Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH
QUICK SEARCH:   [advanced]


     

Published online before print November 7, 2006, 10.1148/radiol.2421051167

(Radiology 2006;242:258.)

A more recent version of this article appeared on December 1, 2006
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow All Versions of this Article:
2421051167v1
242/1/258    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Copley, S. J.
Right arrow Articles by Wells, A. U.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Copley, S. J.
Right arrow Articles by Wells, A. U.
© RSNA, 2006

Thoracic Imaging

Asbestos-induced and Smoking-related Disease: Apportioning Pulmonary Function Deficit by Using Thin-Section CT1

Susan J. Copley, MD, MRCP, FRCR, Y. C. Gary Lee, MBChB, PhD, FRACP, David M. Hansell, MD, FRCP, FRCR, Pathmanathan Sivakumaran, MBBS, FRACP, Michael B. Rubens, DMRD, FRCR, Anthony J. Newman Taylor, MSc, FRCP, FMedSci, Robin M. Rudd, MD, FRCP, A. William Musk, MD, MSc, FRACP and Athol U. Wells, MD, FRACP

1 From the Department of Radiology, Hammersmith Hospital, London, England (S.J.C.); Department of Radiology (D.M.H., M.B.R.) and Interstitial Lung Disease Unit (A.J.N.T., A.U.W.), Royal Brompton Hospital, Emmanuel Kaye Building, Manresa Rd, London, SW6 LR6, England; Center for Respiratory Research, University College London, London, England (Y.C.G.L.); Departments of Medicine and Public Health, University of Western Australia, Perth, Australia (A.W.M.); Department of Respiratory Medicine, Middlemore Hospital and University of Auckland, Auckland, New Zealand (P.S.); and Department of Respiratory Medicine, the London Chest Hospital, London, England (R.M.R.). Received July 12, 2005; revision requested September 21; revision received February 16, 2006; accepted March 21; final version accepted April 3. Y.C.G.L. supported by a Wellcome Advanced Fellowship. Address correspondence to A.U.W. (e-mail: a.wells{at}rbh.nthames.nhs.uk).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 
Purpose: To retrospectively correlate the extent of individual diseases seen at thin-section computed tomography (CT) with pulmonary function in an initial group of patients with asbestos-related parenchymal disease (asbestosis) and to test these findings in a subsequent group of patients whose CT scans were retrospectively identified.

Materials and Methods: This retrospective study had Institutional Review Board approval; informed consent was not required. The study included 133 individuals who had been exposed to asbestos. In the initial study group (81 patients; 79 men, two women; median age, 67 years), two observers used a CT scoring system to quantify the extent of pulmonary fibrosis, diffuse pleural thickening, small-airways disease, and emphysema. Multivariate equations were formulated by using independent CT variables to predict changes in total lung capacity (TLC) and carbon monoxide diffusing capacity (DLCO). The validity of these equations was then tested in a subsequent group of patients (52 patients; all men; median age, 60 years).

Results: At thin-section CT, the extent of asbestos-induced pleuropulmonary disease and emphysema correlated significantly with physiologic impairment (P < .001). Combined CT variables predicted 58% and 57% of the variability in TLC and DLCO, respectively, despite considerable variation in the proportion of coexisting pathologic conditions. When predictive equations with CT variables derived from the initial study group were applied to the subsequent study group, predicted TLC ({rho} = 0.75, P < .001) and DLCO ({rho} = 0.64, P < .001) correlated strongly with measured values.

Conclusion: The proposed CT system provides a semiquantitative method for assessing the relative contribution of asbestos-induced pleuropulmonary disease and smoking-related emphysema to functional impairment.

© RSNA, 2006


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 
Cigarette smoking is common among workers exposed to asbestos, and emphysema frequently complicates the interpretation of physiologic impairment in patients with asbestosis (1–3). This poses a problem for objective evaluation and management. Further, compensation entitlements are largely related to damage from asbestosis but not emphysema, and the relative contribution of asbestosis and emphysema to functional impairment is medicolegally contentious (4).

Traditionally, chest radiography has had an important role in the diagnosis and epidemiologic study of nonmalignant asbestos-induced pleuroparenchymal disease (5). In general, thin-section computed tomography (CT) is more sensitive and more specific than chest radiography in demonstrating interstitial lung disease and has been used to quantify disease, especially in patients who have idiopathic pulmonary fibrosis with or without concurrent emphysema (6–8). Although observational and quantitative studies have been performed in workers who have been exposed to asbestos (9), none of these studies, to our knowledge, has investigated the varying proportion of diseases and their contribution to overall functional abnormality.

