| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the Department of Radiology (S.J.C., M.B.R., F.C., R.E.S., D.M.H.) and the Interstitial Lung Disease Unit (A.U.W.), Royal Brompton Hospital, Sydney St, London SW3 6NP, England; and the Department of Respiratory Medicine, Sir Charles Gairdner Hospital, Nedlands, Australia (A.W.M.). Received August 1, 2000; revision requested September 13; final revision received February 5, 2001; accepted February 12. Address correspondence to D.M.H. (e-mail: d.hansell@rbh.nthames.nhs.uk).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: The extent of pleural thickening and plaques was assessed in 50 patients by using the following: (a) a radiographic score based on the International Labour Office system, (b) a subjective simple computed tomographic (CT) score, (c) a subjective comprehensive CT score, (d) an objective nonautomated method, and (e) an objective computer-aided semiautomated method.
RESULTS: Similar correlations between the extent of diffuse pleural thickening and forced vital capacity were seen for each system (objective CT, r = -0.72, P < .001; simple CT, r = -0.69, P < .001; radiographic, r = -0.67, P < .001; comprehensive CT, r = -0.66, P < .001). Comparable correlations were observed for total lung capacity. After controlling for extent of diffuse pleural thickening, pleural plaque scores were functionally irrelevant.
CONCLUSION: Comparable functional-morphologic correlations were achieved by using different CT and radiographic scoring systems for pleural disease. A subjective simple CT system had the advantages of ease of application and potential to aid in the accurate assessment of the lung parenchyma, which may be important in individuals exposed to asbestos.
Index terms: Asbestos, 66.773 • Pleura, CT, 66.12118 • Pleura, diseases, 66.76, 66.773
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
An understanding of the relationships between structure and function in pleural disease is important for two reasons. First, it would aid in the quantification of the proportion of pulmonary functional impairment due to pleural disease in patients with concurrent parenchymal and pleural involvement, for example in connective tissue disease and asbestosis. Second, compensation is often sought by individuals with asbestos-related benign pleural disease: Correlations between structure and function may be used to determine whether functional impairment is ascribable to pleural disease in individual patients.
The aim of the study was to identify a system for the quantification of pleural thickening, with an acceptable level of interobserver variation and good functional correlation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Spirometric pulmonary volumes (forced expiratory volume in 1 second [FEV1], forced vital capacity [FVC], and the ratio FEV1/FVC) were obtained by using spirometry (Masterlab; E. Jaeger, Leicestershire, England). Plethysmographic pulmonary volumes (residual volume, total lung capacity, and their ratio) were obtained by using body plethysmography (Masterlab; E. Jaeger). Diffusing capacity measurements (single-breath carbon monoxide diffusing capacity corrected for hemoglobin concentration and adjusted for alveolar volume KCO) were acquired with gas transfer equipment (Masterlab; E. Jaeger). Pulmonary function indices were expressed as a percentage of the predicted values of the patient’s age, sex, and height (9).
All CT scans were obtained with an electron-beam scanner (Imatron, San Francisco, Calif). Thirty-four patients underwent thin-section CT scanning (1.5-mm-thick sections at 10-mm intervals). Fourteen patients underwent scanning to obtain 6-mm-contiguous sections imaged on lung windows (window center, -550 HU; window width, 1,500 HU) and mediastinal windows (window center, 0 HU; window width, 450 HU). Two patients underwent both contiguous and thin-section CT. Thin-section CT images were reconstructed with a high-spatial-frequency (bone) algorithm and photographed with relatively wide window settings to allow evaluation of both pleura and parenchyma (window center, -550 HU; window width, 1,500 HU).
Two observers (D.M.H., M.B.R.) scored the chest radiographs and CT scans with the radiographic, simple CT, and comprehensive CT systems independently, in random order, and on different occasions at least 2 weeks apart. The two objective methods of assessing pleural disease were independently performed by two different observers (S.J.C., R.E.S.; who had not scored either the CT images or chest radiographs), in random order, and on different occasions at least 2 weeks apart. The observers had no knowledge of the clinical findings or pulmonary function data.
