Eur Radiol (2003) 13:43–51 DOI 10.1007/s00330-002-1422-2

CHEST

Kyung Soo Lee Tae Sung Kim Kiminori Fujimoto Hiroshi Moriya Hideyuki Watanabe Ukihide Tateishi Kazuoto Ashizawa Takeshi Johkoh Eun A. Kim O Jung Kwon

Thoracic manifestation of Wegener’s granulomatosis: CT findings in 30 patients

Received: 30 October 2001 Revised: 28 January 2002 Accepted: 8 February 2002 Published online: 9 May 2002 © Springer-Verlag 2002

K. Ashizawa Department of Radiology, Nagasaki University School of Medicine, Nagasaki 852–8501, Japan

K.S. Lee (✉) · T.S. Kim · E.A. Kim Department of Radiology, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50, Ilwon-Dong, Kangnam-Ku, Seoul 135–710, Korea e-mail: [email protected] Tel.: +82-2-34102511 Fax: +82-2-34102559 K. Fujimoto Department of Radiology, Kurume University School of Medicine, Kurume 830–0011, Japan H. Moriya Department of Radiology, Fukushima Medical University School of Medicine, Fukushima 960–1295, Japan H. Watanabe Department of Radiology, University of Occupational and Environmental Health, Japan School of Medicine, Kitakyushu 807–8555, Japan U. Tateishi Department of Radiology, Hokkaido University Graduate School of Medicine, Sapporo 060–8638, Japan

T. Johkoh Department of Radiology, Osaka University Graduate School of Medicine, Suita 565–0871, Japan O.J. Kwon Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50, Ilwon-Dong, Kangnam-Ku, Seoul 135–710, Korea

Abstract Our objective was to describe the CT findings of thoracic Wegener’s granulomatosis. At presentation, both conventional and thinsection CT scans were available in 30 patients with Wegener’s granulomatosis. Serial CT scans (range of intervals: 1–25 months, mean 4.5 months) were available in 20 patients. The initial and follow-up CT scans were analyzed retrospectively by two observers in terms of pattern and extent of parenchymal and airway lesions. Positive CT findings were seen in 29 of

Introduction Wegener’s granulomatosis (WG) is characterized clinicopathologically by the triad of granulomatous necrotizing and ulcerative inflammatory process of the respirato-

30 (97%) patients at initial presentation. The most common pattern was nodules or masses seen in 27 of 30 (90%) patients. They were multiple in 23 of 27 (85%) patients, bilateral in 18 (67%), subpleural in 24 (89%), and peribronchovascular in 11 (41%) in distribution. Bronchial wall thickening in the segmental or subsegmental bronchi was seen in 22 (73%) patients. Large airways were also abnormal in 9 (30%) patients. Patchy areas of consolidation and ground-glass opacity were seen in 7 (23%) patients, respectively. In 17 of 20 (85%) patients in whom follow-up CT scans were available, the parenchymal or airway lesion showed complete or partial improvement with treatment. The CT findings of Wegener’s granulomatosis, although multiple and variable, consist mainly of bilateral subpleural or peribronchovascular nodules or masses and bronchial wall thickening in the segmental or subsegmental bronchi. Parenchymal and airway lesions improve with treatment in most patients. Keywords Lung · Granuloma · CT · Granuloma vasculitis · Wegener’s granulomatosis

ry tract and the internal organs, generalized necrotizing granulomatous vasculitis, and a usually focal, necrotizing glomerulonephritis [1]. The lung is the most commonly involved organ and almost all patients have either pulmonary or upper airway involvement [2].

44

Wegener’s granulomatosis in the thorax presents with wide variety of radiologic findings [3, 4, 5, 6, 7, 8, 9, 10, 11]. More than one pattern of radiologic findings may be seen on CT scans at presentation or during the course of the disease. Because chest radiographs fail to show the exact pattern and the extent of thoracic involvement, CT is particularly useful in the assessment of thoracic involvement of the disease [12]. Recent improvement in the treatment of the disease with combined cytotoxic and corticosteroid therapy has led to good prognosis and long-term survival; however, serial CT findings of WG have been published in a limited number of papers [13]. The purpose of this study was to present the initial and follow-up CT findings of thoracic WG in a relatively large number of patients.