Thus, the purpose of the present study was to retrospectively correlate the extent of individual diseases seen at thin-section CT with pulmonary function in an initial group of patients with asbestos-related parenchymal disease (asbestosis) and to test these findings in a subsequent group of patients whose CT scans were retrospectively identified.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 
Patients
Our retrospective study was approved by the Institutional Review Board of Royal Brompton Hospital; informed consent was not required. All patients had undergone clinically indicated CT examinations and pulmonary function tests. Consecutive patients with nonmalignant asbestos-related parenchymal disease (asbestosis) at CT were retrospectively identified from three centers by using hospital radiology and respiratory clinic databases. Of the 147 patients, 14 were excluded because the images were of suboptimal quality or the diagnosis of asbestosis was equivocal with regard to exposure history and clinical criteria. The remaining 133 patients who had been exposed to asbestos had CT features of asbestosis in the context of an appropriate asbestos exposure history; in these patients, clinical diagnosis was made by an experienced occupational physician (A.J.N.T., R.M.R., or A.W.M., with 23–28 years of experience). All patients met the criteria for the diagnosis of asbestosis with respect to the recent American Thoracic Society guidelines (10).

The initial study group included 81 patients (79 men, two women; median age, 67 years; age range, 46–86 years) who were referred from two different specialist centers. All of the patients from one center (n = 42) were asbestos miners. The remainder of the patients (n = 39) had worked in a variety of asbestos-related industries. The subsequent study group consisted of 52 consecutively identified patients (all men; median age, 60 years; age range, 33–82 years) who had been exposed to asbestos and who were referred to a third specialist center. The patients ' records were independently evaluated by several authors (S.J.C., Y.C.G.L., P.S.) for data on demographics and smoking history.

CT Scanning Methods
CT was performed by using an electron-beam scanner (C-100; Imatron, San Francisco, Calif) or helical CT scanners (Somatom Power and Somatom Plus, Siemens, Erlangen, Germany; or Xpress, Toshiba, Otawara, Japan). Thin sections (1–3 mm) were obtained at 10–30-mm intervals in all patients. Patients were scanned in the supine or prone position or in both positions. Images were obtained during suspended full inspiration, were reconstructed with a high-spatial-frequency (bone) algorithm, and were acquired by using appropriate window settings (window center, –600 to –655 HU; window width, 1500–1850 HU).

CT Scoring Method
CT scans were evaluated independently by two experienced thoracic radiologists (D.M.H. and M.B.R., with 18 and 20 years of experience, respectively, in thoracic CT) who were blinded to the clinical and pulmonary function data. By using previously described methods, the observers scored all CT scans for the extent of fibrosis, emphysema, small-airways disease, pleural disease, and bronchiectasis (7,11–13). In addition, other features that have been described in patients with asbestosis were recorded (14,15). The total extent of disease was estimated by a summation of the scores of each feature at five levels: level 1, great vessels; level 2, aortic arch; level 3, carina; level 4, pulmonary venous confluence; and level 5, halfway point between level 4 and the extreme left costophrenic angle. Discrepancies between the observers for the presence or absence of fibrosis, emphysema, diffuse pleural thickening, or bronchial dilatation were resolved by consensus.

Several CT features were quantified at each level. First, the overall extent of pulmonary fibrosis (both reticular pattern and ground-glass attenuation with traction bronchiectasis) was estimated visually to the nearest 5% (6,8). The mean of the two observers' measurements was taken as the extent of fibrosis, irrespective of the predominant pattern.

Second, the relative percentages of ground-glass attenuation and reticular pattern were recorded. From this, the percentage of each pattern was defined. Ground-glass attenuation was defined as a hazy increase in attenuation, with preservation of bronchial and vascular margins. A reticular pattern was defined as innumerable interlacing line shadows (fine, intermediate, or coarse) with associated distortion of lung architecture (16).

Third, the coarseness of fibrosis was graded as 0, ground-glass attenuation with no reticular element or cysts; 1, predominantly fine intralobular fibrosis without cysts; 2, predominantly microcystic reticular pattern (definable airspaces smaller than 3 mm in diameter); or 3, predominantly macrocystic reticular pattern (airspaces larger than 3 mm in diameter) (8). The five scores were summed to provide an overall coarseness score (range, 0–15). In patients with no disease in one or more CT sections, the coarseness score was adjusted proportionately to a five-level score (in order to prevent spurious reductions in the coarseness score due to localized distribution of disease). For example, in a patient with disease in only three of five sections, a coarseness score of 5 would be adjusted by 5 divided by 3 to 8.3.

