| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Department of Radiology, Hammersmith Hospital, London, England (S.J.C.); Department of Radiology (M.B.R., N.P.H., D.M.H.) and Interstitial Lung Disease Unit (A.U.W.), Royal Brompton Hospital, Sydney St, London SW3 6NP, England; Department of Radiology, Green Lane Hospital, Auckland, New Zealand (D.G.M.); and Department of Radiology, Vancouver General Hospital, Vancouver, British Columbia, Canada (N.L.M., J.C.). Received April 10, 2001; revision requested May 29; revision received August 31; accepted October 10. Address correspondence to D.M.H. (e-mail: d.hansell@rbh.nthames.nhs.uk).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Thin-section CT scans of 105 patients with obstructive pulmonary disease (asthma, n = 35; centrilobular emphysema, n = 30; panlobular emphysema, n = 21; and obliterative bronchiolitis, n = 19) and 33 healthy subjects were assessed independently by two observers. The most likely diagnosis and a confidence rating were assigned. Individual thin-section CT features were recorded. Accuracy, sensitivity, specificity, negative predictive value, and positive predictive value for first-choice diagnoses were calculated. The prevalence of CT features between pairs of conditions was compared with the 
2 or Fisher exact test as appropriate.
RESULTS: A correct first-choice diagnosis was made in 199 of 276 (72%) observations. A correct first-choice diagnosis was made in 35 of 38 (92%) observations in patients with obliterative bronchiolitis, in 53 of 60 (88%) observations in patients with centrilobular emphysema, in 53 of 66 (80%) observations in healthy subjects, in 37 of 70 (53%) observations in patients with asthma, and in 20 of 42 (48%) observations in patients with panlobular emphysema. The major sources of diagnostic inaccuracy were differentiation between panlobular and centrilobular emphysema, asthma and normality, and asthma and obliterative bronchiolitis. There were significant increases in prevalence of (a) bronchial wall thickening and vascular attenuation in patients with asthma when compared with healthy subjects and (b) vascular attenuation and decreased attenuation in patients with obliterative bronchiolitis when compared with patients with asthma (P < .001).
CONCLUSION: CT helps to distinguish diseases that cause airflow obstruction. Thin-section CT is particularly accurate in the identification of obliterative bronchiolitis.
© RSNA, 2002
Index terms: Alpha-1-antitrypsin deficiency, 60.7511, 60.7512 • Asthma, 68.754 • Bronchiolitis obliterans, 60.219 • Emphysema, pulmonary, 60.7511, 60.7512
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Thin-section computed tomographic (CT) features of asthma (2,3), centrilobular emphysema (4,5), panlobular emphysema (6,7), and obliterative bronchiolitis (8) have all been described. The superior sensitivity of CT compared with that of chest radiography in the detection of many obstructive pulmonary diseases is undisputed (5,9,10). However, the diagnostic accuracy of thin-section CT has not been evaluated in patients with diseases characterized by chronic airflow obstruction.
The aims of our study were to use thin-section CT to distinguish causes of obstructive pulmonary disease, to determine which distinctions give rise to diagnostic imprecision, and to identify the most useful CT features.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Asthma
Thirty-five consecutive patients from two centers (11 males, 24 females; median age, 37 years; age range, 17–67 years) underwent thin-section CT as part of clinical assessment. All patients had reversible airflow obstruction at spirometry, were nonsmokers or had a smoking history of less than 5 pack-years, and fulfilled American Thoracic Society criteria for the diagnosis of asthma (11).
Centrilobular Emphysema
Thirty consecutive patients with centrilobular emphysema from two centers over a 3-year time period (13 men, 17 women; median age, 60 years, age range, 37–75 years) underwent thin-section CT as part of an investigation for obstructive pulmonary disease or an evaluation for lung volume reduction surgery. The diagnosis was based on histopathologic findings (n = 13) or clinical criteria (n = 17). All patients had a greater than 20 pack-year smoking history, a decreased carbon monoxide diffusing capacity of less than 70% of predicted capacity, a history of irreversible airflow obstruction at pulmonary function testing, and normal 
1-antitrypsin levels.
