance of metastatic pleural disease cannot be distinguished from malignant pleural mesothelioma.
Fig. 5.-Asbestos-related pleural disease. A, i-cm collimation scan in a 62-year-old man with history of asbestos exposure shows bilateral pleural thickening with a calcified plaque (arrow) and a small left pleural effusion. Panetal pleural thickening extends into anterior third of left hemithorax but does not involve mediastinal pleura. B, 1.5-mm collimation CT scan in a 61-yearold man with history of asbestos exposure shows bilateral pleural thickening with calcified parietal and mediastinal (arrow) plaques.
Pleural disease.
effusion is a common manifestation of diffuse pleural The major objective in CT assessment of unexplained pleural effusion is establishment of benign vs malignant cause. As the main mechanism of pleural fluid formation in malignancy is lymphatic obstruction, the site of blockage may be present at any point between the stomata of parietal pleura and the mediastinal lymph nodes [13]. The resulting effusion is usually an exudate. With the possible exception of hemothorax, measurement of fluid attenuation coefficients has proved unreliable in distinguishing fluid composition [4, 14]. Maffessanti et al. [8] suggested that the absence of pleural
thickening
In that
series,
pleura three
had
cases,
neoplastic
does
seven
malignant pleural pleural
not
preclude
a neoplastic
diagnosis.
of 12 patients
with normal-appearing Our results are similar;
effusions.
effusion
involvement.
was
the sole
During
manifestation
thoracotomy
in
of in one of
these patients, nodules were seen over the entire surface of the visceral pleura, although pleural thickening could not be seen on CT even in retrospect. Thus, the absence of pleural thickening on CT does not rule out pleural malignancy. In the presence of pleural thickening, the most useful features in the differentiation of malignant from benign pleural disease are the presence of pleural rind, pleural nodularity, pleural thickening greater than 1 cm, and mediastinal pleural involvement. These features may be seen in mesothelioma and in metastatic pleural disease but are unusual in benign pleural disease. These CT features correlate closely with the pathologic findings. Malignant pleural disease tends to involve the entire pleural surface, whereas reactive pleurisy usually does not affect the mediastinal pleura [15, 16]. The CT differentiation of benign and malignant pleural disease is important because the specific diagnosis is often
CT
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
AJR:154, March 1990
Fig. 6.-Empyema. A, i-cm collimation
CT scan
a nontuberculous empyema
in a patient
OF
DIFFUSE
PLEURAL
DISEASE
491
with
shows a lenticular
fluid collection separating smoothly thickened layers of parietal and visceral pleura. Although pleural thickening is extensive, this does not involve mediastinal pleura (i.e., it is not circumferential). Areas of atelectasis are present in right lower lobe. B, i-cm collimation CT scan in a patient with a tuberculous empyema shows anterior mediastinal pleural involvement (arrow).
difficult to make by clinical criteria, pleurocentesis, and percutaneous pleural biopsy. Not rarely, a diagnostic thoracotomy is required because of the need to assess the growth pattern of disease and to obtain a relatively large amount of tissue for histologic, immunocytochemical, and ultrastructural studies [1 1 ]. Even careful examination and biopsy of the pleura at thoracotomy can fail to establish the diagnosis [17, 1 8]. Ryan et al. [1 7] reported 51 patients in whom no cause for a pleural effusion was found at thoracotomy. In 1 3 of these patients (25%), the diagnosis of malignancy (including lymphoma, carcinoma, and malignant mesothelioma) was established 1 2 days to 5 years after thoracotomy. The difficulty in pathologically differentiating benign reactive mesothelial cells from mesothelioma is a common and well-known probIem, and in many instances, the gross appearance assumes great importance in making this distinction [1 1 ]. Insofar as the CT appearance reflects the gross distribution of disease, we believe that the CT findings are helpful in determining whether thoracotomy is recommended and in guiding the surgeon to the sites most likely to yield a positive diagnosis. These sites would include areas with greater than 1 cm pleural thickening, nodular pleural thickening, and thickening of the mediastinal pleura. The presence of pleural calcification suggests a benign process. In this series, calcification was seen in 1 6 of 35 patients with benign pleural thickening and in only three of 39 patients with malignant pleural disease. Although calcified plaques may be seen in cases of mesothelioma, they are uncommon [3] and were not seen in any of our 1 1 mesothelioma patients. The relative absence of pleural calcification in mesothelioma may be due to absorption of calcification by the developing tumor [1 0] and the fact that a significant number of malignant mesotheliomas (range, 0-87%) are not asbestos-related [19]. Difficulty in distinguishing between mesothelioma and metastatic pleural involvement pathologically [11, 20] and radiologically [3, 5-7] has been recognized. Pathologic diagnosis requires adequate sampling; cytologic examination of pleural fluid and needle biopsy specimens rarely allows a definitive diagnosis of mesothelioma and may provide the misleading interpretation of metastatic carcinoma or benign pleural disease [9, 21, 22]. Discrimination between epithelial types of
