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


     

DOI: 10.1148/radiol.2321030392
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
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 Sette, A.
Right arrow Articles by Müller, N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sette, A.
Right arrow Articles by Müller, N.
(Radiology 2004;232:66-74.)
© RSNA, 2004

Thoracic Imaging

Thin-Section CT Abnormalities and Pulmonary Gas Exchange Impairment in Workers Exposed to Asbestos1

Andrea Sette, MD, PhD, J. Alberto Neder, MD, PhD, Luiz E. Nery, MD, PhD, Jorge Kavakama, MD, Reinaldo T. Rodrigues, MD, Mario Terra-Filho, MD, PhD, Sandra Guimarães, MD, PhD, Ericson Bagatin, MD, PhD and Nestor Müller, MD, PhD

1 From the Pulmonary Function and Clinical Exercise Physiology Unit, Div of Respiratory Diseases (A.S., J.A.N., L.E.N., S.G.), and Dept of Radiology (R.T.R.), Federal Univ of São Paulo, Rua Professor Francisco de Castro 54, Vila Clementino, CEP 04020–050, São Paulo, Brazil; Occupational Health Service, State Univ of Campinas, São Paulo, Brazil (E.B.); Div of Respiratory Diseases, Univ of São Paulo, Brazil (J.K., M.T.F.); and Dept of Radiology, Vancouver General Hospital, BC, Canada (N.M.). Received Mar 17, 2003; revision requested May 30; final revision received Oct 8; accepted Nov 20. Supported in part by a research grant from FAPESP-Brazil (No. 96–10415-6). A.S. supported by a fellowship grant from CNPq-Brazil. S.G. supported by a fellowship grant from FAPESP-Brazil. Address correspondence to J.A.N. (e-mail: albneder@pneumo.epm.br).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
RESULTS
DISCUSSION
 REFERENCES
 
PURPOSE: To evaluate the relationship between abnormalities at thin-section computed tomography (CT) and indexes of pulmonary gas exchange impairment at rest and during moderate exercise in workers exposed to asbestos.

MATERIALS AND METHODS: Eighty-two workers with long-term exposure to asbestos and abnormal thin-section CT findings underwent respiratory physiologic measurements at rest (lung diffusing capacity, DLCO) and during exercise (oxygen uptake–corrected alveolar-arterial pressure difference for oxygen, {Delta}P[A-a]O2/VO2). CT results were compared with physiologic measurements of impairment in gas exchange (DLCO < 70% predicted value and/or {Delta}P[A-a]O2/VO2 > 20 mm Hg · L · min–1). The CT findings were divided into five categories by using a previously described method. Odds ratios and 95% CIs for gas exchange defects were calculated for patients grouped according to CT findings. Logistic regression analysis was performed with gas exchange as the dependent response and CT abnormalities as independent variables.

RESULTS: A significant association was found between extent of disease at CT and impairment of gas exchange (P < .01). Probability of functional impairment was increased with multifocal (class II) and diffuse (class III) CT abnormalities, particularly when several lesion types were found concomitantly. Logistic regression analysis demonstrated significant association of parenchymal bands (odds ratio, 6.20; 95% CI: 1.99, 19.22) and subpleural nodules (odds ratio, 3.83; 95% CI: 1.23, 11.89) with functional impairment. Presence and number of pleural plaques did not improve model accuracy for gas exchange impairment prediction (P > .05).

CONCLUSION: Thin-section CT grading of interstitial lung disease is useful in assessing the likelihood of pulmonary gas exchange impairment at rest (DLCO) and during exercise ({Delta}P[A-a]O2/VO2) in workers with long-term asbestos exposure.

© RSNA, 2004

Index terms: Asbestos • Computed tomography (CT), thin-section, 60.12118 • Lung, function, 60.773, 60.90 • Pneumoconiosis, 60.773


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
 REFERENCES
 
Asbestosis is an occupational lung disease characterized by interstitial fibrosis in association with asbestos bodies or asbestos fibers (1). The disease constitutes a major occupational health problem, especially in underdeveloped countries where effective measures of asbestos control were established only recently (2). Asbestosis may result in mechanical lung abnormalities and, in particular, impairment in intrapulmonary gas exchange as suggested by low diffusing capacity for carbon monoxide (DLCO) and increased exercise alveolar-arterial oxygen gradient (P[A-a]O2) (3–6).

Radiologic evaluation of workers exposed to asbestos plays a critical role in the assessment of disease occurrence and objective quantification of parenchymal and/or pleural involvement. The traditional approach consists of assessment of presence and extent of abnormalities on chest radiographs on the basis of the International Labour Office (ILO) classification of radiographs of pneumoconiosis (7). Although the system was originally described as an epidemiologic tool, it has been used for clinical diagnosis and prediction of functional impairment (8–11).

Since the late 1980s, thin-section computed tomography (CT) has proved to be more sensitive and accurate than conventional radiography in the detection of subtle parenchymatous disease (11–17). Accordingly, emphasis has been placed on the use of thin-section CT in the early detection of asbestos-related parenchymal abnormalities, such as in patients with a score of less than 1/0 according to the ILO scheme (11,13,17). For this purpose, Gamsu and colleagues (18) developed a dual semiquantitative and qualitative scoring system for thin-section CT evaluation of asbestos-exposed workers. In this classification, the reader uses a combination of an extent and severity index (0–4) and a cumulative score, which is used to add the different types of interstitial abnormalities (up to 6). This classification has been shown to correlate better with histopathologic evidence of asbestosis than does the ILO system—notably in milder incipient cases (18).