Fourth, the severity of bronchial dilatation within an area of reticular pattern or ground-glass attenuation (traction bronchiectasis or bronchiolectasis) was graded by comparing the diameter of the airway with that of the adjacent pulmonary artery (0, no dilatation; 1, mild dilatation; or 2, severe dilatation) (8). The five scores were summed to provide an overall traction bronchiectasis score (range, 0–10).

Fifth, other parenchymal abnormalities that were known to be associated with asbestosis were recorded. These included parenchymal bands (areas of nontapering linear opacities 2–5 cm in length that contacted the pleura), short interstitial lines (areas of linear opacities less than 2 cm in length that corresponded to thickened interlobular and intralobular septa), and curvilinear subpleural lines (areas of linear opacities parallel to and within 1 cm of the pleura) (14,15). These features were recorded as either present or absent.

Sixth, the extent of emphysema was visually estimated to the nearest 5% (7). The mean value was taken as the extent of emphysema. Emphysema was defined as areas of decreased attenuation, usually without discrete walls, and nonuniform distribution causing permeative destruction of lung parenchyma (16). The percentage of pure emphysema versus the percentage of emphysema intermingled with fibrosis was also assessed (7).

Seventh, the extent of pleural disease at CT was assessed by using a previously validated scoring system (12). The extent of diffuse pleural thickening was recorded as the percentage of the thorax involved, which was visually estimated to the nearest 5% at each level. The mean thickness of diffuse pleural thickening (in millimeters) was recorded. The thorax at each level was divided into quadrants, and the number of quadrants (0, 1, 2, 3, or 4) that was affected by pleural plaques was recorded. For the purposes of scoring, diffuse pleural thickening was defined as a continuous sheet of pleural thickening, and pleural plaques were defined as circumscribed areas of pleural thickening with well-demarcated edges (12). The number of areas of rounded atelectasis at each level was also recorded. Rounded atelectasis was defined as a round or oval mass abutting a pleural surface (associated with diffuse pleural thickening), with a curving "comet tail" of bronchovascular structures passing into the mass and a loss of volume in the adjacent lung (17,18).

Eighth, the severity of bronchial dilatation in morphologically normal lung at CT was graded semiquantitatively by comparison with the homologous pulmonary artery: 0, no dilatation; 1, mild dilatation (1.5–2.5 times larger in diameter than homologous pulmonary artery); or 2, severe dilatation (more than 2.5 times larger in diameter than homologous pulmonary artery). The same system was used to grade traction bronchiectasis (bronchial dilatation within areas of reticular pattern). Bronchial wall thickening was noted and was rated as follows: 0, same thickness as adjacent pulmonary artery; 1, bronchial wall 0.5 times thicker than adjacent pulmonary artery; 2, bronchial wall 0.5–1.0 times thicker than adjacent pulmonary artery; or 3, bronchial wall more than 1.0 times thicker than adjacent pulmonary artery (13). The scores at each level were summated to give a total bronchiectasis score and a total wall thickening score.

Ninth, the extent of decreased attenuation that was ascribable to small-airways disease was visually estimated to the nearest 5% (11). Small-airways disease was defined as areas of decreased attenuation that were associated with a reduction in the number and caliber of pulmonary vessels but without vascular distortion, as seen in patients with centrilobular emphysema (19).

Pulmonary Function Testing
All pulmonary function tests were performed within 12 weeks of CT. Forced expiratory volume in one second (FEV1), forced vital capacity (FVC), and the ratios of each were obtained by using spirometry (Spiroflow, P.K. Morgan, Gillingham, England; or Masterlab, E. Jaeger, Leicestershire, England). Total lung capacity (TLC) was obtained by using body plethysmography (model 3.01, P.K. Morgan; Pulmostar-SMB, Fenyves and Gut, Basle, Switzerland; or Masterlab, E. Jaeger). Single-breath carbon monoxide diffusing capacity (DLCO) corrected for hemoglobin concentration and DLCO adjusted for alveolar volume (KCO) were acquired with gas transfer equipment (model B, P.K. Morgan; or Masterlab, E. Jaeger).

In 60 patients, ear lobe capillary blood gases were analyzed by using an automatic analyzer that was calibrated according to the manufacturer's instructions. The alveolar-arterial oxygen gradient was obtained by using the simplified alveolar air equation (20). Standardized values were expressed as a percentage of the predicted values with respect to the patient's age, sex, and height (21).