mesothelioma and metastatic adenocarcinoma requires histochemical, immunohistochemical, and ultrastructural analysis [11, 12]. Invasion of the chest wall or mediastinum on CT are specific markers for malignancy but have a low sensitivity; they were present in only 8% of patients with malignant pleural involvement in our series. In addition, this sign was not useful in distinguishing mesothelioma from carcinoma. The limitations of CT in assessing the presence or absence of invasion into the chest wall or mediastinum in mesothelioma also have been demonstrated recently by Rusch et al. [23]. Adams et al. [24] have suggested that the presence of hilar adenopathy may be helpful in differentiating metastases from mesothelioma. However, true hilar involvement with no mediastinal involvement is rare in metastatic pleural disease except for bronchogenic carcinoma, lymphoma, and renal cell carcinoma [25]. In our study, hilar adenopathy was identified in only two patients with malignancy. One case was a nonHodgkin lymphoma and the second case was an adenocarcinoma. The tendency of mesothelioma to involve the inferior hemithorax has been postulated to be a specific sign [1 0]. However, in our study basal involvement was present not only in patients with mesothelioma but also in the majority of patients with metastatic pleural disease. Benign asbestos pleural effusion has been shown by Epler et al. [26] to occur with an approximate prevalence of 3% in the asbestos-exposed population. Effusion is usually unilateral but may be bilateral; other etiologic manifestations of asbestos-related disease may be present or absent. Differentiation between benign asbestos pleural effusion and asbestos-related malignant mesothelioma usually can be made on the basis of the much shorter latency period of benign effusion, which ranges from 1 0 to 20 years from time of initial exposure [26]. However, close follow-up is necessary as the effusion may be the first manifestation of mesothelioma [3], and routine radiologic studies have been advocated for mesothelioma screening in this high-risk population [22]. Nodularity of pleura has been recognized as a helpful discriminating feature [2, 3, 10]. In our study, nodular pleural thickening was seen in seven (63%) of 11 cases of mesothelioma and in none of the cases of benign asbestos-related
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
492
LEUNG
diffuse pleural disease. However, extensive nodular thickening of pleural plaques mimicking mesothelioma has been described [3]. In such cases, open pleural biopsy is the only reliable way to establish the definitive diagnosis of benign disease. Although loss of volume is a common manifestation in mesothelioma [3], it is also a relatively nonspecific finding seen in a variety of other malignant and benign conditions [6]. In our series, volume loss was seen in approximately 50% of patients, irrespective of the cause of the diffuse pleural disease. Benign pleural asbestos-related disease generally is acknowledged to be a symmetrical process [27, 28], and in all 1 6 of our cases, pleural involvement was bilateral. Typically, the changes affect the posterolateral aspect of the pleura [29]. In one of our cases of early malignant mesothelioma, the only sign that suggested possible tumor was the unilaterality of pleural involvement. Although this is not a useful feature in differentiation from other benign diseases, in the select asbestos-exposed subgroup, unilaterality or marked asymmetry of pleural thickening must be considered suggestive of mesothelioma. Asbestos-related pleural disease was the only benign disease that could be specifically suggested from the CT scan by virtue of the presence of bilateral pleural plaques. Although smooth pleural thickening with associated effusion was suggestive of empyema, such features also were observed in a number of noninfectious processes. No characteristic features of fibrothorax were identified. Aberle et al. [29, 30] found that high-resolution CT was more sensitive than conventional CT in detection of asbestosrelated pleural and parenchymal fibrosis. Our study similarly shows that the scans with higher spatial resolution improved detection of subtle asbestos-related parenchymal changes and allowed confirmation of pleural changes deemed equivocal on conventional CT. Our study was limited by the relatively small number of cases of some diseases and by the relatively small number of diseases included. However, as all consecutive cases of confirmed diffuse pleural disease studied during a 4-year period were included, this study is a reflection of the incidence and prevalence of the types of diffuse pleural disease posing diagnostic problems in our institution. We conclude that the CT features most helpful in distinguishing malignant from benign pleural disease are pleural rind, nodular pleural thickening, pleural thickening greater than 1 cm, and mediastinal pleural involvement. Pleural calcification usually implies a benign process. Loss of volume is not a helpful feature. In the majority of patients, mesothelioma cannot be distinguished from metastatic pleural disease on CT. High-resolution CT is superior to conventional CT, particularly in the assessment of asbestos-related pleural disease.
ET AL.
AJR:154,
mesothelioma and asbestosis using CT and conventional chest radiography. Radiology 1982:144:453-460 4. Williford ME, Hidalgo H, Putman CE, Korobkin M, Ram PC. Computed tomography of pleural disease. AJR 1983:140:909-914 5. Grant DC, Seltzer SE, Antman KH, Finberg HJ, Koster K. Computed tomography of malignant pleural mesothelioma. J Comput Assist Tomogr
i983;7:626-632 6. Salonen
0, Kivisaari
K, Tammilehto
7. 8. 9. 1 0. 11.
1 2.
13. 14. 15. 1 6.