Materials and methods Between March 1993 and April 2001, 30 patients with WG were seen in seven tertiary hospitals. The patients were identified retrospectively by computerized search of all patients’ records. They were 16 men and 14 women (age range 15–80 years, mean age 54.3 years). The diagnosis of WG was made in accordance with the 1990 American College of Rheumatology [14] and the Chapel Hill criteria [15]. Histologic verification was done with biopsies in 29 patients. The positive biopsy specimens were obtained in the lungs including airways (n=13 [13 patients]), paranasal sinuses or nasal cavity (n=15), kidneys (n=8), and skins including eyelids (n=2). C-antineutrophil cytoplasmic antibodies (c-ANCA) were evaluated in the peripheral blood at the time of the initial diagnosis in 28 patients. They were elevated in 24 patients. In the remaining 4 patients, they were not elevated. Twenty-nine patients had symptoms and signs at admission including cough (n=15), nasal obstruction (n=11), fever (n=6), fatigue (n=6), dyspnea (n=4), hemoptysis (n=4), proptosis (n=3), headache (n=2), scleral ulcer (n=2), sputum (n=2), generalized edema (n=2), and diarrhea (n=1). The remaining 1 patient, in whom pulmonary abnormalities were detected incidentally on routine chest radiograph, was asymptomatic. The involved organs were the lungs or airways (n=29), the paranasal sinuses (n=26), kidneys (n=14), eyes (n=5), skin (n=1), and oral cavity (n=1). Both cyclophosphamide and corticosteroids were given in 23 patients and corticosterioids only were given in 7 patients. Computed tomography scans were obtained with different machines, but mainly with three: GE 9800 or HiSpeed Advantage scanner (General Electric Medical Systems, Milwaukee, Wis.), a X-Vigor Scanner (Toshiba, Tokyo, Japan), and Somatom Plus 400 scanner (Siemens, Inselin, Germany). At initial evaluation, conventional CT scans were obtained from the lung apices to the middle portion of both kidneys with 7- to 10-mm collimation. The scans were obtained with (n=12) and without (n=18) intravenous injection of contrast medium (a total of 100 ml of 30 g iodine). Thin-section (1- to 2-mm thickness) CT scans were also obtained through the thorax at 20-mm intervals before (n=30, with full inspiration) and after (n=7, with full expiration) conventional CT scans. Expiratory thin-section CT scans were obtained only in 7 patients suspected clinically of airway involvement. Follow-up CT scans were available in 20 patients. When follow-up CT scans were performed more than twice in 1 patient, the last CT scans were analyzed. Intervals between the initial and follow-up CT scans ranged from 1 to 25 months (mean 4.5 months). The imaging data were reconstructed with bone algorithm. The scans were imaged by using lung (window width, 1500–2000 HU; window

level, –700–800 HU) and mediastinal (window width, 300– 400 HU; window level, 0–20 HU) window settings. Two independent chest radiologists of 2- and 7-year experiences, respectively, analyzed the CT scans retrospectively. Final decisions on the findings were reached by a consensus with a third reviewer who took a role of mediator when the two reviewers had different opinions. On initial CT scans, the patterns of parenchymal abnormalities were subdivided into being nodule (3 cm or less in diameter) or mass (>3 cm in diameter), consolidation (increased opacity with obscuration of underlying vessels), and ground-glass opacity (without obscuration of underlying vessels). When nodules <10 mm were identified, the presence of accompanying branching linear structures was also evaluated (centrilobular nodules and branching linear structures). The distribution of each pattern of parenchymal abnormality in each patient was classified as being in upper, middle, and lower lung zone; as being central, subpleural, and random; as being diffuse, patchy, and random; and as being peribronchovascular and random. Lesions were considered to be located in the upper lung zone when they were seen superior to the aortic arch, in the lower lung zone when seen inferior to the inferior pulmonary vein, and in the middle lung zone when seen between the two. Lesions were regarded central when they were located within 3 cm from the hilum, subpleural when located within 3 cm from the visceral pleura, and random otherwise. Lesions were regarded as diffuse when they were widespread and continuous, patchy when they were multiple in small pieces separated from each other, and random otherwise. Lesions were regarded as bronchovascular when they were located along the axial interstitium and random otherwise. As for nodules or masses, the shortaxis diameter of the smallest and largest lesions was recorded in lung window setting. When there were cavitary changes in nodules, the number of patients who had cavitary nodule(s) was counted. The number of cavitary nodules among the total number of nodules was also counted. The presence of halo sign (nodules surrounded by areas of ground-glass opacity) was evaluated. The attenuation of nodules and masses in mediastinal window setting, when they were more than 20 mm in diameter, was compared with that of chest wall muscles and was subdivided into being low, iso, and high. The abnormalities of the large airways from the trachea down to the lobar bronchi were subclassified in its patterns into wall thickening, intraluminal nodule, and extraluminal mass. The walls of trachea as well as main and lobar bronchi were regarded as thickened when its wall was >3.0 mm in thickness on lung window images [16]. The extent of large airway abnormalities was subclassified into diffuse (>3 cm in involved length) or focal (<3 cm). Bronchial wall was regarded as thickened in the segmental and subsegmental bronchi when their walls were >1.5 mm in thickness [16, 17]. Bronchi imaged perpendicular or horizontal to the scan plane were chosen. Bronchi imaged oblique to the scan plane were not measured. When the bronchial wall thickness was difficult to measure on films, the images were magnified into a screen by using an overhead projector and bronchial wall thickness was assessed. Presence of bronchiectasis, hilar and mediastinal lymph node enlargement, as well as pleural and pericardial effusion was also evaluated. Bronchiectasis was regarded as present when the airways distal to lobar bronchi had diameters greater than those of accompanying arteries or did not show tapering with distal branching. Air trapping was regarded as present when expiratory scans showed mosaic hypoattenuation in lung parenchyma. On the follow-up CT scans, overall extent and the extent of each pattern of the parenchymal abnormality and airway abnormalities were evaluated and classified into being totally disappeared, partially disappeared, stable, increased, or newly appeared.

N/A

N/A

7 0 0

N/A

N/A

7 0 5

N/A 3 4 3 2 Bronchiectasis (n=4, 13%)

2

15 12 13 14 8 Bronchial wall thickening (n=22, 73%)

2 5

N/A

2 0 5 4

7 0 1 6 0 2 4 2 5

Consolidation (n=7, 23%) GGO (n=7, 23%)

3

N/A 3 24 0 18 19 21 17 10

5

7 0 0

Random Patchy Diffuse Random Subpleural Central Lower Middle Upper Uni

Bi

Distribution Laterality

Nodule(s)/mass(es) (n=27, 90%)

Fig. 1 Wegener’s granulomatosis in a 48-year-old man who presented with scleral ulcer. Thin-section (1.0-mm collimation) CT scan obtained at level of left innominate vein shows two cavitating nodules (curved arrow) in left upper lobe. Also note small nodules (arrows) in right upper lobe