In the clinical context, however, it would also be interesting to know whether this higher sensitivity of thin-section CT, as compared with conventional radiography, translates to enhanced prediction of relevant physiologic outcomes, such as impairment in gas exchange. Thus, the purpose of our study was to evaluate the relationship between interstitial abnormalities at thin-section CT and indexes of pulmonary gas exchange impairment at rest and during moderate exercise in workers exposed to asbestos.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
DISCUSSION
 REFERENCES
 
Study Design and Subjects
This was a prospective cross-sectional study. Eighty-seven asbestos-cement workers (n = 57) and chrysotile miners (n = 30) were consecutively referred for impairment evaluation and considered for study inclusion. Thin-section CT was performed in all subjects as part of the medicolegal evaluation process. Individuals were excluded from the study if they had other known causes of interstitial lung disease, anemia (hemoglobin concentration less than 14 g · dL–1 [140 g/L], congestive heart failure or other serious cardiovascular disease, or malignancy, or if they could not tolerate exercise on the stationary bicycle. Five subjects were excluded because of advanced cardiovascular disease: All 82 remaining subjects (mean age ± SD, 64.1 years ± 8.6; range, 41–82 years) had abnormal thin-section CT findings, at least 1 year of asbestos exposure (mean, 13.8 years ± 10.1), and more than 15 years of latency from the beginning of asbestos exposure to the time of evaluation. According to a simplified version of the Medical Research Council scale of breathlessness (19), 42 subjects (51%) did not report dyspnea during activities of daily living; in the remaining 40 subjects, this symptom was classified as mild in 30, moderate in five, and severe in five.

Data were pooled for analysis only after certification that there was no significant effect of specific setting of exposure (asbestos-cement vs chrysotile mining) on the pulmonary gas exchange variables (nonpaired t test, P > .05). Informed consent (as approved by the Institutional Medical Ethics Committee, which also approved our study) was obtained from all subjects.

Imaging Evaluation
Thin-section CT.—Thin-section CT scans of lung parenchyma were obtained by using an X-Vision scanner (Toshiba, Tokyo, Japan). Images were recorded at full inspiration with the subjects in the prone position. Images with 2-mm collimation were obtained every 1 cm from the lung apex to the lowest hemidiaphragm. The scanning time was 1.2 seconds. Images were reconstructed by using a high-spatial-frequency algorithm and photographed at two window settings: (a) window level of 700 HU and window width of 1,500 HU for assessment of lung parenchyma and (b) window level of 35 HU and window width of 350 HU for assessment of pleura and mediastinum (20).

The thin-section CT scans were evaluated by three readers (J.K., R.T.R., N.M.) (experience with thin-section CT ranging from 9 to 12 years), and recorded findings were agreed on by means of consensus. The readers were aware that subjects had been exposed to asbestos, but they were blinded to physiologic evaluation and conventional radiographic findings. Presence and number of diaphragmatic, chest wall, and mediastinal pleural plaques were assessed at each hemithorax.

Parenchymal abnormalities were classified according to the method proposed by Gamsu and co-workers (18). This scheme is divided into two components: (a) a semiquantitative extension and severity score and (b) a qualitative score based on the different types of thin-section CT abnormalities. For the extension and/or severity score, a four-point scale was used: grade 0, normal, without evidence of interstitial lung disease; grade I, a few sites (one to four) of interstitial abnormality unlikely to represent diffuse interstitial fibrosis (ie, different from those described below); grade II, multifocal abnormalities (as described later) that are limited in extent (but in both hemithoraces) or occur in at least two levels in one hemithorax and are consistent with asbestosis; and grade III, profuse bilateral interstitial alterations (as described later) on several CT scans.

The qualitative score was based on the following interstitial findings, which have been associated with asbestosis (21): (a) interstitial lines due to thickening of the interlobular septa and centrilobular core structures, (b) parenchymal bands (long scars), (c) subpleural curvilinear opacities (subpleural lines), (d) honeycombing, (e) subpleural nodules, and (f) architectural distortion, as defined by distorted secondary lobules within peripheral areas. These abnormalities had to be bilateral or present on several CT scans in one hemithorax (18).

On the basis of a combination of these parenchymal thin-section CT findings, patients were allocated into five groups: group A (extension and/or severity grade I); group B (extension and/or severity grade II; two or fewer abnormalities); group C (extension and/or severity grade III; three or fewer abnormalities); group D (extension and/or severity grade II; three or more abnormalities); and group E (extension and/or severity grade III; four or more abnormalities).

Chest radiographs.—Standard high-kilovoltage posteroanterior chest radiographs were obtained up to 1 month before CT. They were classified by two respiratory physicians (L.E.N. and E.B., with 15 and 20 years of experience; one B-reader [E.B.]) and one radiologist (J.K., 18 years of experience) on the basis of the 1980 ILO classification of radiographs and pneumoconiosis (7). Each reader evaluated all images. Recorded values were agreed on by means of consensus. Interpretations included determination of category and profusion of parenchymal opacities and width and extent of diaphragmatic, chest wall, and mediastinal pleural plaques. The readers were blinded to physiologic evaluation and CT results.