Statistical Analysis
In the initial study group, interobserver agreement with respect to categorical variables was quantified by using a weighted {kappa} coefficient (22). Variations in quantitative data were expressed as a single determination standard deviation (22). Group comparisons were made by using the Student t test, the Wilcoxon rank sum test, or {chi}2 statistics, as appropriate. Linear regression was examined by using the Spearman rank correlation coefficient ({rho}). The variance of the physiologic variables accounted for by CT features was expressed as the square of the correlation coefficient (R2). The mean values of the two observers' scores were used for multiple regression and linear regression analysis. Multiple regression analysis was performed by using Stata software (Stata, Santa Monica, Calif).

Independent relationships between the extent of disease at CT and pulmonary function indexes were examined by using stepwise forward regression; CT variables that were present in less than 10% of patients and variables with unacceptable observer variation (weighted {kappa} < 0.40) were not evaluated. Correlations with a P value of less than .01 were retained. Variables that were not normally distributed were transformed prior to analysis (categorical or zero-skewness logarithmic transformations, as appropriate). The assumptions of the multiple linear regression analysis were met in all analyses, as judged by testing for heteroscedasticity and omitted variables.

To evaluate a relatively simple system versus a more comprehensive system, multiple regression analyses were performed in the initial study group by using (a) the extent of the three key CT variables only (ie, fibrosis, emphysema, and diffuse pleural thickening) (simplified equation) and (b) all CT variables that met the above criteria (comprehensive equation).

To validate the scoring system in a separate group of patients, the CT scores from the subsequent study group were used in the derived equations to calculate a predicted lung function value. The predicted values were compared with the measured values by using the Spearman rank correlation coefficient.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 
The demographics, smoking history, and radiologic features of both study groups are given in Table 1.


View this table:
[in this window]
[in a new window]

  		Table 1. Demographic Data, Smoking History, Pulmonary Function, and Prevalence of Abnormalities at CT

 
Initial Study Group
Interobserver agreement for CT features in the initial study group varied, with {kappa} values ranging from 0.29 to 0.88 (Table 2). Variations in quantitative data were expressed as a single determination standard deviation (Table 3). Small-airways disease was present in less than 10% of cases and was therefore excluded from further analysis. Univariate correlations between pulmonary function and the selected CT variables (Table 4) showed the strongest correlation between the extent of fibrosis at CT and TLC and DLCO, the extent of emphysema at CT and KCO and DLCO, and the extent of diffuse pleural thickening at CT and FVC and TLC (Fig 1).


View this table:
[in this window]
[in a new window]

  		Table 2. Observer Variation for CT Patterns in Initial Study Group

 

View this table:
[in this window]
[in a new window]

  		Table 3. Observer Variation for CT Patterns in Initial Study Group

 

View this table:
[in this window]
[in a new window]

  		Table 4. Correlations between the Extent of Fibrosis, Emphysema, and Diffuse Pleural Thickening at CT and Functional Indexes

 

Figure 1
View larger version (112K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a: Transverse unenhanced thin-section CT scans obtained in a patient in the prone position who had been exposed to asbestos (initial study group). (a) By using a simple score system at level 2, observer 1 described the extent of fibrosis and emphysema as 30% and 40%, respectively, and observer 2 described the extent of fibrosis and emphysema as 40% and 50%, respectively. No diffuse pleural thickening was noted. (b) By using a simple score system at level 5, observer 1 described the extent of fibrosis, emphysema, and diffuse pleural thickening (percentage of circumference of thorax) as 65%, 15%, and 30%, respectively, and observer 2 described the extent of fibrosis, emphysema, and diffuse pleural thickening as 55%, 25%, and 20%, respectively.

 

Figure 1
View larger version (108K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b: Transverse unenhanced thin-section CT scans obtained in a patient in the prone position who had been exposed to asbestos (initial study group). (a) By using a simple score system at level 2, observer 1 described the extent of fibrosis and emphysema as 30% and 40%, respectively, and observer 2 described the extent of fibrosis and emphysema as 40% and 50%, respectively. No diffuse pleural thickening was noted. (b) By using a simple score system at level 5, observer 1 described the extent of fibrosis, emphysema, and diffuse pleural thickening (percentage of circumference of thorax) as 65%, 15%, and 30%, respectively, and observer 2 described the extent of fibrosis, emphysema, and diffuse pleural thickening as 55%, 25%, and 20%, respectively.

 
To examine the possible interaction between the development of emphysema and fibrosis, patients with the both emphysema and fibrosis (n = 43) were evaluated separately. The extent of fibrosis did not differ between patients with and those without emphysema. Furthermore, there was no correlation between the extent of fibrosis and that of emphysema in individuals with coexisting emphysema and fibrosis (n = 43).