Five Methods for Quantifying Extent of Pleural Disease
Radiographic classification.—The radiographic system used was based on the current International Labour Office classification of pleural disease of the pneumoconioses (10,11). Each hemithorax was evaluated separately for (a) the presence of costophrenic angle obliteration; (b) the maximum width of both noncalcified diffuse pleural thickening and circumscribed pleural plaques (0, <5 mm, 5–10 mm, >10 mm); (c) the extent in terms of the maximum length of chest wall involvement (0%, <25% of lateral chest wall involved, 25%–50% of lateral chest wall involved, >50% of lateral chest wall involved); and (d) the extent of pleural calcification of the chest wall, diaphragm, and other sites, including the mediastinal pleura and pericardial surfaces, graded on a scale of 1 to 3: grade 1 equals an area of pleural calcification with a greatest diameter of up to 20 mm, or a number of areas the sum of whose greatest diameters does not exceed 20 mm; grade 2 equals area (or number of areas) with a greatest diameter of 20–100 mm; grade 3 equals area (or number of areas) with a greatest diameter greater than 100 mm.
Simple CT system.—The simple CT system was developed for use in clinical practice. Five levels were evaluated: the origin of the great vessels from the aortic arch, the most cranial section containing the descending thoracic aorta, the carina, the level of the left inferior pulmonary vein and the section halfway between the level of the left inferior pulmonary vein, and the extreme caudal extent of the left costophrenic angle. At each level, the mean thickness of diffuse pleural disease, the percentage circumference of the thorax involved (combining the left and right hemithoraces and excluding the mediastinal surface), and the presence and number of areas of rounded atelectasis were recorded.
The thorax was divided into four quadrants at each level, and the number of quadrants containing pleural plaques was determined. Pleural plaques were defined as areas of circumscribed pleural thickening with well-demarcated or defined borders. Diffuse pleural thickening was defined as pleural thickening with tapering margins.
Comprehensive CT system.—The comprehensive CT system was based on the scoring systems of Aberle et al (4) and Jarad et al (5). The thorax was divided into thirds on the basis of the number of CT sections from the lung apex to the extreme costophrenic angle; all the sections in each lung third, for both the right and left hemithoraces, were assessed. The maximum, minimum, and mean thickness of diffuse pleural thickening and pleural plaques was assessed in each third. The maximum circumference of each hemithorax individually involved by diffuse pleural thickening and pleural plaques was recorded on the section in each third, with the greatest circumference of hemithorax affected by pleural disease (grade 1, <25%; grade 2, 25%–50%; grade 3, 51%–75%; grade 4 >75%).
The presence of pleural calcification (if present, localized vs extensive), rounded atelectasis (grade 1, crow’s feet; grade 2, rounded atelectasis < 4 cm in diameter; grade 3, rounded atelectasis > 4 cm in diameter), and the extent of fissural thickening (grade 1, <50%; grade 2, 50%–99%, grade 3, 100%) was recorded. Crow’s feet or parenchymal bands were defined as nontapering linear structures 2–5 cm in length, which contacted the pleural surface (12,13). Pleural plaques and diffuse pleural thickening were defined as in the simple CT system. In addition, the presence or absence of interstitial fibrosis was scored.
Cutout objective CT system.—The cutout objective CT method of quantifying pleural disease consisted of tracing the total outline of pleural thickening and/or plaques and aerated lung for each hemithorax onto tracing paper at all five levels used in the simple CT system. The tracing paper outline was then cut out and weighed by using an electronic balance (AJ50; Mettler Toledo AG, Greifensee, Switzerland). The tracing paper contour of diffuse pleural thickening and/or pleural plaques at each level was cut away, and the remainder weighed. The difference, representing the cross-sectional area of pleural disease, was calculated. For both the right and left hemithoraces, the weight of diffuse pleural thickening and/or pleural plaques was expressed as a percentage of the total weight of pleural disease and aerated lung. The score was then derived from a mean of the total values for each hemithorax.
A similar technique using cutout tracing paper has been previously used (14) to analyze the relationships between disease extent and function indices in idiopathic pulmonary fibrosis, but this method, to our knowledge, has not been previously used to evaluate pleural disease.
Image analysis objective CT system.—This method was similar to the semiautomated techniques described by al Jarad et al (7) and Schwartz et al (8). Fully automated segmentation of the main anatomic structures on transverse CT images was achieved with a filtering operator based on thresholding, region growing, and mathematic morphology (15). The technique handles partial volume effect and identifies the pixels representing the lung parenchyma, allowing automatic delineation of the internal contour of pleural thickening. The accurate delineation of diffuse pleural thickening from the adjacent soft tissues of the chest wall was not possible with threshold techniques. Therefore, a semiautomated technique requiring user interaction was designed to achieve this segmentation.