Pulmonary Function Tests
Spirometry.—Spirometric tests were performed by using the CPF system (Medical Graphics, St Paul, Minn) with flow measurements acquired with a calibrated pneumotacograph. The subjects completed three acceptable maximal forced expiratory maneuvers; technical procedures, acceptability, and reproducibility criteria were those recommended by the American Thoracic Society (22). Forced vital capacity and forced expiratory volume in 1 second were recorded and expressed as body temperature, ambient pressure, saturated with water vapor (BTPS) conditions. Values were compared with those predicted by Pereira et al (23)—that is, functional variables were expressed as a percentage of the predicted value according to this source of reference values. No statistical test was performed for this purpose.

Lung diffusing capacity for carbon monoxide.—Carbon monoxide diffusing capacity (DLCO) was measured by means of the modified Krogh technique (single breath) by using a computer-based automated system (PF-DX System; Medical Graphics). The subjects performed two acceptable and reproducible tests, with the results being within 10% or 3 mL CO · min–1 · mm Hg–1. Absolute values were reported in milliliters of carbon monoxide per millimeters of mercury of driving pressure at standard temperature and pressure with dry conditions, or STPD. Absolute values and alveolar volume–corrected values were compared with the standards from Crapo et al (24).

Arterial Blood Gas Tensions
Rest and exercise arterial partial pressure for oxygen (PaO2) values were obtained by means of radial artery puncture in standard anaerobic conditions. Samples were analyzed immediately by using a calibrated tonometer (ABL 330; Radiometer, Copenhagen, Denmark). The alveolar-arterial oxygen pressure gradient (P[A-a]O2) was calculated by using the simplified alveolar gas equation and the measured respiratory exchange ratio, or RER, values (see the Cardiopulmonary Exercise Testing section): P(A-a)O2 = {[FIO2 · (Pb – 47)] – PaCO2 [(FIO2 + 1) – (FIO2/RER)]} – PaO2, where FIO2 is the inspired fraction of oxygen and Pb is the local barometric pressure.

Exercise pulmonary gas exchange variables were related to the approximated metabolic demand (pulmonary oxygen uptake, or VO2) and expressed in relative values—that is, exercise-rest ({Delta}).

A clinically significant pulmonary gas exchange impairment was defined as DLCO less than 70% of the predicted value (4,25) and/or {Delta}P(A-a)O2/VO2 more than 20 mm Hg · L · min–1 (4,26).

Cardiopulmonary Exercise Testing
The exercise tests were performed with an electromagnetically braked cycle ergometer (CPE 2000; Medical Graphics) with gas exchange and ventilatory variables being analyzed breath by breath by using a calibrated computer-based exercise system (MGC-CPX system; Medical Graphics). All tests were performed in the same laboratory at an altitude of 680 m above sea level with a barometric pressure of 685–699 mm Hg and an ambient temperature between 18° and 20°C.

The following data were obtained breath by breath and expressed as 15-second averages: pulmonary oxygen uptake (VO2, in milliliters per minute STPD), carbon dioxide output (VCO2, in milliliters per minute STPD), respiratory exchange ratio, minute ventilation (VE, in liters per minute BTPS), ventilatory equivalent for O2 and CO2 (VE/VO2 and VE/VCO2), and end-tidal partial pressures of O2 and CO2 (PETO2 and PETCO2, in millimeters of mercury).

Initially, a rapidly incremental exercise test was performed with noninvasive determination of lactate threshold (VO2{theta}L). This parameter was obtained by means of the gas exchange method of visually inspecting the inflection point of VCO2 with regard to VO2 (modified V slope) (27) and by means of the ventilatory method, when VE/VO2 and PETO2 increased while VE/VCO2 and PETCO2 remained stable (28). After 1 hour of rest, patients undertook a constant work rate, or WR, test at {theta}L. The test protocol consisted of (a) 3 minutes at rest, (b) 3 minutes of unload, (c) 6-minute loaded phase at WR{theta}L, and (d) recovery period. Arterial blood gas measurements (see above) were acquired during the steady-state phase (5th to 6th minute): The steady-state condition was established on the basis of VO2 and RER (VCO2/VO2) stability (variation less than 100 mL/min and 0.02, respectively).

Statistical Analysis
Skewness was assumed as not present if the skewness value was less than twice its standard error. Considering that no data showed a significantly skewed distribution, all values were expressed as mean ± SD. The nonpaired t test was used to assess comparisons between groups. One-way analysis of variance was used to determine differences in patients classified according to thin-section CT or chest radiography findings: Post hoc analysis was performed by using the Scheffé contrast test. Differences in proportions were assessed by means of {chi}2 tests. Odds ratios and 95% CIs for prediction of gas exchange impairment were calculated for each thin-section CT and chest radiography group (29). The McNemar test was used to analyze the level of agreement between the two imaging methods in the identification of pleural plaques (30).