L, Standertskjold-Nordenstam
L. Computed
tomography
K, Mattson special (Stockh) 1986;
reference to the mediastinal pleura. Acta Radiol (Diagn] 27:527-531 Alexander E, Clark RA, Colley DP, Mitchell SE. CT of malignant pleural mesothelioma. AJR i98i;1 37:287-291 Maftessanti M, Tommasi M, Pellegrini P. Computed tomography of free pleural effusions. Eur J Radio! 1987:7:87-90 Mirvis 5, Dutcher JP, Haney PJ, Whitley NO, Aisner J. CT of malignant pleural mesothelioma. AJR 1983; 1 40: 665-670 Kreel L. Computed tomography in mesothelioma. Semin Oncol 1981; 8:302-312 Roggli VL, Kolbeck J, Sanfillippo F, Shelburne JD. Pathology of human mesothelioma: etiologic and diagnostic considerations. Pathol Annu 1987;2:91-131 Ordonez NG. The immunohistochemical diagnosis of mesothelioma: differentiation of mesothelioma and lung adenocarcinoma. Am J Surg Patho! 1989:13:276-291 Sahn SA. The pleura. Am Rev Respir Dis 1988:138:184-234 Kollins SA. Computed tomography of the pulmonary parenchyma and chest wall. Radio! Clin North Am i977;1 5: 297-308 Herbert A. Pathogenesis of pleurisy. pleural fibrosis, and mesothelial proliferation. Thorax 1986:41 :176-189 Barrett NR. The pleura, with special reference to fibrothorax. Thorax
1970:25:515-524 Ryan CJ, Rodgers RF, Unni KK, Hepper NGG. The outcome of patients with pleural effusion of indeterminate cause at thoracotomy. Mayo Clin Proc 1981:56:145-149 18. Douglass BE, Carr DT, Bernatz PE. Diagnostic thoracotomy in the study of “idiopathic” pleural effusion. Am Rev Tuberc 1956:74:954-957 19. Peterson JT, Greenberg SD, Buffler PA. Non-asbestos related malignant mesothelioma: a review. Cancer i984;54:951-960 20. Szpak CA, Johnston WW, Roggli V, et al. The diagnostic distinction between malignant mesothelioma of the pleura and adenocarcinoma of the lung as defined by a monoclonal antibody (B72.3). Am J Pathol 1986:122:252-260 21 . Shearin JC, Jackson D. Malignant pleural mesothelioma. J Thorac Cardiovasc Surg i976;71 :621-627
22. Strankinga WFM, Sperber M, Kaiser MC, Stam J. Accuracy of diagnostic procedures Respiration
in the initial evaluation and follow-up (Base!) 1987;51 :179-187
of mesothelioma
23. Rusch VW, Godwin JD, Shuman WP. The role of computed
patients.
tomography
scanning
24. 25.
in the initial assessment and the follow-up of malignant pleural mesothelioma. J Thorac Cardiovasc Surg 1988:96; 171-177 Adams VI, Unni KK, Muhm JR. Jeff JR. llstrup DM, Bernatz PE. Diffuse malignant mesothelioma of pleura. Cancer 1986:58: 1540-1 551 Reed JC. Chest radiology: plain film patterns and differential diagnosis. Chicago: Year Book Medical Publishers, Inc. , 1981
26. Epler GA, McLoud TC, Gaensler EA. Prevalence and incidence of benign asbestos pleural effusion in a working population. JAMA 1982;247: 61 7-622 27. Sargent EN, Gordonson J, Jacobson G, Birnbaum W, Shaub M. Bilateral pleural
thickening:
a manifestation
of
asbestos
dust
exposure.
AJR
1978:131:579-585
REFERENCES 30. 1978:13:213-225
C-G, Somer
of pleural lesions with
17.
28.
of the lung and pleura. Semin Roentgeno!
1990
2. Katz D, Kreel L. Computed tomography in pulmonary asbestosis. Clin Radiol 1979:30:207-213 3. Rabinowitz JG, Efremidis SC, Cohen B, et al. A comparative study of
Preger
L. Asbestos-related
disease.
New York:
29. Aberle DR. Gamsu G, Ray CS, Feuerstein
1. Kreel L. Computed tomography
March
Grune
and Stratton,
IM. Asbestos-related
1978
pleural
and parenchymal fibrosis: detection with high-resolution CT. Radiology 1988:166:729-734 Aberle DR, Gamsu G, Ray CS. High-resolution CT of benign asbestosrelated diseases: clinical and radiologic correlation. AJR 1988:151: 883-891
487
CT in Differential Diagnosis Diffuse Pleural Disease
Ann Nestor Roberta
N. Leung1 L. MUller1 R. Miller2
of
The CT features of benign and malignant pleural diseases have been described. However, the accuracy of these features in the differential diagnosis of diffuse pleural disease has not been assessed before. Without knowledge of clinical or pathologic data, we reviewed the CT findings in 74 consecutive patients with proved diffuse pleural disease (39 malignant and 35 benign). The patients included 53 men and 21 women 2378 years old. Features that were helpful in distinguishing malignant from benign pleural disease were (1) circumferential pleural thickening, (2) nodular pleural thickening, (3) panetal pleural thickening greater than 1 cm, and (4) mediastinal pleural involvement. The specificities of these findings were 100%, 94%, 94%, and 88%, respectively. The sensitivities were 41%, 51%, 36%, and 56%, respectively. Twenty-eight of 39 malignant cases (sensitivity, 72%; specificity, 83%) were identified correctly by the presence of one or more of these criteria. Malignant mesothelioma (n = 1 1) could not be reliably differentiated from pleural metastases (n = 24). We conclude that CT is helpful in the differential diagnosis of diffuse pleural disease, particularly
in differentiation
AJR 154:487-492,
March
of malignant
from
benign
conditions.
1990
A number of benign and malignant diseases may cause diffuse pleural abnormalities. The most common causes are asbestos-related pleural fibrosis, fibrothorax, empyema, mesothelioma, and metastatic disease. The characteristic CT appearances of these and other diffuse pleural diseases have been described [110]. However, no study has analyzed the value of CT in the differential diagnosis of diffuse pleural disease. Because diffuse pleural abnormalities usually consist of various degrees of pleural thickening, calcification, and effusion, an overlap of the CT manifestations of the different benign and malignant pleural diseases would be expected. The aim of our study was to determine the CT features most helpful in the differential diagnosis of diffuse pleural disease.
Materials
Received August 31, 1989; accepted after revision October 16, 1989.
Department of Radiology, University of British Columbia, and Vancouver General Hospital, 855 W. 12th Ave., Vancouver, B.C., Canada V5Z 1M9. Address reprint requests to N. L. MUller. 2 Department of Pathology, University of British Columbia, and Vancouver General Hospital, 855 W. 12th Ave., Vancouver, B.C., Canada V5Z 1 M9. 0361 -803x/9o/1 © American
543-0487
Roentgen
Ray Society
and Methods
All patients with diffuse pleural disease March 1989 who had a definitive diagnosis
included
in this
retrospective
study.
seen in our institution between May 1 985 and and in whom CT scans had been obtained were
Seventy-four
patients
were
selected
by
reviewing
the
medical records and CT reports. Fifty-three were men and 21 were women, with a mean age of 60 years (range, 23-78 years). The hemithorax with the most involvement was the right in 41 cases and the left in 33. Definitive pathologic diagnosis based on the results of pleural biopsy was available in 38 of 39 cases of malignant pleural disease (11 mesothelioma, 22 metastatic pleural carcinoma, three non-Hodgkin lymphoma, one metastatic melanoma and one liposarcoma), five of eight cases of fibrothorax, and seven of 11 infectious processes. The diagnosis of metastatic adenocarcinoma in one patient was based on cytologic findings in a sample obtained by thoracentesis as well as the concomitant findings of multiple brain and lung metastases. The
LEUNG
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
488
ET AL.
AJR:154, March 1990
three patients with fibrothorax who did not have biopsy had known previous history of treated tuberculosis and no change in pleural
location
of the
fissural,
or
thickening
pleural
for
1 year
or more.
The 1 6 patients with asbestos-related pleural disease had documented exposure to asbestos and had been referred either for evaluation of questionable radiographic findings with associated abnormal pulmonary function tests or pleural effusion of unknown cause. All 16 patients had diffuse pleural thickening, defined as blunting of the costophrenic sulcus or localized areas of thickening extending for more than 5 cm in both craniocaudal and transverse diameters. Patients with pleural plaques only were excluded. Six months to 3 years of follow-up in these 1 6 patients has shown no evidence
of malignancy.
the
of positive
basis
performed
in three
tuberculous The
of four
empyemas,
CT
patients)
Infectious cultures
scans
or
(64
tuberculous obtained
patients)
disease
was
with
biopsy
fluid
empyemas,
and both cases
were
9800
pleural
of pleural
on
diagnosed
on
confirmation
two
of five
non-
of actinomycosis. a General
scanner
(GE
Electric
Medical
8800
(10
Systems,
Mil-
parietal
between
pleural
nodular. whether perimeter of
scans
were
thickening
ening
length
of
high-resolution
The
scans
were
presence, type, previous studies adenopathy;
thorax; diastinal
10 mm
assessed
for
the
presence
of
pleural
effusion;
and extent of pleural thickening as suggested by 1 -3]; presence or absence of extrapleural invasion;
calcified
nodes;
loss
of
volume
in the
involved
and coexisting pulmonary parenchymal abnormalities. nodes were considered enlarged if they were greater in long-axis
diameter
in the
transverse
hemi-
Methan
plane.
Pleural thickening was classified as either focal plaques or diffuse thickening. Pleural plaques were defined as areas of pleural thickening less than 5 cm in either transverse or craniocaudal extent. The
TABLE
1: Characteristics
>1 cm
Mediastinal
pleural
involvement
Calcification Invasion Unilateral
involvement
Extension through entire hemithorax Lung base involvement Visceral pleural involvement Fissural involvement Effusion Plaques Hilar or mediastinal adenopathy Calcified nodes Loss of volume Note-One category.
as
involve-
thickening
The contour
smooth,
of
irregular,
or
in the
lower
extension,
third
of the
and
(defined
hemithorax).
presence
as pleural
Pleural
or
thick-
calcification
was
All
pathologic
specimens
assessment
basis of standard theliomas
one
one desmoplastic
and cytologic
low-grade
In cases
could
in
comparison
formed
a pathologist. determined
the
mesothelioma,
usual
differential
The on the
The 1 1 meso-
epithelial
and nine more which
epithelial diagnosis
and was
and epithelial mesothelioma, the diagnosis made if neutral mucin and/or carcinoem-
be shown
[11
The clinical data were obtained Statistical
by was
features.
papillary
mesothelioma,
patterns.
reviewed
vs malignant
architectural
included
antigen
were
of benign
between
by using the chi-square
12]. from hospital ,
selected
patient
and office charts. groups
was
per-
test.
Results Seventy-one patients had pleural thickening on CT. In three cases, unilateral pleural effusion was the sole CT manifestation of diffuse pleural malignancy (Table 1). In patients with pleural thickening, the features most suggestive of a malignant cause were pleural rind (a finding seen only in patients with malignant pleural disease), with a specificity of 1 00% and a sensitivity of 41%; nodular pleural thickening, with a specificity of 94% and a sensitivity of 51%; parietal pleural thickening greater than 1 cm, with a specificity
of Pleural Disease
Ch aracteristic Nodularity Rind Thickness
as
was
classified as involving plaques or diffuse thickening. In the cases for which high-resolution CT was available, these additional scans were judged to be of benefit when they enabled a classification change in any of the previously noted characteristics.
bryonic
algorithm,
pleural
pleural
of pleural effusion.
of the lung bases
a high-spatial-resolution
(bone
mediastinal
characterized
visceral, defined
and no attempt
and parietal
craniocaudal
of involvement
between adenocarcinoma of adenocarcinoma was
algorithm
visceral
visceral
was
obtained at selected levels specific to each case. These thin-section images were reconstructed by using a field of view of 20-25 cm and CT). Except for the asbestos-exposed group, contrast material was administered routinely. The CT scans were done as part of the initial assessment of these patients at our institution. The CT scans and accompanying chest radiographs were reviewed by two observers who were unaware of the pathologic diagnosis; a conclusion was reached by consensus.
was
Parietal pleural involvement was further characterized by it was circumferential (defined as involvement of entire of hemithorax including mediastinum, or pleural rind), width
thickening,
absence
as parietal,
involvement
from
was made only in the presence the
classified
the mediastinum,
The distinction
mixed
collimation
bordering
to differentiate
mediastinum.
1 .5-mm
thickening
was
Mediastinal
ment.
histologic
additional
thickening
made
waukee, WI). Contiguous 1 -cm collimation scans were obtained from lung apices to the level of the adrenals. All scans were obtained at end-inspiratory lung volumes by using the standard algorithm and were photographed at windows appropriate for lung parenchyma and In 31 cases,
pleural
mediastinal.
case of pleural liposarcoma
Mesothelioma
Metastases
Lymphoma
(1 1 )
(24)
(3)
7 8 6
12 8 5
1 0 2
Asbestos
Empyema
Fibrothorax
Actinomycosis
(1 6)
(9)
(8)
(2)
0 0 1
0 0 0
2 0 1
8
13
1
1
2
1
0
3
1 9
2 14
0 0
11 0
1 0
4 0
0 0 0 0 0 0
2
0
8
4
2
6 8 7 4 11 3 5 0 7
8 21 11 3 21 3 10 2 13
0 1 0 0 2 0 2 0 1
2 15 6 5 6 10 4 1 7
3 9 9 2 9 0 2 1 7
2 8 5 2 5 1 4 3 3
0 2 0 0 0 0 0 0 0
was not included in this table but was included
in the series.
Numbers
in parentheses
are number
of patients
in each
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
AJR:154,
March
1990
CT
OF
DIFFUSE
of 94% and a sensitivity of 36%; and mediastinal pleural involvement, with a specificity of 88% and a sensitivity of 56% (Table 1 Figs. 1 and 2). All these features were significantly more common in malignant than in benign pleural thickening (p < .01, chi-square test). Twenty-eight of 39 cases of malignant pleural disease had one or more of these features as compared with only six of 35 cases of benign pleural disease, representing a sensitivity of 72% and a specificity of 83% for malignancy (Table 1, Fig. 3). The presence of pleural calcification was suggestive of benign cause, with a sensitivity of 46% and a specificity of 92% (Table 1). In most patients with mesothelioma, the CT findings were identical to those of metastatic pleural disease, including the presence of nodular pleural thickening, pleural rind, mediastinal or chest wall invasion, and loss of volume (Table 1, Fig. 4). Pleural effusion as the only manifestation of pleural malig,
PLEURAL
489
DISEASE
nancy was a feature seen in pleural metastases and not in mesothelioma. Pleural lymphoma was indistinguishable from nant
pleural
involvement.
or mediastinal)
was
In this group,
a more
consistent
(two
of 24)
other
malig-
adenopathy
finding
(hilar
and/
(two
of three); circumferential pleural rind was not seen. The only case of pleural liposarcoma had a virtually pathognomonic appearance: an inhomogeneous pleurally based mass of tissue density intermixed with several areas of fat. Parietal pleural thickening was greater than 1 cm and irregular in contour; invasion into both mediastinum and chest wall was present. Features distinguishing mesothelioma from benign asbestos-related pleural thickening were essentially the same as those differentiating neoplastic from benign disease (Table 1). The two most useful features in differentiating asbestosrelated pleural disease from other benign conditions were the presence of pleural plaques, with a specificity of 95% (p < .01) and bilateral pleural involvement, with a specificity of 74% (p < .01) for asbestos-related pleural disease (Fig. 5). All nine cases of empyema were seen on CT as a round or lenticular fluid collection separating slightly thickened visceral and parietal pleural surfaces (Fig. 6). The thickened pleura was usually smooth although sometimes irregular; neither nodularity nor thickening greater than 1 cm was seen (Table 1). High-resolution CT was helpful in 19 of 31 cases. In 10 cases, better definition of asbestos-related parenchymal changes was obtained; in the remainder, the higher resolution scans confirmed equivocal mediastinal or fissural involvement and nodular pleural thickening (Fig. 2).
Discussion Fig. 1.-i-cm collimation CT scan at level of left atrium in a patient with metastatic melanoma shows right-sided nodular pleural thickening >1 cm. Nodular thickening also involves mediastinal pleura (arrow) and major fissure.
Pleural to three thickening,
response to a variety of diseases is limited largely radiologically detectable manifestations: effusion, and calcification. .‘ I;
Fig. 2.-CT scans at level of tracheal carina in a 59-year-old man with desmoplastic mesothelioma. A, i-cm collimation scan at time of presentation shows circumferential right-sided pleural thickening and subtle mediastinal involvement (arrow). B, 1.5-mm collimation scan at time of presentation shows improved definition of the circumferential pleural changes and nodular pleural thickening (arrow). C, 1-cm collimation scan 4 months after presentation shows an increase in width of pleural thickening.
LEUNG
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
490
Fig. 3-1.5-mm
collimation
tient with fibrothorax
CT scan
caused
in a pa-
by Mycobacterium
avium-intracellulare infection shows nodular pleural thickening with involvement of right major fissure and mediastinal pleura. Anterior surface of parietal pleura was normal. Nodular pleural thickening and mediastinal pleural thickening are usually indicative of malignant pleural disease; they are seen in 14% or less of patients with benign pleural thickening.
ET AL.
AJR:154, March 1990
Fig. 4.-Pleural rind. A, 1-cm collimation CT scan in a 75-year-old man with mesothelioma shows circumferential pleural thickening with mediastinal involvement (pleural rind). Pleural thickening is >1 cm in maximal width (arrow). B, 1-cm collimation CT scan in a 45-year-old woman with metastatic adenocarcinoma shows a pleural rind with focal areas of nodularity (arrows). CT appearance of metastatic pleural disease cannot be distinguished from malignant pleural mesothelioma.
Fig. 5.-Asbestos-related pleural disease. A, i-cm collimation scan in a 62-year-old man with history of asbestos exposure shows bilateral pleural thickening with a calcified plaque (arrow) and a small left pleural effusion. Panetal pleural thickening extends into anterior third of left hemithorax but does not involve mediastinal pleura. B, 1.5-mm collimation CT scan in a 61-yearold man with history of asbestos exposure shows bilateral pleural thickening with calcified parietal and mediastinal (arrow) plaques.
Pleural disease.
effusion is a common manifestation of diffuse pleural The major objective in CT assessment of unexplained pleural effusion is establishment of benign vs malignant cause. As the main mechanism of pleural fluid formation in malignancy is lymphatic obstruction, the site of blockage may be present at any point between the stomata of parietal pleura and the mediastinal lymph nodes [13]. The resulting effusion is usually an exudate. With the possible exception of hemothorax, measurement of fluid attenuation coefficients has proved unreliable in distinguishing fluid composition [4, 14]. Maffessanti et al. [8] suggested that the absence of pleural
thickening
In that
series,
pleura three
had
cases,
neoplastic
does
seven
malignant pleural pleural
not
preclude
a neoplastic
diagnosis.
of 12 patients
with normal-appearing Our results are similar;
effusions.
effusion
involvement.
was
the sole
During
manifestation
thoracotomy
in
of in one of
these patients, nodules were seen over the entire surface of the visceral pleura, although pleural thickening could not be seen on CT even in retrospect. Thus, the absence of pleural thickening on CT does not rule out pleural malignancy. In the presence of pleural thickening, the most useful features in the differentiation of malignant from benign pleural disease are the presence of pleural rind, pleural nodularity, pleural thickening greater than 1 cm, and mediastinal pleural involvement. These features may be seen in mesothelioma and in metastatic pleural disease but are unusual in benign pleural disease. These CT features correlate closely with the pathologic findings. Malignant pleural disease tends to involve the entire pleural surface, whereas reactive pleurisy usually does not affect the mediastinal pleura [15, 16]. The CT differentiation of benign and malignant pleural disease is important because the specific diagnosis is often
CT
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
AJR:154, March 1990
Fig. 6.-Empyema. A, i-cm collimation
CT scan
a nontuberculous empyema
in a patient
OF
DIFFUSE
PLEURAL
DISEASE
491
with
shows a lenticular
fluid collection separating smoothly thickened layers of parietal and visceral pleura. Although pleural thickening is extensive, this does not involve mediastinal pleura (i.e., it is not circumferential). Areas of atelectasis are present in right lower lobe. B, i-cm collimation CT scan in a patient with a tuberculous empyema shows anterior mediastinal pleural involvement (arrow).
difficult to make by clinical criteria, pleurocentesis, and percutaneous pleural biopsy. Not rarely, a diagnostic thoracotomy is required because of the need to assess the growth pattern of disease and to obtain a relatively large amount of tissue for histologic, immunocytochemical, and ultrastructural studies [1 1 ]. Even careful examination and biopsy of the pleura at thoracotomy can fail to establish the diagnosis [17, 1 8]. Ryan et al. [1 7] reported 51 patients in whom no cause for a pleural effusion was found at thoracotomy. In 1 3 of these patients (25%), the diagnosis of malignancy (including lymphoma, carcinoma, and malignant mesothelioma) was established 1 2 days to 5 years after thoracotomy. The difficulty in pathologically differentiating benign reactive mesothelial cells from mesothelioma is a common and well-known probIem, and in many instances, the gross appearance assumes great importance in making this distinction [1 1 ]. Insofar as the CT appearance reflects the gross distribution of disease, we believe that the CT findings are helpful in determining whether thoracotomy is recommended and in guiding the surgeon to the sites most likely to yield a positive diagnosis. These sites would include areas with greater than 1 cm pleural thickening, nodular pleural thickening, and thickening of the mediastinal pleura. The presence of pleural calcification suggests a benign process. In this series, calcification was seen in 1 6 of 35 patients with benign pleural thickening and in only three of 39 patients with malignant pleural disease. Although calcified plaques may be seen in cases of mesothelioma, they are uncommon [3] and were not seen in any of our 1 1 mesothelioma patients. The relative absence of pleural calcification in mesothelioma may be due to absorption of calcification by the developing tumor [1 0] and the fact that a significant number of malignant mesotheliomas (range, 0-87%) are not asbestos-related [19]. Difficulty in distinguishing between mesothelioma and metastatic pleural involvement pathologically [11, 20] and radiologically [3, 5-7] has been recognized. Pathologic diagnosis requires adequate sampling; cytologic examination of pleural fluid and needle biopsy specimens rarely allows a definitive diagnosis of mesothelioma and may provide the misleading interpretation of metastatic carcinoma or benign pleural disease [9, 21, 22]. Discrimination between epithelial types of
mesothelioma and metastatic adenocarcinoma requires histochemical, immunohistochemical, and ultrastructural analysis [11, 12]. Invasion of the chest wall or mediastinum on CT are specific markers for malignancy but have a low sensitivity; they were present in only 8% of patients with malignant pleural involvement in our series. In addition, this sign was not useful in distinguishing mesothelioma from carcinoma. The limitations of CT in assessing the presence or absence of invasion into the chest wall or mediastinum in mesothelioma also have been demonstrated recently by Rusch et al. [23]. Adams et al. [24] have suggested that the presence of hilar adenopathy may be helpful in differentiating metastases from mesothelioma. However, true hilar involvement with no mediastinal involvement is rare in metastatic pleural disease except for bronchogenic carcinoma, lymphoma, and renal cell carcinoma [25]. In our study, hilar adenopathy was identified in only two patients with malignancy. One case was a nonHodgkin lymphoma and the second case was an adenocarcinoma. The tendency of mesothelioma to involve the inferior hemithorax has been postulated to be a specific sign [1 0]. However, in our study basal involvement was present not only in patients with mesothelioma but also in the majority of patients with metastatic pleural disease. Benign asbestos pleural effusion has been shown by Epler et al. [26] to occur with an approximate prevalence of 3% in the asbestos-exposed population. Effusion is usually unilateral but may be bilateral; other etiologic manifestations of asbestos-related disease may be present or absent. Differentiation between benign asbestos pleural effusion and asbestos-related malignant mesothelioma usually can be made on the basis of the much shorter latency period of benign effusion, which ranges from 1 0 to 20 years from time of initial exposure [26]. However, close follow-up is necessary as the effusion may be the first manifestation of mesothelioma [3], and routine radiologic studies have been advocated for mesothelioma screening in this high-risk population [22]. Nodularity of pleura has been recognized as a helpful discriminating feature [2, 3, 10]. In our study, nodular pleural thickening was seen in seven (63%) of 11 cases of mesothelioma and in none of the cases of benign asbestos-related
Downloaded from www.ajronline.org by Biblio Astrazeneca Des Science De La Sante Du Chul on 07/15/14 from IP address 132.203.227.61. Copyright ARRS. For personal use only; all rights reserved
492
LEUNG
diffuse pleural disease. However, extensive nodular thickening of pleural plaques mimicking mesothelioma has been described [3]. In such cases, open pleural biopsy is the only reliable way to establish the definitive diagnosis of benign disease. Although loss of volume is a common manifestation in mesothelioma [3], it is also a relatively nonspecific finding seen in a variety of other malignant and benign conditions [6]. In our series, volume loss was seen in approximately 50% of patients, irrespective of the cause of the diffuse pleural disease. Benign pleural asbestos-related disease generally is acknowledged to be a symmetrical process [27, 28], and in all 1 6 of our cases, pleural involvement was bilateral. Typically, the changes affect the posterolateral aspect of the pleura [29]. In one of our cases of early malignant mesothelioma, the only sign that suggested possible tumor was the unilaterality of pleural involvement. Although this is not a useful feature in differentiation from other benign diseases, in the select asbestos-exposed subgroup, unilaterality or marked asymmetry of pleural thickening must be considered suggestive of mesothelioma. Asbestos-related pleural disease was the only benign disease that could be specifically suggested from the CT scan by virtue of the presence of bilateral pleural plaques. Although smooth pleural thickening with associated effusion was suggestive of empyema, such features also were observed in a number of noninfectious processes. No characteristic features of fibrothorax were identified. Aberle et al. [29, 30] found that high-resolution CT was more sensitive than conventional CT in detection of asbestosrelated pleural and parenchymal fibrosis. Our study similarly shows that the scans with higher spatial resolution improved detection of subtle asbestos-related parenchymal changes and allowed confirmation of pleural changes deemed equivocal on conventional CT. Our study was limited by the relatively small number of cases of some diseases and by the relatively small number of diseases included. However, as all consecutive cases of confirmed diffuse pleural disease studied during a 4-year period were included, this study is a reflection of the incidence and prevalence of the types of diffuse pleural disease posing diagnostic problems in our institution. We conclude that the CT features most helpful in distinguishing malignant from benign pleural disease are pleural rind, nodular pleural thickening, pleural thickening greater than 1 cm, and mediastinal pleural involvement. Pleural calcification usually implies a benign process. Loss of volume is not a helpful feature. In the majority of patients, mesothelioma cannot be distinguished from metastatic pleural disease on CT. High-resolution CT is superior to conventional CT, particularly in the assessment of asbestos-related pleural disease.
ET AL.
AJR:154,
mesothelioma and asbestosis using CT and conventional chest radiography. Radiology 1982:144:453-460 4. Williford ME, Hidalgo H, Putman CE, Korobkin M, Ram PC. Computed tomography of pleural disease. AJR 1983:140:909-914 5. Grant DC, Seltzer SE, Antman KH, Finberg HJ, Koster K. Computed tomography of malignant pleural mesothelioma. J Comput Assist Tomogr
i983;7:626-632 6. Salonen
0, Kivisaari
K, Tammilehto
7. 8. 9. 1 0. 11.
1 2.
13. 14. 15. 1 6.
L, Standertskjold-Nordenstam
L. Computed
tomography
K, Mattson special (Stockh) 1986;
reference to the mediastinal pleura. Acta Radiol (Diagn] 27:527-531 Alexander E, Clark RA, Colley DP, Mitchell SE. CT of malignant pleural mesothelioma. AJR i98i;1 37:287-291 Maftessanti M, Tommasi M, Pellegrini P. Computed tomography of free pleural effusions. Eur J Radio! 1987:7:87-90 Mirvis 5, Dutcher JP, Haney PJ, Whitley NO, Aisner J. CT of malignant pleural mesothelioma. AJR 1983; 1 40: 665-670 Kreel L. Computed tomography in mesothelioma. Semin Oncol 1981; 8:302-312 Roggli VL, Kolbeck J, Sanfillippo F, Shelburne JD. Pathology of human mesothelioma: etiologic and diagnostic considerations. Pathol Annu 1987;2:91-131 Ordonez NG. The immunohistochemical diagnosis of mesothelioma: differentiation of mesothelioma and lung adenocarcinoma. Am J Surg Patho! 1989:13:276-291 Sahn SA. The pleura. Am Rev Respir Dis 1988:138:184-234 Kollins SA. Computed tomography of the pulmonary parenchyma and chest wall. Radio! Clin North Am i977;1 5: 297-308 Herbert A. Pathogenesis of pleurisy. pleural fibrosis, and mesothelial proliferation. Thorax 1986:41 :176-189 Barrett NR. The pleura, with special reference to fibrothorax. Thorax
1970:25:515-524 Ryan CJ, Rodgers RF, Unni KK, Hepper NGG. The outcome of patients with pleural effusion of indeterminate cause at thoracotomy. Mayo Clin Proc 1981:56:145-149 18. Douglass BE, Carr DT, Bernatz PE. Diagnostic thoracotomy in the study of “idiopathic” pleural effusion. Am Rev Tuberc 1956:74:954-957 19. Peterson JT, Greenberg SD, Buffler PA. Non-asbestos related malignant mesothelioma: a review. Cancer i984;54:951-960 20. Szpak CA, Johnston WW, Roggli V, et al. The diagnostic distinction between malignant mesothelioma of the pleura and adenocarcinoma of the lung as defined by a monoclonal antibody (B72.3). Am J Pathol 1986:122:252-260 21 . Shearin JC, Jackson D. Malignant pleural mesothelioma. J Thorac Cardiovasc Surg i976;71 :621-627
22. Strankinga WFM, Sperber M, Kaiser MC, Stam J. Accuracy of diagnostic procedures Respiration
in the initial evaluation and follow-up (Base!) 1987;51 :179-187
of mesothelioma
23. Rusch VW, Godwin JD, Shuman WP. The role of computed
patients.
tomography
scanning
24. 25.
in the initial assessment and the follow-up of malignant pleural mesothelioma. J Thorac Cardiovasc Surg 1988:96; 171-177 Adams VI, Unni KK, Muhm JR. Jeff JR. llstrup DM, Bernatz PE. Diffuse malignant mesothelioma of pleura. Cancer 1986:58: 1540-1 551 Reed JC. Chest radiology: plain film patterns and differential diagnosis. Chicago: Year Book Medical Publishers, Inc. , 1981
26. Epler GA, McLoud TC, Gaensler EA. Prevalence and incidence of benign asbestos pleural effusion in a working population. JAMA 1982;247: 61 7-622 27. Sargent EN, Gordonson J, Jacobson G, Birnbaum W, Shaub M. Bilateral pleural
thickening:
a manifestation
of
asbestos
dust
exposure.
AJR
1978:131:579-585
REFERENCES 30. 1978:13:213-225
C-G, Somer
of pleural lesions with
17.
28.
of the lung and pleura. Semin Roentgeno!
1990
2. Katz D, Kreel L. Computed tomography in pulmonary asbestosis. Clin Radiol 1979:30:207-213 3. Rabinowitz JG, Efremidis SC, Cohen B, et al. A comparative study of
Preger
L. Asbestos-related
disease.
New York:
29. Aberle DR. Gamsu G, Ray CS, Feuerstein
1. Kreel L. Computed tomography
March
Grune
and Stratton,
IM. Asbestos-related
1978
pleural
and parenchymal fibrosis: detection with high-resolution CT. Radiology 1988:166:729-734 Aberle DR, Gamsu G, Ray CS. High-resolution CT of benign asbestosrelated diseases: clinical and radiologic correlation. AJR 1988:151: 883-891