Two observers (S.J.C., R.E.S.) viewing the CT scans on a workstation defined an approximate polygonal outline of the external contour of the pleura (Fig 2). An algorithm then deformed the contour, to fit the maximum attenuation gradient found in the neighborhood of the original outline. As a result, at points where the ribs were visible, the contour was automatically fitted to follow the interface between bone and pleura, where a high-attenuation gradient can be observed.
|
|
Statistical Analysis
Data are expressed as the mean ± SD or as a median with range as appropriate. A P value of less than .05 indicated a statistically significant difference. Interobserver variation in semicategorical variables was quantified by using a weighted kappa (
w) coefficient. Variation in quantitative data were single-determination SD (16). Univariate relationships were tested with the Spearman rank correlation. Independent relationships between the CT extent of pleural disease and pulmonary function indices were examined by using stepwise forward regression; non-normally distributed variables were logarithmically transformed prior to analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
w > 0.60) was shown for most observations in all systems. The radiographic scoring system was characterized by moderate interobserver agreement for the width (w = 0.44) and extent (w = 0.45) of pleural plaques and pleural calcification (w = 0.53). The simple CT scoring system showed the least interobserver variation (an example of the observers’ scoring is shown in Figure 3). The observer variation for the cutout objective CT method was 1.16% single-determination SD (score range, 0%–28.0%) for diffuse pleural thickening and 0.48% single-determination SD (score range, 0%–7.6%) for pleural plaques; for the image analysis objective CT method, a 1.08% single-determination SD (score range, 0%–29.8%) for diffuse pleural thickening and 0.30% single-determination SD (score range, 0%–4.6%) for pleural plaques were observed.
|
|
Relationships between Pleural Disease Extent and Pulmonary Function
There were strong relationships between indices of diffuse pleural disease and reductions in FVC and total lung capacity (Spearman r values, -0.66 to -0.72; P < .001 for FVC) for all scoring systems, with little difference between systems. The parameters for each system that correlated most strongly with function indices are shown in Table 3. Increases in KCO (denoting extrapulmonary restriction) were less strongly associated with increasingly extensive diffuse pleural thickening.
|
|
|
|
Comparisons of Scoring Systems for Diffuse Pleural Thickening
To compare scoring systems, the parameter of diffuse pleural thickening most closely linked to functional impairment was used. Each scoring system was evaluated against the image analysis objective CT system as the standard; there was almost perfect correlation with the cutout objective CT method (r = 0.97; P < .001) and good correlation with the simple CT system (r = 0.89; P < .001), the comprehensive CT system (r = 0.87; P < .001), and the radiographic system (r = 0.75; P < .001).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The simple CT score system was clearly the CT system of choice; there was excellent interobserver agreement for both pleural plaques and diffuse pleural thickening, the functional-morphologic correlation was virtually as good as both objective CT methods, and the system was straightforward and not time-consuming. After controlling for extent of diffuse pleural thickening, pleural plaques were functionally silent, confirming the traditionally held view that pleural plaques are of little consequence. We also found that for a given total extent of diffuse pleural thickening, there was no statistically significant difference between unilateral and bilateral disease.
The strength of functional-morphologic correlations is dependent on two factors: the intrinsic strength of an association and the noise of measurement. We have shown that minor differences in observer variation exist between scoring systems; however, this is not a major consideration as the correlations between structure and function differ remarkably little among the five systems. Thus, the choice between scoring systems is governed by other factors, particularly in the ease of application. In principle, in patients with isolated pleural disease, a chest radiograph may be sufficient. However, CT may still be required to identify other disease processes, where the detail of the lung parenchyma is important (eg, in suspected asbestosis).
Interobserver agreement was good for diffuse pleural thickening in all systems; however, chest radiography performed less well than the CT systems, particularly in scoring the extent of pleural plaques. An explanation could be that, while the extent of diffuse pleural thickening is usually straightforward to assess in relation to the lateral chest wall, pleural plaques (often seen en face) are more difficult to quantify in extent and thickness. In addition, it is probably more difficult to separate diffuse pleural thickening from pleural plaques with chest radiography (20). Previous studies of pleural disease are difficult to compare with the current study in terms of interobserver agreement, because coefficients were not stated (4,5,8) or single observers were used (6,21).