Independent relationships between individual pulmonary function and the extent of the three main processes at CT (fibrosis, emphysema, and diffuse pleural thickening) showed that the predicted percentages for TLC and DLCO were excellent overall measures of morphologic abnormality (Table 5); these indexes were therefore chosen as the dependent variables for the multivariate equations.


View this table:
[in this window]
[in a new window]

  		Table 5. Independent Relationships between Functional Indexes and the Extent of Three Main CT Features in the Initial Study Group

 
The two simplified equations were as follows: TLC (percentage predicted) = 96.4 – (0.46 · Efibrosis) + (0.22 · Eemphysema) – (0.39 · CEdiffuse) and DLCO (percentage predicted) = 84.7 – (0.21 · CEdiffuse) – (0.49 · Eemphysema) – (0.55 · Efibrosis), where Eemphysema is the extent of emphysema, Efibrosis is the extent of fibrosis, and CEdiffuse is the circumferential extent of diffuse pleural thickening.

More comprehensive multiple regression analyses were also performed by using all the variables scored by the observers. The R2 values were 0.58 for TLC, 0.57 for DLCO, and 0.41 for KCO. Therefore, the combined CT variables predicted 58% and 57% of the variability in TLC and DLCO, respectively, despite considerable variation in the proportion of coexisting pathologic conditions. Furthermore, by including all ancillary CT features, there was little increased correlation relative to the three key features.

Subsequent Study Group
The derived equations were then tested in a subsequent group of patients. For the simplified equation, the correlation between observed and predicted TLC for the three key CT variables was 0.74, 0.67, and 0.44 for {rho}, r, and R2, respectively (P < .001), and the correlation between observed and predicted DLCO was 0.64, 0.68, and 0.46 for {rho}, r, and R2, respectively (P < .001). Therefore, up to 44% of the variance in TLC and 46% of the variance in DLCO can be accounted for by the CT scores for fibrosis, emphysema, and diffuse pleural thickening.

TLC, DLCO, and FVC values that were predicted by using the comprehensive equations for all CT variables also showed excellent correlation with observed values (Fig 2); {rho}, r, and R2 values were 0.75, 0.72, and 0.52, respectively, for TLC (P < .001); 0.58, 0.63, and 0.40, respectively, for DLCO (P < .001); and 0.64, 0.68, and 0.46, respectively, for FVC (P < .001). The comprehensive equations were able to explain 52% of the variance in TLC, 40% of the variance in DLCO, and 46% of the variance in FVC.


Figure 2
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2a: Graphs are shown for (a) predicted TLC (derived from multiple regression analyses) versus measured TLC and (b) predicted DLCO versus measured DLCO in the subsequent group of patients who had been exposed to asbestos (n = 52). Equation accuracy was largely unaffected by different acquisition parameters for CT and pulmonary function testing.

 

Figure 2
View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2b: Graphs are shown for (a) predicted TLC (derived from multiple regression analyses) versus measured TLC and (b) predicted DLCO versus measured DLCO in the subsequent group of patients who had been exposed to asbestos (n = 52). Equation accuracy was largely unaffected by different acquisition parameters for CT and pulmonary function testing.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ADVANCE IN KNOWLEDGE
 References
 
Our results demonstrate that CT substantially increases the precision with which the results of pulmonary function tests can be interpreted in patients who have been exposed to asbestos.

Our findings may also have medicolegal relevance. In individuals who have been exposed to asbestos, the definition of the functional consequences of asbestos (as opposed to those of smoking) is an important medicolegal problem because emphysema is not generally compensable. No robust method is currently available for this purpose, and the degree of physiologic impairment attributed to asbestos exposure in smokers is often estimated arbitrarily. By using multiple regression analysis, we have identified the combination of CT features most closely linked to individual pulmonary function indexes. Thus, for a given reduction in TLC or DLCO, the proportion of pulmonary deficit that can be ascribed to fibrosis, diffuse pleural thickening, and emphysema can be preliminarily estimated within a 95% confidence interval. The derived multiple regression analyses were validated by their accurate prediction of pulmonary function indexes in the subsequent study group.

An advantage of the described technique is that the proportion of deficit is assigned independent of the normal variation in predicted pulmonary function values. Depending on the starting values, which range from 80% to 120% of the predicted values, a reduction to 75% represents a decrease of between 5% and 45% in a given individual. CT scoring takes no account of this variation but merely assigns the proportions of the observed functional deficit that are ascribable to fibrosis, emphysema, and diffuse pleural thickening whether the level of functional impairment is 75% or 50% of the predicted value. Thus, the application of our methods consists of quantifying each CT feature, adjusting the scored extent feature for its regression coefficient, and then computing the ratios (Table 6). In individuals who have been exposed to asbestos, it is relatively straightforward to quantify the extent of disease at CT, to apply the regression coefficient, and to calculate the relative functional effect of individual CT features (eg, fibrosis and emphysema).


View this table:
[in this window]
[in a new window]

  		Table 6. Application of the Population DLCO Equation in an Individual Patient

 
A further advantage is the requirement to score only the three cardinal CT abnormalities that characterize asbestos exposure and cigarette smoking–related damage. A large number of ancillary CT features are present in individuals who have been exposed to asbestos, but the inclusion of these features did not materially diminish the relationship between structure and function. The exclusion of a multiplicity of ancillary features has obvious advantages in terms of reducing interobserver variation, thereby increasing the general applicability of our techniques.

Stronger functional-morphologic correlations than those currently observed may be difficult to achieve, because variability in CT scoring and pulmonary function testing is inescapable. The use of three pulmonary function tests and a variety of CT scanners can be viewed as strengths of the study, because a pragmatic CT scoring system needs to be widely applicable. However, variation between centers must necessarily weaken functional-morphologic relationships. Interlaboratory variability in pulmonary function data is a greater source of inaccuracy than intralaboratory variability (23). Similarly, there are variations in image definition between CT scanners. The observed functional-morphologic relationships are surprisingly strong when these constraints are taken into account.

Criteria for diagnosis and compensation in cases of nonmalignant asbestos pleuroparenchymal disease vary from country to country. The method described in our study provides a unifying method for the apportionment of functional deficit, which could aid medicolegal assessment. However, although our CT scoring system can be readily applied by experienced thoracic radiologists, it is not yet clear whether it should be widely used by less experienced radiologists and occupational physicians. The situation is analogous to the reading of chest radiographs in patients with occupational lung disease; such readings are largely reserved for designated trained readers, as in the International Labor Office system (5).

Ultimately, the solution may lie in computer-aided techniques, but these are not yet sufficiently advanced (and may not be for some years) to supplant the human eye. Early attempts to quantify the extent of pulmonary fibrosis at CT by using computer-aided techniques have had limited success and are particularly problematic when there are coexisting disease processes (24).

The accuracy of the explanatory equations in the subsequent study group is reassuring because, in this group, disease was milder and more representative of the type of disease encountered in routine practice; the initial study group was dominated by miners with heavy asbestos exposure. Furthermore, in patients with milder disease, variations in normal predictive values are a serious source of noise, which might be expected to diminish relationships between structure and function. The findings indicate that our techniques are applicable across a wide range of disease severity.

Similar to other series (1–3), 85% of the patients exposed to asbestos in our study were present or past smokers. Hence, coexisting emphysema was common (32% of our study population). The interaction between cigarette smoking and asbestos exposure is complicated. Smoking retards the lung clearance of asbestos fibers and may contribute to the severity and progression of asbestosis (25,26).

The results of some studies have shown an increased prevalence of emphysema in both nonsmoking and smoking workers who had been exposed to asbestos (2); however, there was no relationship between the extent of fibrosis and that of emphysema on CT scans in the present study. Nonetheless, even if it is eventually established that asbestos exposure can cause emphysema in nonsmokers, our equations discriminate between functional impairment that is definitely ascribable to asbestos exposure (fibrosis and diffuse pleural thickening) and functional impairment that is largely the result of cigarette smoking, with a possible contribution from occupational exposure.

Apart from the retrospective nature of our study, other possible limitations are the absence of histopathologic proof of asbestosis and the width of the 95% confidence intervals in the explanatory equations. The current consensus is that there is no justification for open lung biopsy in patients who are suspected of having asbestosis; diagnosis is usually made on the basis of clinical data and an appropriate and substantial history of asbestos exposure (10). Lack of histopathologic proof is a common limitation to many other radiologic studies of asbestosis (27). Furthermore, the histopathologic demonstration of asbestos (or ferruginous) bodies in patients with pulmonary fibrosis is not entirely specific for the condition and may reflect incidental environmental exposure (28).

The application of population data to individuals is always problematic. However, unlike previous evaluations with chest radiography and pulmonary function testing, the derivation of 95% confidence intervals allows the intrinsic variability of our approach to be stated with confidence. The confidence intervals in the present study are acceptable. For example, a 95% confidence interval ranging from 23.2% to 54.4% (median, 31.2%) for the functional effect of emphysema may appear wide but equates to a standard deviation of less than 10% around the mean value of 39.2%. This method is suitable for adaptation to medicolegal use. One pragmatic approach would be to quantify the functional deficit by using the regression coefficient itself rather than by arbitrarily selecting either end of the confidence interval.

The important point is that, compared with previous evaluations that used chest radiographs or the instinctive view of a chest physician with lung function tests in isolation, the importance of the 95% confidence interval is that one can state the moderate variability of this procedure with great confidence. In the example we cite, a more reassuring expression is that the deficit ascribable to smoking was 40% ± 12 (standard deviation) of the 95% confidence interval. We believe that this degree of precision is a considerable advance.

Our findings highlight the difficulty in quantifying the severity of asbestos-related lung damage, whether for clinical or medicolegal purposes, by using pulmonary function indexes in isolation. Emphysema and asbestosis both produce reductions in gas transfer, whereas the opposing effects of emphysema and asbestosis for spirometric and plethysmographic measurements result in a deceptive preservation of lung volumes (29). This phenomenon was quantified by multiple regression analysis, with an increase in TLC seen with increasingly extensive emphysema after adjustment for other CT abnormalities. The essential implication is that neither disease severity nor compensation should be quantified with lung volumes alone (which may seriously understate the amount of asbestos-induced disease) or with isolated measurements of gas transfer (which are confounded by concurrent emphysema).

To overcome this problem, it is necessary to quantify the separate functional consequences of emphysema and pulmonary fibrosis in patients. The method described in our study provides a robust semiquantitative method for the estimation of the extent of fibrosis and emphysema and may be of use in the assessment of individuals who have been exposed to asbestos.


   ADVANCE IN KNOWLEDGE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
References
 


   FOOTNOTES
 

Abbreviations: DLCO = carbon monoxide diffusing capacity • FEV1 = forced expiratory volume in one second • FVC = forced vital capacity • KCO = DLCO adjusted for alveolar volume • TLC = total lung capacity

Authors stated no financial relationship to disclose.

Author contributions: Guarantors of integrity of entire study, D.M.H., A.U.W.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, S.J.C., P.S., R.M.R., A.W.M., A.U.W.; clinical studies, Y.C.G.L., P.S., M.B.R., A.J.N.T., A.W.M., A.U.W.; statistical analysis, S.J.C., A.U.W.; and manuscript editing, all authors