Diagnostic accuracy of individual tests in the prediction of gas exchange impairment was analyzed in terms of sensitivity (TP/[TP + FN]), specificity (TN/[TN + FP]), positive predictive value (TP/[TP + FP]), and negative predictive value (TN/[TN + FN]), where FN is number of false-negative findings, FP is number of false-positive findings, TN is number of true-negative findings, and TP is number of true-positive findings. Logistic regression analysis was performed with gas exchange impairment as the dependent response and thin-section CT findings as independent variables. Estimated logit coefficients (ß) for the significant variables were calculated, along with the Wald statistic (ß/standard error) (31). The probability of type I error was established at .05 for all tests.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
DISCUSSION
 REFERENCES
 
CT and Chest Radiographic Findings
The overall extent of disease on thin-section CT scans was classified as grade I in 21 patients, grade II in 36, and grade III in the remaining 25 patients (Table 1). The number of patterns of abnormality varied considerably between individuals; most workers, however, had one to three abnormalities (59 of 82 workers) (Table 1). Interstitial lines were the most common finding: All but one patient had this abnormality. Frequency of the remaining imaging patterns was as follows: subpleural nodules in 36 subjects (44%), parenchymal bands in 32 (39%), subpleural lines in 27 (33%), honeycombing in 27 (33%), and architectural distortion in 23 (28%).


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

  		TABLE 1. Relationship between Extension and Severity of Parenchymal Involvement and Cumulative Number of Thin-Section CT Abnormalities

 
There was a significant association between extent of disease and cumulative number of abnormalities (P < .01). Eighteen of 21 workers with grade I findings at CT had only one or two types of parenchymal abnormalities, and none had more than four (Table 1). In contrast, four or more types of abnormalities were found on 14 of 25 CT scans with grade III findings (Table 1). According to CT findings, patients were grouped as follows: Group A had 21 patients; group B, 21 patients; group C, 11 patients; group D, 15 patients; and group E, 14 patients (Table 2).


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

  		TABLE 2. Relationship between Thin-Section CT and Chest Radiography Scoring Systems

 
Similar to thin-section CT, chest radiographs (according to the ILO classification) showed that the workers exhibited a wide range of abnormalities. The workers were grouped on the basis of profusion: group 0 (0/0 or 0/1, n = 32), group Ia (1/0, n = 16), group Ib (1/1, n = 18), group Ic (1/2, n = 8) and group II (2/1 or greater, n = 8) (Table 2).

Although we found a statistically significant association between thin-section CT and chest radiographic findings (P < .01, Table 2), there were some relevant discrepancies. Overall, nine of 32 patients with extensive thin-section CT findings (groups D and E) were classified as disease-free according to the ILO scheme (group 0). However, better agreement was found in cases in which more extensive radiographic involvement was depicted: Six of eight patients with group II findings at chest radiography were classified as belonging to group E at thin-section CT (Table 2).

Pleural plaques were found in 46 of 82 subjects (56%) on chest radiographs and in 61 of 82 subjects (74%) on thin-section CT scans (McNemar test, P < .05). Discordant results were found in 17 subjects: In 16 of them, pleural plaques were identified only at thin-section CT. These abnormalities were typically multiple (38 of 46 subjects) and bilateral (53 of 61 subjects) at both chest radiography and thin-section CT, with a higher prevalence of diaphragmatic and chest wall plaques. There was a significant association between parenchymal disease extension on chest radiographs and presence of pleural plaques ({chi}2 test, P < .01). In contrast, the occurrence of pleural plaques was similar among the thin-section CT groups ({chi}2 test, P > .05).

Relationship between CT and Clinical and Physiologic Parameters
There was no correlation between the extent of abnormalities at thin-section CT and any of the clinical variables, including smoking history and years of asbestos exposure (Tables 3 and 4, respectively; P > .05). Similarly, there was no significant correlation between extent of abnormalities at thin-section CT and individual spirometric values. However, there was a significant association between frequency of abnormal test results (forced vital capacity and forced expiratory volume in 1 second values below the lower limit of normality) and more advanced thin-section CT (groups D and E) and radiographic involvement (groups Ic and II) ({chi}2 analysis, P < .05).


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

  		TABLE 3. Descriptive Statistics for Patients Grouped according to Thin-Section CT Findings

 

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

  		TABLE 4. Population Characteristics according to Chest Radiography (ILO classification of profusion)

 
Several indexes of impaired pulmonary gas exchange at rest and during exercise differed among groups, independent of the imaging method. Groups D and E (thin-section CT) and group II (chest radiography) presented lower DLCO values (percentage predicted) and higher exercise {Delta}P(A-a)O2/VO2 values than those of their counterparts with less severe disease (P < .01) (Tables 3, 4; Fig 1). On the other hand, differences in resting arterial blood gas tensions among groups did not reach statistical significance (Tables 3, 4).



View larger version (36K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Plots show individual values of pulmonary gas exchange variables at rest (DLCO) and during exercise ({Delta}P[A-a]O2/VO2) in asbestos-exposed workers grouped according to thin-section CT (HRCT) abnormalities ({bullet}) and chest radiographic abnormalities (according to the ILO classification, {circ}). Note the increased frequency of abnormal test results with progresive parenchymal involvement (groups A to E and groups 0 to II), independent of the imaging method. Dashed lines represent the limit for normality. See text and Table 2 for group definitions.

 
Prediction of Pulmonary Gas Exchange Impairment
A clinically significant pulmonary gas exchange impairment was defined as DLCO less than 70% predicted value (4,24) and/or {Delta}P(A-a)O2/VO2 more than 20 mm Hg · L · min–1 (4,25). According to these criteria, 23 patients had clinically significant pulmonary gas exchange impairment—that is, prevalence of 28% with pretest odds of 0.38 (prevalence/[1 – prevalence]).