Panlobular Emphysema
Twenty-one consecutive patients from two centers (16 men, five women; median age, 49 years; age range, 34–66 years) were identified as having panlobular emphysema according to a search of clinical databases. All had 1-antitrypsin levels lower than 80 mg/dL [0.8 g/L] (normal range, 150–350 mg/dL [1.5–3.5 g/L]) and a homozygous phenotype of PiZZ according to American Thoracic Society criteria (12).
Obliterative Bronchiolitis
Nineteen consecutive patients from two centers (three men, 16 women; median age, 51 years; age range, 19–75 years) underwent CT as part of clinical assessment. Inclusion criteria were (a) unexplained fixed airflow obstruction, (b) flow-volume loop compatible with small-airways disease, and (c) ratio of residual volume to total lung capacity of 115% or more of the predicted value (13). Exclusion criteria were modified from those used in the study of Turton et al (14): (a) smoking history (5 or more pack-years), (b) history of asthma, (c) purulent sputum production of more than 6 months’ duration, or (d) features of interstitial lung disease or overt emphysema at CT.
Healthy Subjects
Thirty-three healthy subjects (21 men, 12 women; median age, 23 years; age range, 18–63 years) had been recruited prospectively as part of two studies that required thin-section CT (local ethics committee approval was granted for these studies). Informed consent was obtained from all subjects. All were lifelong nonsmokers with no respiratory symptoms (as determined by questionnaire). Nineteen subjects underwent pulmonary function tests as part of the study protocol (all results were normal). Fourteen patients were from another study in which pulmonary function tests were not performed within 3 months of CT. To the best of our knowledge, no patient had a respiratory infection at the time of scanning.
CT Scanning
CT scans were obtained with an electron-beam CT scanner (Imatron, San Francisco, Calif) or a HiSpeed Advantage CT scanner (GE Medical Systems, Milwaukee, Wis). In 128 individuals (95 patients and 33 subjects), inspiratory 1-mm or 1.5-mm sections were obtained, and in 10 patients, 3-mm sections were obtained. End-expiratory images were available for 111 of 138 individuals. In 105 patients and 14 subjects, sections were obtained at 10-mm intervals, and in 19 subjects, sections were obtained every 30 mm. In five patients from each disease group (ie, asthma, centrilobular emphysema, panlobular emphysema, and obliterative bronchiolitis), sections were obtained every 30 mm. Images were viewed at settings appropriate for lung parenchyma (window width, 1,500 HU; window level, -700 to -550 HU).
Image Evaluation
Two observers (each with more than 10 years of experience in thin-section CT interpretation) scored the CT scans independently, in random order, and on separate occasions without knowledge of clinical data. The observers were initially asked to record the most likely diagnosis, one other differential diagnosis, and their confidence (ie, confident vs uncertain) in making the first-choice diagnosis on the basis of their own experience with thin-section CT. Subsequently, in a separate scoring session and on a separate score sheet, they were asked to record the individual thin-section CT features in each case, without knowledge of their initial diagnosis.
To identify any differences in the prevalence of individual CT features between specific conditions, the observers recorded individual thin-section CT features at three levels: (a) the origin of the great vessels, (b) the carina, and (c) the right inferior pulmonary vein. At each of these levels, the observers recorded (a) bronchial dilatation by using the external diameter of the bronchus (0, no dilatation; 1, dilatation less than two times the diameter of the homologous pulmonary artery; 2, dilatation two to three times the diameter of the homologous pulmonary artery; and 3, dilatation more than three times the diameter of the homologous pulmonary artery), (b) bronchial wall thickening (0, none; 1, wall thickening less than 50% of the diameter of the homologous pulmonary artery; 2, wall thickening between 50% and 100% of the diameter of the homologous pulmonary artery; and 3, wall thickening greater than 100% of the diameter of the homologous pulmonary artery), (c) parenchymal destruction (0, none; 1, mild; 2, moderate; and 3, severe), (d) vascular distortion (0, none; 1, mild; 2, moderate; and 3, severe), (e) vascular attenuation (ie, thinning of vessels) (0, none; 1, mild; 2, moderate; and 3, severe), (f) percentage of decreased parenchymal attenuation as part of a mosaic attenuation pattern (0, none; 1, <25%; 2, 25%–50%; and 3, >50%), (g) air trapping on expiratory images (0, none; 1, mild; 2, moderate; and 3, severe), and (h) "long lines" (ie, adjoining visible interlobular septa, not linear atelectasis) (15) (0, none; 1, mild; 2, moderate; and 3, severe). For each of these features (except the edges of decreased attenuation), the overall distribution (in upper, middle, and lower zones) was recorded.