All five scoring systems correlated well with indices of pulmonary function. The objective CT methods correlated most strongly with function indices, but these methods are relatively laborious, particularly the cutout objective CT method and, therefore, impractical for clinical or epidemiologic studies. However, the simple CT scoring system correlates remarkably well with functional impairment compared with both objective methods, especially in subjects with less extensive diffuse pleural thickening (Fig 5). A possible explanation is that the objective CT methods could be used to detect and quantify subtle disease, which was functionally unimportant. The present investigation did not demonstrate a significant relationship between pleural calcification and indices of pulmonary function, which is consistent with the findings of the previous study by Broderick et al (22).
The main functional determinant of the simple CT scoring system was the circumference of diffuse pleural thickening rather than the thickness; however, as the circumference involved increases, so does pleural thickness. Previous quantification systems that differentiated diffuse pleural thickening from functionally irrelevant pleural plaques primarily in terms of thickness of pleural disease are, therefore, likely to be less accurate (4). Logically, it might be expected that subjects with bilateral diffuse pleural thickening show significantly decreased lung volumes compared with individuals with unilateral diffuse pleural thickening in that the unaffected side may be able to compensate. Surprisingly, the current study findings showed no statistically significant difference between the two forms of involvement, however, decreased volumes were significantly associated with the costal surface circumference of the thorax involved.
Our results are consistent with those of previous studies that correlated pulmonary functional impairment with the extent of pleural disease at CT. Aberle et al (4) used a graded CT system for severity of diffuse pleural thickening and found that significant inverse correlations with FVC and single-breath diffusing capacity, both function indices of restrictive interstitial lung disease, were demonstrated in subjects with asbestos-related pleuroparenchymal disease. However, patients with isolated pleural disease were not evaluated separately and although correlations were statistically significant, relationships were not as strong as shown in the current study.
A possible explanation may be that diffuse pleural thickening was defined as a thickened sheet of pleura with a transverse extent of 5 cm or more, cephalocaudal extent of at least 8 cm, and thickness greater than 3 mm, excluding cases with diffuse thickening less than 3 mm (which may, nevertheless, be functionally significant) (4). Another graded CT scoring system for pleural disease was used by Hillerdal et al (23), who found that individuals with bilateral pleural thickening demonstrated with CT or chest radiography had statistically significant decreases in lung volumes. The number of patients included in this study was small, and although good correlation was recorded between the degree of right-sided diffuse pleural thickening and an index of pulmonary function (maximal expiratory flow-to–flow at 50% of FVC ratio), relationships between other pulmonary function indices were not as statistically significant, as in the present study.
A study by Jarad et al (5) including patients with asbestosis, emphysema, and asbestos-induced pleural disease showed similar strong independent relationships between the pleural CT score and FEV1, FVC, and total lung capacity, as in those of our study, despite apparent lack of discrimination between diffuse pleural thickening and pleural plaques.
Valkila et al (6) showed that diffuse pleural thickening evaluated with both CT and chest radiographic scoring systems was significantly associated with decreased static and dynamic compliance but not statistically significant decreases in lung volumes. Reasons for the discrepancy between our findings and those of Valkila et al could be the relatively high proportion of patients with confounding chronic obstructive pulmonary disease or asthma assessed, a difference in smoking histories (the current study having a higher proportion of nonsmokers), or differences in disease extent between the two populations in their study.
More recently, Kee et al (21) examined 53 subjects with diffuse pleural thickening at CT and found that the presence of diffuse pleural thickening was associated with a substantially reduced FVC and single-breath carbon monoxide diffusing capacity independent of interstitial fibrosis or pleural plaques. Correlations of function indices with the extent of pleural disease at CT were not performed, which makes comparison with our study difficult.
Semi-objective methods for evaluating pleural disease similar to our image analysis objective CT method have also been used to correlate extent of pleural disease with function indices (7,8). al Jarad et al (7) delineated pleural disease by using a computer-aided method that calculated the cross-sectional area of pleural disease. They found that substantial inverse correlations were observed between the CT score and FEV1, FVC, total lung capacity, functional residual capacity, single-breath carbon monoxide diffusing capacity, and alveolar volume, with similar strong correlations with total lung capacity, as shown in the present study.
Schwartz et al (8) also used a quantitative method for assessing asbestos-related pleural disease with thin-section CT. The volume of pleural thickening in relation to the total chest cavity was calculated by using three-dimensional image reconstruction. The volume of pleural thickening was inversely related to the total lung capacity, although the relationship was not as strong as shown in the present study (r = -0.40; P = .002). This difference may be due to the fact that a large proportion of subjects examined by Schwartz et al had limited pleural thickening, which may have been functionally inconsequential.
It is unlikely that our results were substantially affected by selection bias, borne out by the range of disease included in this study. However, retrospective selection of patients by using CT reports may have potentially identified patients with more severe pleural disease (investigated with CT and pulmonary function tests). Moreover, patients with minor pleural disease may have been overlooked if pleural disease was not reported at the original CT examination. However, we also included cases in which the principal indication for investigation was unrelated to pleural disease, which was reported as an ancillary and generally minor finding at CT.
All the individuals in our study were identified on the basis that pleural disease was shown with CT. This may have provided a spurious advantage to the radiographic scoring system, as the chest radiograph is less specific for the differentiation between pleural thickening and extrapleural fat than CT (11,24–26). A combination of thick- and thin-collimation CT images were evaluated by the observers, but the same images were used for each scoring method.
In individuals exposed to asbestos, the application of a CT scoring system for pleural disease is desirable as a variety of coexisting disease processes may each independently contribute to a deficit in pulmonary function. Diffuse pleural thickening and interstitial fibrosis both contribute to a restrictive pulmonary function deficit (17,18), while small-airway disease and emphysema may result in variable degrees of airflow obstruction (27,28). To deconstruct and analyze the complex relationships in individuals exposed to asbestos, a robust system for quantifying each individual component is required.
Our study has identified a straightforward and reliable CT visual scoring system for the accurate quantification of pleural disease, which may be of particular relevance in individuals exposed to asbestos. Using this system, our results are in broad agreement with those of previous CT studies and studies based on scoring of the chest radiograph in that a reduction of lung volumes and an increase in the transfer coefficient KCO correlate strongly with the extent of diffuse pleural thickening. While pleural plaques were not independently associated with pulmonary function deficit.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, S.J.C., D.M.H.; study concepts and design, A.U.W., A.W.M., D.M.H.; literature research, S.J.C.; clinical studies, S.J.C.; data acquisition, S.J.C., F.C., R.E.S., M.B.R.; data analysis/interpretation, M.B.R., S.J.C., F.C., R.E.S., A.U.W.; statistical analysis, S.J.C., A.U.W.; manuscript preparation, S.J.C., F.C.; manuscript definition of intellectual content, S.J.C., R.E.S., M.B.R., A.U.W., A.W.M., D.M.H.; manuscript editing, S.J.C., D.M.H.; manuscript revision/review, A.U.W., S.J.C., A.W.M., M.B.R., D.M.H.; manuscript final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. J. Copley, Y. C. G. Lee, D. M. Hansell, P. Sivakumaran, M. B. Rubens, A. J. Newman Taylor, R. M. Rudd, A. W. Musk, and A. U. Wells Asbestos-induced and Smoking-related Disease: Apportioning Pulmonary Function Deficit by Using Thin-Section CT Radiology, December 1, 2006; 242(1): 258 - 266. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D R Warakaulle and Z C Traill Imaging of pleural disease Imaging, October 1, 2004; 16(1): 10 - 21. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Sette, J. A. Neder, L. E. Nery, J. Kavakama, R. T. Rodrigues, M. Terra-Filho, S. Guimaraes, E. Bagatin, and N. Muller Thin-Section CT Abnormalities and Pulmonary Gas Exchange Impairment in Workers Exposed to Asbestos Radiology, July 1, 2004; 232(1): 66 - 74. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. W. Cugell and D. W. Kamp Asbestos and the Pleura: A Review Chest, March 1, 2004; 125(3): 1103 - 1117. [Full Text] [PDF]
|
|
|
|
|
|
S. J. Copley, A. U. Wells, P. Sivakumaran, M. B. Rubens, Y. C. Gary Lee, S. R. Desai, S. L. S. MacDonald, R. I. Thompson, T. V. Colby, A. G. Nicholson, et al. Asbestosis and Idiopathic Pulmonary Fibrosis: Comparison of Thin-Section CT Features Radiology, December 1, 2003; 229(3): 731 - 736. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. D. Roach, G. J. Davies, R. Attanoos, M. Crane, H. Adams, and S. Phillips Asbestos: When the Dust Settles—An Imaging Review of Asbestos-related Disease RadioGraphics, October 1, 2002; 22(90001): S167 - 184. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