The likelihood of impairment in gas exchange increased with parenchymal disease severity, independent of the imaging method (P < .05). Overall, 15 of 29 thin-section CT group D and E subjects had gas exchange impairment (positive predictive value, 51.7%)—that is, test sensitivity of 65.2% (15 of 23 subjects with impairment in gas exchange). On the other hand, 45 of 53 subjects in thin-section CT groups A, B, and C did not manifest this abnormality (negative predictive value, 84.9%)—that is, the test had a specificity of 76.3% (45 of 59 subjects without impairment in gas exchange). Therefore, the odds ratio for the presence of impairment in gas exchange was significantly increased in groups D and E; in contrast, probability of this negative outcome was reduced in group A (P < .05, Table 5, Fig 2). Consequently, thin-section CT results were not conclusive about impairment in gas exchange occurrence in the intermediate groups (groups B and C, 39% of subjects [32 of 82]). Interestingly, however, 81% of these workers (26 of 32) did not exhibit pulmonary gas exchange dysfunction (Table 5, Fig 2).


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

  		TABLE 5. Pulmonary Gas Exchange Impairment in Relation to Thin-Section CT and Chest Radiography Scores

 


View larger version (35K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Plots show relationship between selected indexes of resting (DLCO) and exercise-induced ({Delta}P[A-a]O2/VO2) gas exchange impairment in asbestos-exposed workers grouped according to thin-section CT (HRCT) and chest radiographic findings (according to ILO classification). There was a significant association between gas exchange impairment (shaded area) and more extensive CT (groups D and E) and radiographic (group II) involvement. Conversely, the probability of a negative functional outcome decreased with less severe disease (group A and group 0). Dashed lines represent the limit for normality. See text and Table 2 for group definitions.

 
In relation to chest radiography, six of eight group II subjects depicted impairment in gas exchange (positive predictive value of 75%)—that is, the test had a sensitivity of only 26% (six of 23 subjects with impairment in gas exchange). Conversely, 57 of 74 group 0 to group Ic workers did not manifest abnormal gas exchange (negative predictive value of 77%)—that is, the test had a specificity of 96.6% (57 of 59 subjects without impairment in gas exchange). The odds ratio for altered intrapulmonary gas exchange was significantly reduced in group 0; conversely, its probability was increased only in subjects with more extensive disease (group II; P < .05; Table 5, Fig 2). Chest radiography results were therefore equivocal in relation to impairment in gas exchange occurrence in 51% of the workers (42 of 82) (groups Ia, Ib, and Ic): Normal pulmonary gas exchange was found in 31% of them (Table 5, Fig 2).

In contrast with the results mentioned earlier, we were unable to find a significant association between pleural plaques and pulmonary gas exchange impairment that was independent of the imaging method ({chi}2 test, P > .05). As shown in Table 6, the positive predictive value and specificity for gas exchange dysfunction were particularly low for both methods. As a consequence, the odds ratios for gas exchange impairment according to the chest radiography and thin-section CT groups were not enhanced after addition of presence and/or extension of pleural plaques to the parenchymal abnormalities.


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

  		TABLE 6. Relationship between Pleural Plaques at Thin-Section CT and Chest Radiography with Pulmonary Gas Exchange Impairment

 
Gas Exchange Impairment and CT Abnormalities
We assessed the individual likelihood of the different types of thin-section CT abnormalities in the prediction of impairment in gas exchange. The category of parenchymal bands demonstrated the best combination of sensitivity and specificity for impairment in gas exchange prediction (65.2% and 71.2%, respectively), followed by subpleural nodules (60.8% and 62.7%), honeycombing (47.8% and 72.8%), architectural distortion (47.8% and 79.6%), and subpleural lines (43.4% and 71.2%). In fact, results of logistic regression analysis, which was used to consider all abnormalities, including pleural plaques, indicated that only parenchymal bands and subpleural nodules were significantly associated with impairment in gas exchange (P > .05) (Table 7).


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

  		TABLE 7. Final Logistic Regression Analysis Model with Pulmonary Gas Exchange Impairment as the Dependent Variable and CT Abnormalities as Independent Variables

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
 
Our results demonstrate that thin-section CT findings can be successfully used to estimate individual probability of impairment in gas exchange in workers exposed to asbestos, even when chest radiography results are equivocal. Therefore, thin-section CT might be valuable not only for early recognition of disease (11–17) but also for providing clinically useful estimates of respiratory impairment secondary to asbestos exposure.

CT and Asbestosis
Clinical diagnosis of asbestosis depends on a reliable history of exposure, an appropriate latency period between exposure and disease detection (usually longer than 15–20 years), and, crucially, presence of small irregular opacities on chest radiographs (profusion greater than or equal to 1/0, based on the ILO classification) (1,2,7). Because thin-section CT depicts fine morphologic alterations in the lungs, it has been used increasingly for early recognition of asbestos-related lung fibrosis and presence of pleural plaques (11–17).

Although the value of thin-section CT in the diagnosis of subtle interstitial disease in patients with suspected asbestosis and for evaluation of suspected lung masses seems to be well established, little is known about its usefulness in predicting the development of negative functional outcomes. Pulmonary gas exchange, in particular, is prone to be affected early by the morphologic derangements associated with parenchymal disease (4–6). The fibrotic process of asbestosis characteristically develops first in the peribronchiolar areas and then progressively involves the adjacent alveoli. Akira et al (21) have demonstrated that the initial parenchymal changes are subpleural isolated dotlike abnormalities connected to the most peripheral branch of the pulmonary artery system. Parenchymal bands then develop as fibrotic tissue along the bronchovascular sheath or interlobular septa, with distortion of the parenchyma. In this context, it is worth noting that the two thin-section CT abnormalities related more closely to impairment in gas exchange in the present study were subpleural nodules and parenchymal bands.

Functional Consequences of Asbestos Exposure
We can speculate that the aforementioned thin-section CT abnormalities were associated with ventilation-perfusion mismatch and/or thickening of the alveolar-capillary membrane; these conditions have long been recognized to impair the gas exchanging role of the lungs (32). This is particularly true with increased pulmonary blood flow, such as during exercise. In these circumstances, the decreased contact time between red blood cells and alveoli shortens the available time for gas exchange completion. Consequently, complete diffusion equilibrium might not develop, especially when mixed venous partial pressure for O2 is low (33).

Not surprisingly, an exercise-related index of pulmonary gas exchange impairment ({Delta}P[A-a]O2/VO2) proved to be more sensitive than resting variables in the indication of disease extension. In line with these findings, Smith and Agostoni (6) showed that a P(A-a)O2/VO2 value of more than 35 mm Hg · L–1 · min–1 of VO2 presented the highest discriminatory power in the differentiation of asbestosis from chronic obstructive pulmonary disease.

It should be noted that in the current study, we used {Delta}P(A-a)O2 and not only P(A-a)O2 to assess exercise-induced impairment in gas exchange (4,34). This value represents the exercise-induced physiologic modification—that is, "{Delta}" (exercise-rest). It should also be recognized that impairment in gas exchange was evaluated at an exercise intensity (lactate threshold, or {theta}L) that is likely to be representative of the metabolic demands associated with activities of daily living rather than a high exercise level (35). In this context, it is conceivable that evaluation of submaximal rather than maximal exercise may have negatively influenced the test sensitivity for detection of physiologic impairment.

In contrast with impairment in gas exchange, it has long been documented that reductions in lung volumes are markedly insensitive in the indication of disease extension and progression, especially in milder cases (36–38). Lung diffusing capacity values are also relatively well preserved in asbestosis when compared with other fibrosing lung diseases with similar volume restriction (5,39). In the present study, given the relatively mild extent of abnormalities in most subjects, it is not surprising that several indexes of resting functional impairment were altered only modestly. We interpret these results as evidence that rest-to-exercise changes in gas exchange efficiency should be valued particularly for early detection of the pathophysiologic consequences of interstitial fibrosis associated with lung deposition of asbestos fibers.

Clinical Value of CT and Chest Radiography in the Prediction of Gas Exchange Impairment
In the present study, the practical value of identifying specific thin-section CT features was demonstrated through the advantage of combining semiquantitative and qualitative findings to improve the discriminatory value of the classificatory scheme. Likelihood of clinically relevant impairment in gas exchange was increased when (a) interstitial disease extension and/or severity was found to be multifocal and there were three or more abnormalities (group D) and, in particular, (b) profuse bilateral alterations were described with four or more abnormalities (group E). In contrast, a slightly altered thin-section CT scan was predictive of normal pulmonary gas exchange (group A). Note that although the thin-section CT data could not be used to conclusively predict presence of impairment in gas exchange in subjects in groups B and C (n = 32), 26 of these patients did not manifest this functional abnormality. In contrast, chest radiography results were equivocal in relation to impairment in gas exchange occurrence in 42 subjects (groups Ia, Ib, and Ic): Normal pulmonary gas exchange was found in only 16 of these workers. These results suggest that if a thin-section CT scan is not classified as group D or E, probability of impairment in gas exchange at rest and during moderate exercise is small.

Another noticeable finding was that the likelihood of impairment in gas exchange was significantly higher in group D (extension and/or severity, grade II; three or more abnormalities) than in group C (extension and/or severity, grade III; three or fewer abnormalities). In other words, less profuse disease was occasionally more likely to indicate gas exchange impairment, provided that it was associated with a higher number of cumulative abnormalities. As mentioned, presence of parenchymal bands and subpleural nodules should be especially valued in this regard. Conversely, presence and extension of pleural plaques did not enhance the predictive role of the parenchymal thin-section CT findings for prediction of this functional abnormality (see Study Limitations).

In the study of Gamsu and co-workers (18), chest radiography (according to the ILO scheme) was used to predict asbestosis with less frequency than that with thin-section CT. Similar findings were reported by Jarad et al (40) and, more recently, by Huuskonen and colleagues (41); these investigators also developed scoring systems for asbestos-associated pulmonary fibrosis. The latter authors, for instance, found that a thin-section CT fibrosis score performed better than the ILO classification, especially for enhanced sensitivity in the detection of incipient disease (41).

Although a systematic comparison between thin-section CT and chest radiography was not the primary objective of our study, we also found that thin-section CT was used to identify parenchymal and pleural abnormalities more frequently than was chest radiography. In fact, nine patients with severe changes at thin-section CT (groups D and E) had normal chest radiographs; in these patients, the lesions were located in peripheral subpleural areas, especially in the posterior zones of the lower lobes.

Nonetheless, these findings should be analyzed with extreme caution. As mentioned in Materials and Methods, the study inclusion criterion was presence of abnormality at thin-section CT. It is clear, therefore, that this design feature provided a bias for thin-section CT to perform better—that is, sensitivity of thin-section CT in the detection of disease was necessarily higher than that for chest radiography. Despite this important limitation, our data add to the available evidence that thin-section CT is valuable in the identification of parenchymal and pleural abnormalities that can be missed at conventional chest radiography.