Pattern

Positive CT findings were seen in 29 of 30 (97%) patients. In 1 patient in whom CT findings were normal, the diagnosis of WG was made with positive paranasal sinus biopsy and elevated c-ANCA level. The most common finding was nodules or masses seen in 27 of 30 (90%) of patients (Figs. 1, 2, 3, 4). Twenty-seven patients, each having 1–32 nodules or masses (mean 8), had a total of 216 nodules or masses ranging in size from 2 to 80 mm. The nodules or masses were multiple in 23 (85%) patients and bilateral in 18 (67%; Figs. 1, 2, 3). They were predominantly subpleural (Figs. 1, 2, 4) in 24 (89%) patients and peribronchovascular (Fig. 3) in 11 (41%; Table 1). They were isolated findings in 8 (30%) patients and mixed with other findings in the remaining 19 (70%) patients. Cavitation (Fig. 1) was noticed in 33 of 216 (15%) nodules or masses and was seen in 13 (48%) of 27 patients who had nodules or masses. The CT halo sign was found in 4 (15%) of 27 patients (Fig. 3). In 11 of 12 patients in whom enhanced scans were obtained, the nodules or masses were present in the lungs and showed almost always extensive central low attenuation with (n=4) or without (n=7) peripheral higher attenuation (Fig. 4). In 16 of 18 patients in whom unenhanced scans were obtained, they were present in the lungs and showed isoattenuation. Consolidation (Fig. 5) was seen in 7 patients and associated with other abnormalities in 6 (86%) patients. It was subpleural (6 of 7 patients, 86%) and patchy (7 patients, 100%). Ground-glass opacity was seen also in 7 patients (23%) and was associated with other abnormalities in 6 (86%) patients. It was random (5 of 7 patients, 71%) and patchy (7 patients, 100%; Table 1).

Table 1 Initial CT findings of thoracic Wegener’s granulomatosis in 30 patients. Uni unilateral; Bi bilateral; GGO ground-glass opacity; N/A not applicable

Initial CT findings

11

Bronchovascular Random

Results

16

45

46

Fig. 2a–c Wegener’s granulomatosis in a 57-year-old man who presented with blood-tinged sputum and cough. a Mediastinal window of enhanced (7.0-mm collimation) CT scan obtained at level of thoracic inlet shows eccentric wall thickening (arrow) and luminal narrowing of trachea with extraluminal component (asterisk) of lesion. Also note mass (curved arrow) in right upper lobe. Pneumomediastinum (arrowhead) was due to previous tracheal biopsy. Bronchoscopic biopsy specimen (not shown here) from tracheal wall disclosed necrotizing inflammation and granulation tissue along with microabscess formation. b Follow-up CT obtained at similar level to and 5 months after a shows improved tracheal lesion with luminal widening. Nodularity and wall thickening (arrow) still remain. Also note improvement of previous mass lesion in right upper lobe. c Follow-up CT obtained at level of basal trunk shows recurrent nodules in both lower lobes and nodular pleural thickening in right hemithorax (arrows). Nodules in left lower lobe have surrounding groundglass opacity (halo sign; small arrows)

Fig. 3a, b Wegener’s granulomatosis in a 56-year-old man who presented with cough, sputum, and fever. a, b Thin-section (1.5-mm collimation) CT scans obtained at level of right upper lobar bronchus shows nodules and masses (curved arrows) mainly along bronchovascular bundles. Some nodules show halo sign (small arrows). Also note marked bronchial wall thickening in segmental and subsegmental bronchi (arrows)

N/A

N/A

2 3 0 0

N/A

N/A

0 1 0 2

N/A 1 1 3 2 Bronchiectasis (n=4, 13%)

1

1 Bronchial wall thickening (n=22, 73%)

6

3

2

3

N/A

2 1 0 0 0 2 2 2 0 2 2 2 Consolidation (n=7, 23%) GGO (n=7, 23%)

0 1

6 Nodule(s)/mass(es) (n=27, 90%)

9

11

9

7

0

13

2

N/A

2 2

Patchy Diffuse Subpleural Central Lower Middle Upper Uni

Bi

Distribution Laterality Pattern

Table 2 Follow-up CT findings of thoracic Wegener’s granulomatosis in 20 patients

Centrilobular small nodules and branching linear structures were found in three patients. In 2 patients the lesions were mixed with nodules, masses, ground-glass opacity, and bronchial wall thickening. In the remaining patient, they were associated with nodules, consolidation, ground-glass opacity, and bronchial wall thickening. Bronchial wall thickening in the segmental and subsegmental bronchi (Figs. 3, 5, 6) was discovered in 22 (73%) patients. The bronchial lesion was bilateral in 14 (64%) and unilateral in 8 (36%). Bronchial wall thickening was accompanied by other abnormalities in 21 patients (92%) and was isolated finding in 1 (8%) patient. Large airways were abnormal in 9 (30%) patients. The main bronchi were involved in 8 patients, the trachea in 4, the bronchus intermedius in 2, and the right lower lobar bronchus in 1 patient. The abnormalities were always associated with other findings. Wall thickening with luminal narrowing (Fig. 5) was seen in 8 patients. Nodular wall thickening with luminal obliteration (Fig. 6) was seen in 1 patient. Extraluminal soft tissue lesion was seen in 1 patient (Fig. 2). The extent of large airway involvement was focal in all 9 patients. Bronchiectasis was found in 4 (13%) patients. It was unilateral in 2 and bilateral in another 2 patients. Patchy air trapping was seen in 2 (29%) of 7 patients in whom expiratory scans were obtained. Mediastinal lymph node enlargement was discovered in 6 (20%) patients. Right lower paratracheal (n=4), subcarinal (n=3), and left lower paratracheal (n=2) lymph nodes were enlarged. Hilar (right hilum) node enlargement was seen in 1 (3%) patient. A small

Random

Fig. 4 Wegener’s granulomatosis in a 53-year-old woman with rhinorrhea. Mediastinal window of enhanced CT (1-mm collimation) scan shows large mass (arrows) in anterior segment of right upper lobe with extensive geographic central low-attenuation area

0 0

14 1

Random

Bronchovascular Random

47

48

Fig. 5a, b Wegener’s granulomatosis in a 78-year-old woman who presented with dyspnea and rhinorrhea. a Thin-section (1.0-mm collimation) CT scan obtained at level of bronchus intermedius shows consolidation (curved arrows) in superior segment of left lower lobe and cavitating and noncavitating nodules (arrowhead) in lingular division of left upper lobe. Also note

marked bronchial wall thickening in bronchus intermedius (arrow) and subsegmental bronchus (small arrows). b A CT scan obtained at level of liver dome shows patchy areas of ground-glass opacity, cavitating nodules, small centrilobular nodules (arrowheads), and parenchymal bands (small arrows). Also note marked bronchial wall thickening in subsegmental bronchi (arrows)

Fig. 6a–c Wegener’s granulomatosis in a 41-year-old woman who presented with hemoptysis and dyspnea. a, b Thin-section (1.5-mm collimation) CT scans obtained at levels of a right middle lobar bronchus and b right basal trunk, respectively, show obliteration (arrows) of right lower lobar bronchus in a and nodular thickening (curved arrow) of basal trunk in b. Also note bronchiectasis with bronchial wall thickening (small arrows) in right middle and lower lobes and nodules (arrowheads) in right middle lobe. Bronchoscopy showed inflammatory wall thickening with granulation tissue in right basal trunk (not shown here). c Followup CT obtained at similar level to and 23 months after b shows enlarged basal trunk (arrows) with disappearance of bronchial wall thickening. Also note bronchiectasis in right middle and lower lobes. Compared with b, bronchial wall thickening has decreased in extent at subsegmental level

amount of pericardial effusion was seen in 5 (17%) patients. Pleural effusions were seen in 4 (13%) patients (bilateral in 3 and unilateral in 1). They were small in amount. Nodular pleural thickening was seen in 2 (7%) patients. Follow-up CT findings Of 20 patients with follow-up CT studies, 5 showed totally disappeared parenchymal and airway lesions with some areas of irregular linear opacity at the site of previous lung lesions. In 12 patients the parenchymal and airway lesion improved with partial disappearance. In 3 patients the lesions increased in extent (Table 2). In 1 patient eccentric and nodular bronchial wall thickening in

49

the right basal trunk on the initial scans decreased in extent with apparent bronchial widening and normalized wall thickness (Fig. 6); however, eccentric and nodular bronchial wall thickening appeared newly in the lower lobar bronchus of the contralateral left lung. In the other patient, in whom tracheal and parenchymal lesions were seen on initial CT scans, tracheal and some parenchymal lesions decreased in extent; however, parenchymal lesion appeared newly in other areas (Fig. 2). In the third patient, in whom the second follow-up CT showed partial disappearance of the initial lesions of multiple nodules and bronchial wall thickening, the final follow-up CT showed extensive patchy areas of consolidation and ground-glass opacity along with centrilobular nodules and branching linear structures, suggesting relapse of parenchymal lesion. Bronchoalveolar lavage fluid demonstrated pulmonary hemorrhage.

Discussion Pulmonary WG is characterized histopathologically by vasculitis, necrosis, and granulomatous inflammation accompanied by a mixed cellular infiltration of neutrophils, lymphocytes, plasma cells, histiocytes, and eosinophils. Other histologic findings include neutrophilic microabscesses and geographic necrosis. Tissue eosinophilia, as well as focal acute and chronic pulmonary hemorrhage, areas of endogenous lipoid pneumonia, and lymphoid aggregates may also be associated. Pulmonary WG appears grossly as gray-white, solid or cavitary nodules or masses throughout the lungs. The lesions often coalesce to produce large geographic areas containing brown-red areas of necrosis. In approximately 25% of cases, diffuse pattern of reddish pulmonary hemorrhage, tan fibrotic parenchyma, or yellow consolidation from endogeneous lipid pneumonia is seen instead of discrete nodules [18]. On CT, WG appears as nodules or masses. Central cavitation typically occurs in nodules measuring greater than 2 cm in diameter. The walls of centrally cavitated lesion may be thick and irregular or thin and smooth [7, 12]. Feeding vessel sign (vessels heading to nodular lesions) suggesting angiocentric distribution of the disease may be frequently seen [6, 12]. Pleural-based, wedgeshaped lesions simulating pulmonary infarction are also common [6]. The CT halo sign (a rim of ground-glass opacity surrounding a pulmonary lesion) [19, 20] has been described in WG. In our study the nodules or masses were predominantly subpleural or peribronchovascular in distribution. Another important finding we saw was extensive central low attenuation with or without peripheral higher attenuation on mediastinal window of enhanced scans. The low attenuation may reflect pathologic findings of extensive geographic necrosis of the disease [18].

Bronchial abnormalities, including bronchiectasis and peribronchial thickening of the small airways, once regarded as unusual CT manifestations of WG, have been reported in approximately 40% of cases [5, 7, 12]. In the current study, however, bronchial wall thickening at segmental and subsegmental bronchial levels was seen in 22 (73%) of 30 patients. Because the involvement of the segmental and subsegmental bronchi was so striking, these were one of the main CT findings of WG along with nodules or masses. Because vasculitis, parenchymal necrosis, and granulomatous inflammation are regarded as major pathologic manifestations of WG, bronchial and bronchiolar abnormalities have received little attention; however, Travis et al. [18] saw frequent bronchial and bronchiolar changes including nonspecific chronic inflammation (53 of 82 specimens, 64%), acute inflammation (42 specimens, 51%), bronchiolitis obliterans (26 specimens, 31%), follicular bronchiolitis (24 specimens, 28%), and other inflammatory lesions on pathologic study. These inflammatory changes in the bronchi on pathologic examinations may corroborate our frequent observation of bronchial wall thickening. Bronchiectasis was seen in 4 (13%) of 30 patients. Assessment of bronchial wall thickness on thin-section CT is quite subjective and depends on the window settings. Phantom study shows that accurate assessment of bronchial thickness is achieved when we use a display level of –450 HU with window width of 1500 [21, 22]; therefore, our study (using –700 or –800 HU) may not exactly reflect the real thickness of the measured bronchi and overestimate the thickness. However, in clinical practice, because airways as well as lung parenchyma should be evaluated, we used –700 or –800 HU window level. Large airway involvement of WG appears as focal or elongated segments of stenosis, as well as intra- and extraluminal soft tissue masses or thickening with or without lobar or segmental atelectasis [3]. Tracheobronchial abnormalities were seen in 9 (30%) of 30 patients in our study. Wall thickening and mild luminal narrowing were main findings; however, nodular wall thickening with luminal narrowing and extraluminal soft tissue lesion were also seen. Main bronchi, trachea, bronchus intermedius, and lower lobar bronchus were involved in the order of frequency. The abnormality was always focal in extent with short segment involvement. Unlike parenchymal lesions, the airway lesions have been usually reported not to show improvement with drug treatment [7, 9]; however, in our study, except in 1 patient in whom recurrent disease occurred in other airway with improvement of the initial lesion, airway disease showed improvement with treatment (Fig. 6). Centrilobular nodules and branching linear structures, which may suggest the underlying pathologic findings of small vessel vasculitis, were seen in 3 patients on initial CT scans and in 1 with recurrence on follow-up CT

50

scans. Similar findings of centrilobular nodules and branching linear structures were also reported in patients with Churg-Strauss syndrome (allergic angiitis and granulomatosis) [23]; however, in the setting of disseminated bronchial wall thickening as in the current study, the nodules and branching linear structures may represent bronchiolar inflammatory changes rather than vasculitis. Pathologic correlation on the centrilobular nodules and branching linear structures on CT remains to be studied. Aberle et al. [3] had a concern about frequent recurrence of WG, frequent pulmonary infection during treatment of the disease, and drug (cyclophosphamide)-related alveolar damage and interstitial fibrosis that may have led to complicated CT findings to interpret. However, recent improvement of cytotoxic and corticosteroid therapy has mitigated the frequency of recurrence of WG and complication; therefore, in follow-up CT studies, we identified recurrence of the disease in only 3 (15%) of 20 patients who had follow-up CT findings of increased extent of parenchymal or airway lesions. Recurrent WG may appear as multiple cavitary nodules, multifocal or diffuse ground-glass opacity or consolidation, and multifocal or long-segment airway stenosis with or without attendant atelectasis [3]. The recurrent pattern has been reported to be different from the initial pattern at presentation [3, 24]. In our study, we saw both the same and different patterns of recurrence from those at initial presentation (Figs. 2, 6).

Our study was limited by a retrospective review of cases from a large number of institutions. Although thinsection and conventional CT images were available, the parameters of CT scanning were not unified. We got lung biopsy specimens only from 13 patients; therefore, all radiographic findings might not have truly represented those of WG. Some of the findings might have been due to nonspecific inflammation, infection, or any other disease processes. The follow-up period of our study was somewhat short. Long-term follow-up results should be addressed.

Conclusion In conclusion, CT findings of WG, although multiple and variable, consist mainly of subpleural or peribronchovascular nodules or masses in both lungs and bronchial wall thickening in the segmental or subsegmental bronchi. The nodules and masses are seen in 90% and bronchial wall thickening in 73% of patients with the disease. The nodules and masses usually show extensive central low attenuation with or without surrounding enhancing area on enhanced scans. Parenchymal consolidation, ground-glass opacity, and large airway diseases may also be accompanied in one-third or one-fourth of patients. Parenchymal and airway lesions improve with treatment in most of patients.

References 1. Wegener F (1990) The histopathological definition of Wegener’s granulomatosis. Acta Pathol Microbiol Immunol Scand 98:13–14 2. Leavitt RY, Fauci AS (1986) Pulmonary vasculitis. Am Rev Respir Dis 134:149–166 3. Aberle DR, Gamsu G, Lynch D (1990) Thoracic manifestations of Wegener granulomatosis: diagnosis and course. Radiology 174:703–709 4. Cordier JF, Valeyre D, Guillevin L, Loire R, Brechot JM (1990) Pulmonary Wegener’s granulomatosis: clinical and imaging study of 77 cases. Chest 97:906–912 5. Foo SS, Weisbrod GL, Herman SJ, Chamberlain DW (1990) Wegener granulomatosis presenting on CT with atypical bronchovasocentric distribution. J Comput Assist Tomogr 14:1004–1006 6. Kuhlman JE, Hruban RH, Fishman EK (1991) Wegener granulomatosis: CT features of parenchymal lung disease. J Comput Assist Tomogr 15:948–952

7. Maskell FG, Lockwood CM, Flower CDR (1993) Computed tomography of the lung in Wegener’s granulomatosis. Clin Radiol 48:377–380 8. Wadsworth DT, Siegel MJ, Day DL (1994) Wegener’s granulomatosis in children: chest radiographic manifestations. Am J Roentgenol 163:901–904 9. Daum TE, Specks U, Colby TV et al. (1995) Tracheobronchial involvement of Wegener’s granulomatosis. Am J Respir Crit Care Med 151:522–526 10. Frazier AA, Rosado-de-Christenson ML, Galvin JR, Fleming MV (1998) Pulmonary angiitis and granulomatosis: radiologic–pathologic correlation. Radiographics 18:687–710 11. Reuter M, Schnabel A, Wesner F et al. (1998) Pulmonary Wegener’s granulomatosis: correlation between highresolution CT findings and clinical scoring of disease activity. Chest 114:500–506 12. Papiris SA, Manoussakis MN, Drosos AA, Kontogiannis D, Constantopoulos SH, Moutsopoulos HM (1992) Imaging of thoracic Wegener’s granulomatosis: the computed tomographic appearance. Am J Med 93:529–536

13. Attali P, Ben Romdhane H, Valeyre D, Guillevin L, Brauner MW (1998) Pulmonary Wegener’s granulomatosis: changes at follow-up CT. Eur Radiol 8:1009–1113 14. Leavitt RY, Fauci AS, Bloch DA et al. (1990) The American College of Rheumatology 1990 criteria for the classification of Wegener’s granulomatosis. Arthritis Rheum 33:1101–1107 15. Jennette CJ, Falk RJ, Andrassi K et al. (1994) Nomenclature of systemic vasculitides: proposal of an international consensus conference. Arthritis Rheum 37:187–192 16. Naidich DP, Webb WR, Müller NL, Krinsky GA, Zerhouni EA, Siegelman SS (1999) Computed tomography and magnetic resonance of the thorax, 3rd edn. Lippincott-Raven, Philadelphia, pp 175–181, 193–194

51

17. Bankier AA, Fleischmann D, Mallek R et al. (1996) Bronchial wall thickness: appropriate window settings for thin-section CT and radiologic– anatomic correlation. Radiology 199:831–836 18. Travis WD, Hoffman GS, Leavitt RY, Pass HI, Fauci AS (1995) Surgical pathology of the lung in Wegener’s granulomatosis: review of 87 open lung biopsies from 67 patients. Am J Surg Pathol 19:545–552

19. Primack SL, Hartman TE, Lee KS, Müller NL (1994) Pulmonary nodules and the CT halo sign. Radiology 190:513–515 20. Kim Y, Lee KS, Jung K-J, Han J, Kim TS, Suh JS (1999) Halo sign on high resolution CT: findings in spectrum of pulmonary diseases with pathologic correlation. J Comput Assist Tomogr 23:622–626 21. McNamara AE, Müller NL, Okazawa M, Arntorp J, Wiggs BR, Pare PD (1992) Airway narrowing in excised canine lungs measured by high-resolution computed tomography. J Appl Physiol 73:307–316

22. Webb WR, Gamsu G, Wall SD, Cann CE, Proctor E (1984) CT of a bronchial phantom. Factors affecting appearance and size measurements. Invest Radiol 19:394–398 23. Choi YH, Im J-G, Han BK, Kim JH, Lee KY, Myoung NH (2000) Thoracic manifestation of Churg-Strauss syndrome: radiologic and clinical findings. Chest 117:117–124 24. Dugowson CE, Aitken ML (1991) Unusual presentation of recurrent Wegener’s granulomatosis. Chest 99:781–784