Statistical Analysis
The diagnostic observations for the first-choice diagnoses of the two observers were added (n = 276). Accuracy, sensitivity, specificity, negative predictive value, and positive predictive value for the first-choice CT diagnoses were calculated. Diagnostic accuracy for confident and uncertain ratings was assessed. Diagnostic accuracy according to the severity of airflow obstruction was also assessed. Agreement between observers was assessed by using the 
statistic. Agreement between observers for individual thin-section CT features was assessed by using the weighted (w) statistic. values have been defined as the following: less than 0.2 = poor, 0.21–0.40 = fair, 0.41–0.60 = moderate, 0.61–0.80 = good, and 0.81–1.00 = very good (16). The prevalence of thin-section CT features between pairs of conditions was compared with the 2 or Fisher exact test as appropriate. The Wilcoxon rank sum test was used to compare the median extents of individual CT features. Logistic regression was used to define independent predictors of presence of thin-section CT features between conditions. The mean of observations for individual thin-section CT features was obtained to assess the features that allowed differentiation between selected pairs of conditions (ie, asthma vs normality, obliterative bronchiolitis vs panlobular emphysema, obliterative bronchiolitis vs asthma, and panlobular emphysema vs centrilobular emphysema).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
= 0.68).
|
|
W = 0.34).
|
|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The high accuracy of CT in the diagnosis of obliterative bronchiolitis and centrilobular emphysema and its specificity in ruling out these entities is of particular importance because CT is currently used most often to assess these disorders. The utility of thin-section CT in the detection of centrilobular emphysema (5,20) and obliterative bronchiolitis (10) is well recognized. Clinical applications of CT are the detection of obliterative bronchiolitis after lung transplantation (10,21,22) and the identification of areas of centrilobular emphysema before lung volume reduction surgery (23,24). Asthma is usually diagnosed clinically, but CT is sometimes helpful in providing an alternative explanation, such as obliterative bronchiolitis.
The way in which our study was conducted differs from clinical practice in one important respect: Almost half of the patients with centrilobular emphysema had histopathologic confirmation of diagnosis. This apparent advantage overlooks the fact that CT, rather than histologic analysis, is widely used for the identification of centrilobular emphysema. Statements about the use of CT in the diagnosis of centrilobular emphysema in the absence of histopathologic confirmation may therefore be ambiguous. This problem is unavoidable if cases of mild or moderate emphysema are to be included in such a study, since histopathologic confirmation is infrequently obtained. Thus, we believe that the main value of our results is the demonstration that CT may be used to discriminate between causes of airflow obstruction other than centrilobular emphysema.
Similar accuracies for different functional severities of obstruction were seen. This may seem surprising, since making the distinction between centrilobular and panlobular emphysema at CT would be expected to be more difficult with increasingly extensive disease. However, both observers had similar accuracies for patients with severe obstruction, possibly reflecting reliance on ancillary CT findings. In patients with normal pulmonary function (FEV1 values greater than 80% of predicted values), patients with treated or mild asthma were included to provide a spread of disease severity. In these cases, it is unlikely that thin-section CT features such as bronchial wall thickening and areas of decreased attenuation would have been either present or of sufficient severity to allow differentiation from healthy subjects. Undoubtedly, if these patients were excluded, our results would have been better, since a major source of diagnostic confusion was encountered between patients with asthma and healthy subjects.
The accuracy of diagnosis for individual conditions was good (88%–93%). Thin-section CT was specific for all individual conditions, but low sensitivities were seen for asthma (53%) and panlobular emphysema (48%). Panlobular emphysema was incorrectly diagnosed as centrilobular emphysema in almost half of observations, but centrilobular emphysema was far less frequently diagnosed as panlobular emphysema. This may be partly explained by the fact that many of the patients who had panlobular emphysema as a result of 1-antitrypsin deficiency were also smokers, and so they may have had conspicuous centrilobular emphysema in the upper lobes. Furthermore, in extensive and advanced disease it may be hard to discriminate between the two forms of emphysema. However, thin-section CT was surprisingly accurate, sensitive, and specific for the diagnosis of obliterative bronchiolitis, with only three false-negative observations. False-positive diagnoses for obliterative bronchiolitis were mostly attributable to asthma. This is not surprising, given that the peripheral airways are also the major site of obstruction in asthma (25,26), with relatively less inflammatory involvement of the central airways (27). Overall, observer agreement for individual thin-section CT features was good, with the exception of bronchial wall thickening, and was similar to that reported in previous articles (28,29).
There were significant differences in the presence and severity of CT features between certain conditions. Despite low observer agreement for bronchial wall thickening, there were significant differences in the presence and severity of this feature in patients with asthma when compared with healthy subjects, which confirms results of previous studies (2,3). There was also a significantly higher frequency and severity of vascular attenuation and decreased parenchymal attenuation in patients with asthma when compared with healthy subjects. In contrast to results of previous studies (2,3), we did not observe statistically significant differences in extent or presence of bronchial dilatation in patients with asthma when compared with healthy subjects, which may reflect less severe disease in our study population. Patients with asthma were most often falsely diagnosed as being healthy, and vice versa. A likely explanation for this finding may be that if patients with asthma undergo scanning between asthma attacks, reversible features such as bronchial wall thickening may be absent (30). The differentiating features of increased bronchial abnormalities, decreased attenuation, and vascular attenuation between patients with asthma and obliterative bronchiolitis may also be related to the fact that patients with asthma generally do not undergo scanning during an acute exacerbation, when similar (but less extensive) thin-section CT features to those of obliterative bronchiolitis may be expected.
There were surprisingly few misdiagnoses between patients with obliterative bronchiolitis and those with panlobular emphysema in our study. It may be difficult for even experienced observers to differentiate between severe obliterative bronchiolitis and panlobular emphysema because both conditions are characterized by bronchial wall thickening and, particularly in advanced stages, by generalized decreased attenuation of the lung parenchyma and bronchial dilatation (1). A possible explanation is that the observers were able to assess the entire lung and identify areas of mosaic attenuation in patients with less severe obliterative bronchiolitis. There was also a significantly higher frequency and greater extent of parenchymal destruction in patients with panlobular emphysema in comparison to patients with obliterative bronchiolitis, which may have aided differentiation, despite the difficulty in identifying parenchymal destruction when the whole secondary pulmonary lobule is involved (as in panlobular emphysema).
The increased prevalence and extent of long lines in patients with panlobular emphysema (in comparison to patients with obliterative bronchiolitis), to the best of our knowledge, is a new observation. In the absence of pathologic correlation, the nature of the lines is a matter of speculation, but we believe that these lines may represent linked interlobular septa that were made more conspicuous by extensive adjacent parenchymal destruction.
The distribution of thin-section CT features was helpful only in discriminating centrilobular emphysema from panlobular emphysema. Centrilobular emphysema is usually considered to involve predominantly the upper zones, while panlobular emphysema is a predominantly lower-zone disease. We observed some overlap in distribution between the two conditions, and a considerable proportion of patients had widespread disease with no particular zonal predominance. When a zonal predominance of airway abnormalities and decreased attenuation were observed in patients with obliterative bronchiolitis, a lower-zone distribution was seen (similar to panlobular emphysema), another potential reason for misdiagnoses between the two entities in advanced cases.
There were drawbacks to our study. First, the observers had no more than five possible diagnoses to choose from. However, this reflects clinical practice, as the common causes of obstructive pulmonary disease were included apart from bronchiectasis. Bronchiectasis, per se, is likely to be an underdiagnosed cause of airflow obstruction, and 29% of patients in primary care with a clinical diagnosis of chronic obstructive pulmonary disease had CT evidence of bronchiectasis without emphysema in another study (31). However, the clinical presentation and CT features of bronchiectasis rarely result in diagnostic difficulty.
Second, there may be selection bias, since our patients were selected from tertiary referral centers, which may account for the relatively large groups of patients with less common causes of obstruction (ie, panlobular emphysema due to 1-antitrypsin deficiency and obliterative bronchiolitis).
Another possible confounder of our observations is that different conditions may coexist in the same patient. For example, a patient with 1-antitrypsin deficiency may also be a smoker and have both panlobular and centrilobular emphysema (32,33). In patients in whom histopathologic specimens were obtained (during lung volume reduction surgery), sampling error may have been a factor, since upper-zone lung parenchyma are usually excised. Furthermore, histopathologic differentiation between centrilobular or panlobular emphysema may be problematic in some cases (33).
Our study has shown that thin-section CT is accurate in the differentiation of different causes of obstructive lung disease, particularly in the diagnosis of obliterative bronchiolitis. There were significant differences in the prevalence and extents of some thin-section CT features between conditions; however, with the exception of centrilobular and panlobular emphysema, the distribution of disease was less helpful.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, S.J.C., D.M.H.; study concepts, A.U.W., D.M.H., N.L.M.; study design, S.J.C., D.M.H., N.L.M.; literature research, S.J.C., N.P.H.; clinical studies, S.J.C., N.P.H., J.C.; data acquisition, S.J.C., N.P.H., J.C.; data analysis/interpretation, N.P.H., A.U.W., M.B.R., D.G.M., J.C.; statistical analysis, A.U.W., S.J.C.; manuscript preparation, S.J.C.; manuscript definition of intellectual content, all authors; manuscript editing, S.J.C., D.M.H., A.U.W.; manuscript revision/review, D.M.H., N.L.M., A.U.W.; manuscript final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. A. Bankier, S. Mehrain, D. Kienzl, M. Weber, M. Estenne, and P. A. Gevenois Regional Heterogeneity of Air Trapping at Expiratory Thin-Section CT of Patients with Bronchiolitis: Potential Implications for Dose Reduction and CT Protocol Planning Radiology, June 1, 2008; 247(3): 862 - 870. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Hansell, A. A. Bankier, H. MacMahon, T. C. McLoud, N. L. Muller, and J. Remy Fleischner Society: Glossary of Terms for Thoracic Imaging Radiology, March 1, 2008; 246(3): 697 - 722. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T Stone, J H Reynolds, and H J Williams Imaging of large and small airway diseases Imaging, September 1, 2006; 18(3): 139 - 150. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. S. Kim, N. Tanaka, J. D. Newell, M. A. DeGroote, K. Fulton, G. Huitt, and D. A. Lynch Nontuberculous Mycobacterial Infection: CT Scan Findings, Genotype, and Treatment Responsiveness Chest, December 1, 2005; 128(6): 3863 - 3869. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. A. Aziz, A. U. Wells, S. R. Desai, S. M. Ellis, A. E. Walker, S. MacDonald, and D. M. Hansell Functional Impairment in Emphysema: Contribution of Airway Abnormalities and Distribution of Parenchymal Disease Am. J. Roentgenol., December 1, 2005; 185(6): 1509 - 1515. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Cottin, H. Nunes, P-Y. Brillet, P. Delaval, G. Devouassoux, I. Tillie-Leblond, D. Israel-Biet, I. Court-Fortune, D. Valeyre, J-F. Cordier, et al. Combined pulmonary fibrosis and emphysema: a distinct underrecognised entity Eur. Respir. J., October 1, 2005; 26(4): 586 - 593. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. J. Pipavath, D. A. Lynch, C. Cool, K. K. Brown, and J. D. Newell Radiologic and Pathologic Features of Bronchiolitis Am. J. Roentgenol., August 1, 2005; 185(2): 354 - 363. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. S. Patel, I. Vlahos, T. M. A. Wilkinson, S. J. Lloyd-Owen, G. C. Donaldson, M. Wilks, R. H. Reznek, and J. A. Wedzicha Bronchiectasis, Exacerbation Indices, and Inflammation in Chronic Obstructive Pulmonary Disease Am. J. Respir. Crit. Care Med., August 15, 2004; 170(4): 400 - 407. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. C. Sciurba Physiologic Similarities and Differences Between COPD and Asthma Chest, August 1, 2004; 126(2_suppl_1): 117S - 124S. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