Study Limitations
This study has several important limitations in addition to that mentioned above. We restricted our analysis to presence and extension of pleural plaques. Future studies are therefore warranted to investigate whether inclusion of diffuse pleural thickening at thin-section CT adds to clinical prediction of impairment in gas exchange at submaximal exercise in this patient population (38,42,43). Pulmonary emphysema was not scored formally in the present study; although we failed to demonstrate an association between disease severity and smoking, this confounding factor may have influenced our results.

In addition, it was surprising that the predictive role of some well-known indexes of gross parenchymal destruction, such as honeycombing and architectural distortion, was not as substantial as that demonstrated for less severe lesions (subpleural nodules and parenchymal bands). This was probably related to the fact that those advanced abnormalities were limited in extent and had a patchy feature—at least in this sample of mostly mild to moderate cases. Future research should therefore be undertaken to analyze the predictive power of specific abnormalities to indicate gas exchange impairment in patients with more severe conditions and to assess whether it would be valid to add an extension score for each thin-section CT finding.

In conclusion, a dual semiquantitative and qualitative thin-section CT classification of lung parenchymal abnormalities (18) can be used successfully to estimate the individual likelihood of pulmonary gas exchange impairment at rest and during moderate exercise in workers with long-term exposure to asbestos.


   FOOTNOTES
 
Abbreviation: DLCO = diffusing capacity for carbon monoxide, ILO = International Labour Office

Author contributions: Guarantor of integrity of entire study, E.B.; study concepts and design, all authors; literature research, A.S., J.A.N., E.B.; clinical studies, A.S., S.G., L.E.N., J.K., R.T.R.; data acquisition and analysis/interpretation, all authors; statistical analysis, J.A.N.; manuscript preparation and definition of intellectual content, all authors; manuscript editing, J.A.N., L.E.N., N.M.; manuscript revision/review and final version approval, all authors