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 

  1. Lee YC, Singh B, Pang SC, de Klerk NH, Hillman DR, Musk AW. Radiographic (ILO) readings predict arterial oxygen desaturation during exercise in subjects with asbestosis. Occup Environ Med 2003;60:201–206.[Abstract/Free Full Text]
  2. Begin R, Filion R, Ostiguy G. Emphysema in silica- and asbestos-exposed workers seeking compensation: a CT scan study. Chest 1995;108:647–655.[Abstract/Free Full Text]
  3. Osinubi OY, Afilaka AA, Doucette J, et al. Study of smoking behaviour in asbestos workers. Am J Ind Med 2002;41:62–69.[CrossRef][Medline]
  4. Guidotti TL. Legislative controls and compensation: North America. In: Hendrick DJ, Burge PS, Beckett WS, Churg A, eds. Occupational disorders of the lung. London, England: Saunders, 2002; 545–553.
  5. International Labour Office. International classification of radiographs of pneumoconioses. Revised edition 2000. Occupational Health and Safety series no. 22. Geneva, Switzerland: International Labour Office, 2002.
  6. Wells AU, Hansell DM, Rubens MB, Cullinan P, Black CM, du Bois RM. The predictive value of appearances on thin-section computed tomography in fibrosing alveolitis. Am Rev Respir Dis 1993;148:1076–1082.[Medline]
  7. Wells AU, King AD, Rubens MB, Cramer D, du Bois RM, Hansell DM. Lone cryptogenic fibrosing alveolitis: a functional-morphologic correlation based on extent of disease on thin-section computed tomography. Am J Respir Crit Care Med 1997;155:1367–1375.[Abstract]
  8. Wells AU, Desai SR, Rubens MB, et al. Idiopathic pulmonary fibrosis: a composite physiologic index derived from disease extent observed by computed tomography. Am J Respir Crit Care Med 2003;167:962–969.[Abstract/Free Full Text]
  9. Jarad NA, Wilkinson P, Pearson MC, Rudd RM. A new high resolution computed tomography scoring system for pulmonary fibrosis, pleural disease, and emphysema in patients with asbestos related disease. Br J Ind Med 1992;49:73–84.[Medline]
  10. American Thoracic Society. Diagnosis and initial management of nonmalignant diseases related to asbestos. Am J Respir Crit Care Med 2004;170:691–715.[Free Full Text]
  11. Hansell DM, Rubens MB, Padley SP, Wells AU. Obliterative bronchiolitis: individual CT signs of small airways disease and functional correlation. Radiology 1997;203:721–726.[Abstract/Free Full Text]
  12. Copley SJ, Wells AU, Rubens MB, et al. Functional consequences of pleural disease evaluated with chest radiography and CT. Radiology 2001;220:237–243.[Abstract/Free Full Text]
  13. Roberts HR, Wells AU, Milne DG, et al. Airflow obstruction in bronchiectasis: correlation between computed tomography features and pulmonary function tests. Thorax 2000;55:198–204.[Abstract/Free Full Text]
  14. Aberle DR, Gamsu G, Ray CS, Feuerstein IM. Asbestos-related pleural and parenchymal fibrosis: detection with high-resolution CT. Radiology 1988;166:729–734.[Abstract/Free Full Text]
  15. Yoshimura H, Hatakeyama M, Otsuji H, et al. Pulmonary asbestosis: CT study of subpleural curvilinear shadow—work in progress. Radiology 1986;158:653–658.[Abstract/Free Full Text]
  16. Austin JH, Muller NL, Friedman PJ, et al. Glossary of terms for CT of the lungs: recommendations of the Nomenclature Committee of the Fleischner Society. Radiology 1996;200:327–331.[Free Full Text]
  17. McHugh K, Blaquiere RM. CT features of rounded atelectasis. AJR Am J Roentgenol 1989;153:257–260.[Abstract/Free Full Text]
  18. Lynch DA, Gamsu G, Ray CS, Aberle DR. Asbestos-related focal lung masses: manifestations on conventional and high-resolution CT scans. Radiology 1988;169:603–607.[Abstract/Free Full Text]
  19. Stern EJ, Swensen SJ, Hartman TE, Frank MS. CT mosaic pattern of lung attenuation: distinguishing different causes. AJR Am J Roentgenol 1995;165:813–816.[Abstract/Free Full Text]
  20. Riley RL, Cournand A. Ideal alveolar air and the analysis of ventilation-perfusion relationships in the lungs. J Appl Physiol 1949;1:825–847.[Free Full Text]
  21. Quanjer PH. Standardized lung function testing. Eur Respir J 1993;6(suppl):1–100.[Medline]
  22. Brennan P, Silman A. Statistical methods for assessing observer variability in clinical measures. BMJ 1992;304:1491–1494.[Free Full Text]
  23. Kangalee KM, Abboud RT. Interlaboratory and intralaboratory variability in pulmonary function testing: a 13-year study using a normal biologic control. Chest 1992;101:88–92.[Abstract/Free Full Text]
  24. Rodriguez LH, Vargas PF, Raff U, et al. Automated discrimination and quantification of idiopathic pulmonary fibrosis from normal lung parenchyma using generalized fractal dimensions in high-resolution tomography images. Acad Radiol 1995;2:10–18.[CrossRef][Medline]
  25. Churg A, Stevens B. Enhanced retention of asbestos fibers in the airways of human smokers. Am J Respir Crit Care Med 1995;151:1409–1413.[Abstract]
  26. Lilis R, Miller A, Godbold J, Chan E, Selikoff IJ. Radiographic abnormalities in asbestos insulators: effects of duration from onset of exposure and smoking—relationships of dyspnea with parenchymal and pleural fibrosis. Am J Ind Med 1991;20:1–15.[Medline]
  27. Lynch DA. CT for asbestosis: value and limitations. AJR Am J Roentgenol 1995;164:69–71.[Abstract/Free Full Text]
  28. Craighead JE, Abraham JL, Churg A, et al. The pathology of asbestos-associated diseases of the lungs and pleural cavities: diagnostic criteria and proposed grading schema—report of the Pneumoconiosis Committee of the College of American Pathologists and the National Institute for Occupational Safety and Health. Arch Pathol Lab Med 1982;106:544–596.[Medline]
  29. Barnhart S, Hudson LD, Mason SE, Pierson DJ, Rosenstock L. Total lung capacity: an insensitive measure of impairment in patients with asbestosis and chronic obstructive pulmonary disease? Chest 1988;93:299–302.[Abstract/Free Full Text]




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow All Versions of this Article:
2421051167v1
242/1/258    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Copley, S. J.
Right arrow Articles by Wells, A. U.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Copley, S. J.
Right arrow Articles by Wells, A. U.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE