| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Dept of Radiology, Osaka Univ Medical School, 2-2 Yamadaoka, Suita, Osaka, 565-0871, Japan (T.J., N.M., O.H., N.T., H.N.); the Dept of Radiology, Univ of British Columbia and Vancouver Hospital and Health Sciences Centre, Canada (N.L.M.); the Dept of Radiology, National Kinki Chuo Hospital for Chest Disease, Osaka, Japan, (M.Akira); the First Dept of Internal Medicine, Kumamoto University School of Medicine, Japan (K.I., M.S., M.Ando); and the Pulmonary Div, Kumamoto Chu-oh Hospital, Japan (T.Y., T.K.). Received Jun 14, 1999; revision requested Aug 3; final revision received Jan 19, 2000; accepted Feb 1. Address correspondence to T.J. (e-mail: johkoh@radiol.med.osaka-u.ac.jp).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Thin-section CT scans in 111 patients with eosinophilic lung diseases—40 with chronic eosinophilic pneumonia, 16 with Churg-Strauss syndrome, 16 with allergic bronchopulmonary aspergillosis (ABPA), 13 with acute eosinophilic pneumonia, 12 with simple pulmonary eosinophilia, 11 with drug-induced eosinophilic pneumonia, and three with hypereosinophilic syndrome—were assessed independently by two observers. The observers recorded the abnormalities, diagnosis, and degree of confidence in the diagnosis.
RESULTS: The two observers made a correct first-choice diagnosis on average in 61% of readings. The correct diagnosis was made in 78% of cases of chronic eosinophilic pneumonia; 81%, acute eosinophilic pneumonia; 44%, Churg-Strauss syndrome; 84%, ABPA; 17%, simple pulmonary eosinophilia; 27%, drug-induced eosinophilic pneumonia; and 33%, hypereosinophilic syndrome. The two observers made a correct diagnosis with a high degree of confidence in 36% of readings. There was moderate agreement between the observers for the correct diagnosis (
, 0.47) and for the correct diagnosis with a high degree of confidence (, 0.59).
CONCLUSION: Although eosinophilic lung diseases often can be differentiated by means of thin-section CT, correlation between CT findings and careful clinical evaluation are required for a definitive diagnosis.
Index terms: Allergic bronchopulmonary aspergillosis (ABPA), 60.634 • Aspergillosis, 60.634 • Churg-Strauss syndrome, 60.623 • Computed tomography (CT), clinical effectiveness, 60.12118, 60.12119 • Computed tomography (CT), thin-section, 60.12118, 60.12119 • Lung, CT, 60.12118, 60.12119 • Lung, diseases, 60.632, 60.634, 60.64 • Lung, infection, 60.632, 60.634, 60.64 • Pneumonia, eosinophilic, 60.632 • Pneumonitis, hypersensitivity, 60.624
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A number of recent studies (2–6) have described the diagnostic accuracy of computed tomography (CT) in diffuse infiltrative lung diseases. These studies, however, have included only a small number of patients with eosinophilic lung diseases. Furthermore, although it has been suggested that eosinophilic lung diseases have characteristic CT findings (7–18), this conclusion is based on interpretation of CT findings in patients with known diagnoses. To our knowledge, there has been no study to assess the value of thin-section CT in distinguishing these entities. The aim of this study was to determine whether the subtypes of eosinophilic lung diseases can be differentiated on the basis of the pattern and distribution of abnormalities at thin-section CT.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
All patients fulfilled the diagnostic criteria for the specific diagnosis of each of the entities (Table 1) (1,19–23). Pulmonary tissue eosinophilia was diagnosed by means of bronchoalveolar lavage in 16 patients, both bronchoalveolar lavage and transbronchial biopsy in 77 patients, and open lung biopsy in 18 patients. The drugs that induced eosinophilic pneumonia were interferon with an herbal drug (sho-saiko-to; n = 4), methotrexate (n = 3), tetracycline (n = 2), chlorpropamide (n = 1), and nitrofurantoin (n = 1). All patients with simple pulmonary eosinophilia did not need treatment. The remaining 99 patients were treated with corticosteroids and responded promptly.
|
The CT scans were randomized and then retrospectively reviewed independently by two chest radiologists (T.J., K.I.). The observers were unaware of any clinical or histopathologic findings other than the age and sex of the patients. The observers were aware that only seven types of eosinophilic lung diseases were included in the study, but they did not know the frequency of the entities.
The CT scans were assessed for the presence, extent, and anatomic distribution of areas of ground-glass attenuation (opacity), consolidation, nodules, interlobular septal thickening, thickening of bronchovascular bundles, bronchiectasis, bronchial wall thickening, mucous plugging, pleural effusion, and lymph node enlargement. Lymph nodes were considered enlarged if the short-axis diameter at CT exceeded 10 mm. If pleural effusion was mentioned, unilateral or bilateral distribution also was recorded.
A nodule was defined as a focal, rounded area of high attenuation less than 3 cm in diameter, and it could be well defined or ill defined. When the nodule was located in the center of the lobule or lobular core, it was defined as a centrilobular nodule. Mucous plugging was distinguished from nodules by using the following criteria: Mucous plugging was defined strictly as linear or branching structures. Even if the nodular areas of high attenuation had the cross-sectional appearance of mucous plugging, they were classified as nodules.
The anatomic distribution was noted to be central if there was a predominance of abnormalities in the inner third of the lung, peripheral if there was a predominance of abnormalities in the outer third of the lung, and random if there was no predominance. Zonal predominance was assessed as being upper, lower, or random. Upper lung zone predominance was considered present when most of the abnormalities were above the level of the tracheal carina, and lower zone predominance was considered present when most of the abnormalities were below this level.
The CT findings were interpreted on the basis of previously published data (7–18) on the CT appearance of the seven subtypes of eosinophilic lung diseases, as summarized in Table 2. Before interpretation, an initial orientation session was held for both observers to review the descriptions of previously published data by using other cases not included in the present study. Subsequently, each observer noted the most appropriate diagnosis for each patient and graded the degree of confidence of the diagnosis as high (level 1) or low (level 2). Agreement between the two observers was assessed by using the statistic (24): poor, = 0–0.20; fair, = 0.21–0.40; moderate, = 0.41–0.60; good, = 0.61–0.80; excellent, = 0.81–1. The mean numbers of correct diagnoses and mean numbers of CT findings for the two observers were tabulated.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
The confident diagnosis was correct in 10.5 (91%) of the 11.5 cases of ABPA, 6.5 (93%) of the seven cases of acute eosinophilic pneumonia, 18.5 (93%) of the 20 cases of chronic eosinophilic pneumonia, four (62%) of the 6.5 cases of Churg-Strauss syndrome, and 0.5 (33%) of the 1.5 cases of drug-induced eosinophilic pneumonia (Table 3). In none of the three cases of hypereosinophilic syndrome or the 12 cases of simple pulmonary eosinophilia was a correct diagnosis made with a high degree of confidence. There was moderate agreement between the observers for the correct diagnosis (, 0.47) and for the correct diagnosis with a high degree of confidence (, 0.59).
The frequencies of CT findings are summarized in Table 4. Areas of ground-glass attenuation were found in the majority of cases (Figs 1–7). Airspace consolidation was found in all patients with chronic eosinophilic pneumonia (Fig 1) and drug-induced eosinophilic pneumonia (Fig 6), in the majority of patients with acute eosinophilic pneumonia (24 [92%] of the 26 readings) and ABPA (26 [81%] of the 32 readings), and less commonly in the other entities.
|
|
|
|
|
|
|
|
Thickening of bronchovascular bundles was seen in four (67%) of the six readings of hypereosinophilic syndrome, 16 (62%) of the 26 readings of acute eosinophilic pneumonia (Fig 2), and 14 (44%) of the 32 readings of Churg-Strauss syndrome (Fig 3), and it was found less commonly in the other diseases. Bronchiectasis was found in 24 (75%) of the 32 readings of ABPA (Fig 4) and was seen less commonly in the other diseases. Mucous plugging was found in 20 (63%) of the 32 readings of ABPA (Fig 4) and was not seen in the other diseases. In the patients with ABPA, both bronchiectasis and mucous plugging were distributed predominantly in the upper lung zones.
Pleural effusion was found in 18 (69%) of the 26 readings in patients with acute eosinophilic pneumonia and was seen less commonly in the other entities. All patients who had pleural effusion showed bilateral pleural effusion. Lymph node enlargement was seen in two (33%) of the six readings of hypereosinophilic syndrome, eight (25%) of the 32 readings of Churg-Strauss syndrome, and six (23%) of the 26 readings of acute eosinophilic pneumonia, and it was found less commonly in the other diseases.
Upper lung zone predominance was found in 14 (44%) of the 32 readings of ABPA, 10 (42%) of the 24 readings of simple pulmonary eosinophilia, 30.5 (38%) of the 80 readings of chronic eosinophilic pneumonia, eight (31%) of the 26 readings of acute eosinophilic pneumonia, and four (13%) of the 32 readings of Churg-Strauss syndrome. None of the patients with drug-induced eosinophilic pneumonia or hypereosinophilic syndrome showed upper lung zone predominance. A predominantly peripheral distribution was found in 68 (85%) of 80 readings of chronic eosinophilic pneumonia and was seen less commonly in the other diseases. Central distribution was present in 18 (56%) of the 32 readings of ABPA and less commonly in the other diseases. Acute eosinophilic pneumonia, Churg-Strauss syndrome, hypereosinophilic syndrome, simple pulmonary eosinophilia, and drug-induced eosinophilic pneumonia frequently showed random distribution in both cephalocaudal and cross-sectional planes.
Frequencies of characteristic combinations of CT findings are summarized in Table 5. A combination of airspace consolidation and peripheral predominance was prominent in chronic eosinophilic pneumonia (68 [85%] of the 80 readings) (Fig 1) and was seen less commonly in the other entities. A combination of bronchial wall thickening and bronchiectasis was found in 17 (53%) of the 32 readings of ABPA (Fig 4) and was less common in the other diseases. A combination of interlobular septal thickening, thickening of bronchovascular bundles, and pleural effusion was found in eight (31%) of the 26 readings of acute eosinophilic pneumonia, five (16%) of the 32 readings of Churg-Strauss syndrome, and none of the other eosinophilic lung diseases.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although acute eosinophilic pneumonia is a relatively new disease entity (1,20), the CT findings usually are fairly extensive and characteristic and include areas of ground-glass attenuation, airspace consolidation, poorly defined nodules, interlobular septal thickening, and pleural effusion (9,10). These CT features were also commonly seen in the present study, together with additional findings of bronchovascular bundle thickening and random distribution in both cephalocaudal and cross-sectional planes.
A lower accuracy in diagnosis was observed in the remaining four entities of eosinophilic lung disease considered in this study, with a correct diagnosis in 44% of 16 cases of Churg-Strauss syndrome, 33% of three cases of hypereosinophilic syndrome, 27% of 11 cases of drug-induced eosinophilic pneumonia, and 17% of 12 cases of simple pulmonary eosinophilia. To our knowledge, there is relatively little information in the existing literature in regard to the CT manifestations of these entities, the descriptions being based on the findings in a small number of patients (11,12,16–18).
From both present and previous studies, airspace consolidation, areas of ground-glass attenuation, nodules, interlobular septal thickening, bronchial wall thickening, and bronchiectasis are common CT findings in Churg-Strauss syndrome (12). In two previous reports (17,18) and in the present study, patchy areas of airspace consolidation, nodules, diffuse areas with ground-glass attenuation, interlobular septal thickening, thickening of the bronchovascular bundles, bronchial wall thickening, and random distribution were characteristic in hypereosinophilic syndrome. Common CT findings in drug-induced eosinophilic pneumonia are areas of ground-glass attenuation, airspace consolidation, nodules, and interlobular septal thickening. In the previous (18) and present studies, common CT findings in simple pulmonary eosinophilia were areas of ground-glass attenuation, nodules, airspace consolidation, and bronchial wall thickening.
In the present study, bilateral pleural effusion was seen in more than half of patients with acute eosinophilic pneumonia and was seen less commonly in drug-induced eosinophilic pneumonia, Churg-Strauss syndrome, and chronic eosinophilic pneumonia. Interlobular septal thickening was seen in association with pleural effusion in acute eosinophilic pneumonia, Churg-Strauss syndrome, and drug-induced eosinophilic pneumonia.
Eosinophils are known to infiltrate the interstitium, including the subpleural space and interlobular septa (25). These interstitial compartments also contain abundant lymphatic vessels, which are the main pathway for the drainage of parenchymal fluid (26). In some cases, the findings of septal thickening and pleural effusion may reflect direct infiltration of eosinophils, alterations of lymphatic drainage, or both. The basis for lymph node enlargement is not known, but it is of interest that lymph node enlargement was most prominent in hypereosinophilic syndrome, Churg-Strauss syndrome, and acute eosinophilic pneumonia, entities in which interlobular septal thickening, pleural effusion, or both were also present.
Identification of characteristic combinations of CT findings may be of additional assistance in the differentiation of eosinophilic lung diseases. In the present study, a combination of airspace consolidation and peripheral predominance was more prominent in patients with chronic eosinophilic pneumonia, seen in 85% (68 of 80) of the readings, than in those with other eosinophilic lung diseases. A combination of bronchial wall thickening and bronchiectasis was seen more commonly in patients with ABPA—53% (17 of 32) of the readings. The combination of interlobular septal thickening, thickening of bronchovascular bundles, and pleural effusion was seen only in patients with acute eosinophilic pneumonia (31% [eight of 26 readings]) and Churg-Strauss syndrome (16% [five of 32 readings]).
Although the presence of individual or combined CT findings was often very helpful, there was still considerable overlap in these findings among the eosinophilic lung diseases. In all entities, areas with ground-glass attenuation and airspace consolidation were seen commonly. Nodules, bronchial wall thickening, and bronchiectasis were common CT findings in both Churg-Strauss syndrome and ABPA. Therefore, ABPA was misdiagnosed most frequently as Churg-Strauss syndrome. Interlobular septal thickening, thickening of bronchovascular bundles, and random distribution were seen commonly in the patients with acute eosinophilic pneumonia, Churg-Strauss syndrome, drug-induced eosinophilic pneumonia, and hypereosinophilic syndrome. Consequently, Churg-Strauss syndrome, drug-induced eosinophilic pneumonia, and hypereosinophilic syndrome were misdiagnosed most frequently as acute eosinophilic pneumonia.
Precise diagnosis ultimately depends on careful correlation of all CT findings with clinical information, which was not available to the two observers in the present study. In chronic eosinophilic pneumonia, the symptoms are insidious and are present for at least 1 month prior to diagnosis (19,27). ABPA is characterized by a history of asthma and evidence of an allergic reaction to Aspergillus fumigatus (22). Specific clinical findings in acute eosinophilic pneumonia are an acute febrile illness for 1–5 days accompanied by hypoxemic respiratory failure, which often requires mechanical ventilation (10,20). Churg-Strauss syndrome is characterized by systemic vasculitis occurring in patients with asthma and allergic rhinitis (21,28).
Drug reaction is one of the most common causes of pulmonary abnormalities with blood and/or tissue eosinophilia (16,18). Therefore, it is essential for the diagnosis of drug-induced eosinophilic pneumonia to find out the history of exposure to drugs known to induce eosinophilia. Hypereosinophilic syndrome is a rare and fatal multiorgan disorder characterized by blood eosinophilia of more than 0.15 x 109/L for more than 6 months and symptoms and signs of pulmonary damage related to increased eosinophils (29). Patients with simple pulmonary eosinophilia do not need any treatment because it typically resolves spontaneously within 1 month (1).
Our study had several limitations. First, this study was retrospective. Second, because no clinical information was given, the observers were at an unrealistic disadvantage. Third, the study was biased because in clinical practice the differential diagnosis must include entities other than eosinophilic lung diseases. Radiologists will have to consider that there are many noninfectious or infectious disorders that should be differentiated from various eosinophilic lung diseases in daily clinical practice. However, the main goal of the study was to determine whether the findings in the eosinophilic lung diseases are characteristic enough to allow their distinction at thin-section CT. This goal was achieved, we believe, by having the observers review the findings without awareness of the clinical data and by committing to a diagnosis based on previous data in the literature.
On the basis of the results of the present study, we consider the most helpful findings in distinguishing the eosinophilic lung diseases to be (a) the presence of peripheral airspace consolidation to diagnose chronic eosinophilic pneumonia; (b) a combination of bronchial wall thickening, bronchiectasis, and mucous plugging to diagnose ABPA; and (c) a combination of interlobular septal thickening, bronchovascular bundle thickening, and pleural effusion as most suggestive of acute eosinophilic pneumonia. However, the study results show considerable overlap between the CT findings in the various entities, which precludes a confident diagnosis in the majority of cases. In conclusion, although eosinophilic lung diseases often can be differentiated by means of thin-section CT, correlation between CT findings and careful clinical evaluation are required for a definitive diagnosis.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, T.J.; study concepts, T.J., N.L.M.; study design, T.J., M.Akira; definition of intellectual content, T.J., K.I.; literature research, T.J., N.M.; clinical studies, T.J., M.Akira, T.Y., T.K.; experimental studies, T.J., O.H., N.T.; data acquisition, T.J., K.I.; data analysis, T.J., N.M.; statistical analysis, T.J., N.M.; manuscript preparation, T.J., N.L.M.; manuscript editing, T.J., N.L.M., M.S.; manuscript review, N.L.M., M.Ando, H.N.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  F. Okada, Y. Ando, S. Yoshitake, A. Ono, S. Tanoue, S. Matsumoto, M. Wakisaka, T. Maeda, and H. Mori Clinical/Pathologic Correlations in 553 Patients With Primary Centrilobular Findings on High-Resolution CT Scan of the Thorax Chest, December 1, 2007; 132(6): 1939 - 1948. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. J. Jeong, K.-I. Kim, I. J. Seo, C. H. Lee, K. N. Lee, K. N. Kim, J. S. Kim, and W. J. Kwon Eosinophilic Lung Diseases: A Clinical, Radiologic, and Pathologic Overview RadioGraphics, May 1, 2007; 27(3): 617 - 637. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. A. Souza, N. L. Muller, T. Johkoh, and M. Akira Drug-Induced Eosinophilic Pneumonia: High-Resolution CT Findings in 14 Patients Am. J. Roentgenol., February 1, 2006; 186(2): 368 - 373. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Philit, B. Etienne-Mastroianni, A. Parrot, C. Guerin, D. Robert, and J.-F. Cordier Idiopathic Acute Eosinophilic Pneumonia: A Study of 22 Patients Am. J. Respir. Crit. Care Med., November 1, 2002; 166(9): 1235 - 1239. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