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
DISCUSSION
 REFERENCES
 

  1. Begin R. Asbestos-related diseases. Eur Respir Mon 1999; 11:158-177.
  2. Becklake MR. The epidemiology of asbestosis In: Mineral fibers and health. Miami, Fla: CRC, 1991; 104-119.
  3. Howard J, Mohsenifar Z, Brown HV, Koerner SK. Role of exercise testing in assessing functional respiratory impairment due to asbestos exposure. J Occup Med 1982; 24:685-689.[CrossRef][Medline]
  4. Risk C, Epler GR, Gaensler EA. Exercise alveolar-arterial oxygen pressure difference in interstitial lung disease. Chest 1984; 85:69-74.[Abstract/Free Full Text]
  5. Agusti AG, Roca J, Rodriguez-Roisin R, Xaubet A, Agusti-Vidal A. Different patterns of gas exchange response to exercise in asbestosis and idiopathic pulmonary fibrosis. Eur Respir J 1988; 1:510-516.[Abstract]
  6. Smith DD, Agostoni PG. The discriminatory value of the P(A-a)O2 during exercise in the detection of asbestosis in asbestos exposed workers. Chest 1989; 95:52-55.[Abstract/Free Full Text]
  7. International Labour Office. Guidelines for the use of ILO international classification of radiographs of pneumoconiosis Geneva, Switzerland: International Labour Office, 1980.
  8. Rosenstock L, Barnhart S, Heyer NJ, Pierson DJ, Hudson LD. The relation among pulmonary function, chest roentgenographic abnormalities, and smoking status in an asbestos-exposed cohort. Am Rev Respir Dis 1988; 138:272-277.[Medline]
  9. Oksa P, Huuskonen MS, Jarvisalo J, et al. Follow-up of asbestosis patients and predictors for radiographic progression. Int Arch Occup Environ Health 1998; 71:465-471.[CrossRef][Medline]
  10. Miller A, Lilis R, Godbold J, Wu X. Relation of spirometric function to radiographic interstitial fibrosis in two large workforces exposed to asbestos: an evaluation of the ILO profusion score. Occup Environ Med 1996; 53:808-812.[Abstract]
  11. Harkin TJ, McGuinness G, Goldring R, et al. Differentiation of the ILO boundary chest roentgenograph (0/1 to 1/0) in asbestosis by high-resolution computed tomography scan, alveolitis, and respiratory impairment. J Occup Environ Med 1996; 38:46-52.[CrossRef][Medline]
  12. Aberle DR, Gamsu G, Ray CS. High-resolution CT of benign asbestos-related diseases: clinical and radiographic correlation. AJR Am J Roentgenol 1988; 151:883-891.[Abstract/Free Full Text]
  13. Staples CA, Gamsu G, Ray CS, Webb WR. High-resolution computed tomography and lung function in asbestos-exposed workers with normal chest radiographs. Am Rev Respir Dis 1989; 139:1502-1508.[Medline]
  14. Aberle DR. High-resolution computed tomography of asbestos-related diseases. Semin Roentgenol 1991; 26:118-131.[CrossRef][Medline]
  15. Bégin R, Ostiguy G, Filion R, Groleau S. Recent advances in the early diagnosis of asbestosis. Semin Roentgenol 1992; 27:121-139.[CrossRef][Medline]
  16. Bégin R, Ostiguy G, Filion R, Colman N, Bertrand P. Computed tomography in the early detection of asbestosis. Br J Ind Med 1993; 50:689-698.[Medline]
  17. Neri S, Boraschi P, Antonelli A, Falaschi F, Baschieri L. Pulmonary function, smoking habits, and high-resolution computed tomography (HRCT) early abnormalities of lung and pleural fibrosis in shipyard workers exposed to asbestos. Am J Ind Med 1996; 30:588-595.[CrossRef][Medline]
  18. Gamsu G, Salmon CJ, Warnock ML, Blanc PD. CT Quantification of interstitial fibrosis in patients with asbestosis: a comparison of two methods. AJR Am J Roentgenol 1995; 164:63-68.[Abstract/Free Full Text]
  19. Cotes JE, Steel J. Work related lung disorders Oxford, England: Blackwell Scientific Publications, 1987.
  20. Webb WR, Muller NL, Naidich DP. High-resolution CT technique. In: Webb WR, Muller NL, Naidich DP, eds. High-resolution CT of the lung. New York, NY: Raven, 1992; 4-13.
  21. Akira M, Yamamoto S, Yokohama K, et al. Asbestosis: high-resolution CT–pathologic correlation. Radiology 1990; 176:389-394.[Abstract/Free Full Text]
  22. American Thoracic Society. Standardization of spirometry, 1994 update. Am J Respir Crit Care Med 1995; 152:1107- 1136.[Medline]
  23. Pereira CA, Barreto SP, Simões JG, Pereira FW, Gerstler JG, Nakatani J. Valores de referência para espirometria em uma amostra da população brasileira adulta. J Pneumol 1992; 18:10-22.
  24. Crapo RO, Morris AH, Gardner RM. Reference values for pulmonary tissue volume, membrane diffusing capacity, and pulmonary capillary blood volume. Bull Eur Physiopathol Respir 1982; 18:893-899.[Medline]
  25. Sue DY, Oren A, Hansen JE, Wasserman K. Diffusing capacity for carbon monoxide as a predictor of gas exchange during exercise. N Engl J Med 1987; 316:1301-1306.[Abstract]
  26. Jones NL, McHardy GJR, Naimark A, Campbell NL. Physiological dead space and alveolar-arterial gas pressure differences during exercise. Clin Sci 1966; 31:19-29.[Medline]
  27. Beaver WL, Wasserman K, Whipp BJ. A new method for detecting the anaerobic threshold by gas exchange. J Appl Physiol 1986; 60:2020-2027.[Abstract/Free Full Text]
  28. Reinhard U, Muller PH, Schmulling RM. Determination of anaerobic threshold by the ventilation equivalent in normal individuals. Respiration 1979; 38:36-42.[Medline]
  29. Bland JM, Altman DG. Statistics notes: the odds ratio. BMJ 2000; 320:1468.[Free Full Text]
  30. Altman DG. Diagnostic tests. In: Altman DG, Machin D, Bryant TN, Gardner MJ, eds. Statistics with confidence. 2nd ed. London, England: BMJ Books, 2000; 105-119.
  31. Hosmer DH, Jr, Lerneshow S. Applied logistic regression 2nd ed. New York, NY: Wiley, 2000.
  32. Poole DC, Musch TL. Pulmonary and peripheral gas exchange during exercise. In: Roca J, Rodriguez-Roisin R, Wagner PD, eds. Pulmonary and peripheral gas exchange in health and disease. New York, NY: Marcel Dekker, 2001; 69-123.
  33. Wagner PD. An integrated view of the determinants of maximum oxygen uptake. Adv Exp Med Biol 1988; 227:245-256.[Medline]
  34. Medinger AE, Khouri S, Rohatgi PK. Sarcoidosis: the value of exercise testing. Chest 2001; 120:93-101.[Abstract/Free Full Text]
  35. Wasserman K, Beaver WL, Whipp BJ. Gas exchange theory and the lactic acidosis (anaerobic) threshold. Circulation 1990; 81(suppl 1):II14-II30.
  36. Kilburn KH, Warshaw R, Thornton JC. Asbestosis, pulmonary symptoms and functional impairment in shipyard workers. Chest 1985; 88:254-259.[Abstract/Free Full Text]
  37. Sue DY, Oren A, James EH, Wasserman K. Lung function and exercise performance in smoking and nonsmoking asbestos-exposed workers. Am Rev Respir Dis 1985; 132:612-618.[Medline]
  38. Miller A, Bhuptani A, Sloane MF, Brown LK, Teirstein AS. Cardiorespiratory responses to incremental exercise in patients with asbestos-related pleural thickening and normal or slightly abnormal lung function. Chest 1993; 103:1045-1050.[Abstract/Free Full Text]
  39. Markos J, Musk AW, Finucane KE. Functional similarities of asbestosis and cryptogenic fibrosing alveolitis. Thorax 1988; 43:708-714.[Abstract]
  40. 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]
  41. Huuskonen O, Kivisaari L, Zitting A, Taskinen K, Tossavainen A, Vehmas T. High-resolution computed classification of lung fibrosis for patients with asbestos-related diseases. Scand J Work Environ Health 2001; 27:106-112.[Medline]
  42. 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]
  43. 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]




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
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 Sette, A.
Right arrow Articles by Müller, N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sette, A.
Right arrow Articles by Müller, N.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE