| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Thoracic Imaging |
1 From the School of Professional and Continuing Education, Dept of Radiology (G.C.O., P.L.K., H.N., I.W.T.H., F.L.C.), Depts of Diagnostic Radiology and Medicine (K.W.T.T., M.S.M.I., W.K.L., M.C.Y.), and Div of Traditional Chinese Medicine (T.F.C.), Univ of Hong Kong, Queen Mary Hosp, F/4, Block K, Pokfulam, Hong Kong SAR, China; and Hong Kong Wanchai Chest Clinic, Dept of Health (C.M.T.). Received May 15, 2002; revision requested Jul 26; final revision received Nov 22; accepted Jan 14, 2003. Supported by a CRCG grant from the Univ of Hong Kong and by the Hong Kong Pneumoconiosis Fund Board. Address correspondence to G.C.O. (e-mail: cgcooi@hkucc.hku.hk).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Seventy-six men with silicosis underwent volumetric and thin-section CT of the thorax. Lung function, Borg scale dyspnea grade, silica exposure duration, and cigarette consumption were determined. Nodular profusion (NP) at chest radiography was graded according to the International Labor Organization radiographic classification system; NP and PMF at CT were visually graded by using five-point (ie, grades 0–4) and four-point (grades 0–3) scales, respectively. Emphysema and NP, which together are defined as the NP index, were quantified by using attenuation threshold values of less than -950 HU and greater than -100 HU, respectively. Mean lung attenuation was also determined. Relationships among the CT, chest radiographic, and clinical parameters were analyzed by using Spearman correlation.
RESULTS: NP at chest radiography correlated (r > 0.50) with all CT parameters of nodularity. CT PMF had the highest correlation with emphysema (r = 0.58, P < .001). NP at chest radiography and all CT parameters were inversely related to lung function. At multiple regression analysis, PMF and emphysema index (both at CT) were significant determinants of forced expiratory volume in 1 second (FEV1) (P = .006 and .03, respectively) and FEV1 to forced vital capacity (FVC) ratio (P = .007 and .02, respectively). Mean lung attenuation remained related to FVC (P = .03), diffusing capacity of lung for carbon monoxide (P = .04), and Borg scale grade (P = .01). Cigarette consumption and silica exposure duration had no independent effects on lung function.
CONCLUSION: Qualitative and quantitative CT parameters can be used as indirect measures of functional impairment in silicosis. PMF and emphysema are independently related to airflow obstruction, whereas mean lung attenuation is related to clinical dyspnea and reduced lung volume.
© RSNA, 2003
Index terms: Emphysema, pulmonary, 60.751 • Lung, CT, 60.12111, 60.12118 • Lung, fibrosis, 60.6113 • Lung, function • Silicosis, 60.771
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Emphysema with silicosis is thought to be associated with progressive massive fibrosis (PMF), although other causative factors, such as smoking and exposure to dust such as that from coal and asbestos, often coexist in those who have silicosis (9–11). However, emphysema with silicosis has also been observed to occur independently of smoking (13–15). Some investigators (9–12) have reported that silicosis per se without PMF does not contribute to emphysema. In the ILO classification system, emphysema is not graded, but its appearance is noted. It has also been suggested that it is the degree of emphysema with silicosis rather than the severity of silicotic nodule profusion that determines the level of lung function (4,11,12). Some studies have addressed this issue by means of visual scoring of emphysema at CT according to the extent of lung involvement (3,4,10–12). Thus, the purpose of our study was to use qualitative and quantitative CT to test the hypothesis that impaired lung function with silicosis is due to PMF and associated emphysema.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
1/0 according to ILO classification). All patients gave their written informed consent. The period of recruitment was May 1997 to August 1998. The total number of patients consecutively invited to participate in the study was 200, 76 of whom consented to participate. Other inclusion criteria were the absence of asthma and unstable systemic diseases and a diagnosis of steady-state silicosis, which was defined on the basis of a less than 20% alteration in 24-hour sputum volume, forced expiratory volume in 1 second (FEV1), and forced vital capacity (FVC) and the lack of worsening respiratory symptoms for at least 3 consecutive weeks before recruitment. The study protocol was approved by the institutional ethics committee of the University of Hong Kong, Queen Mary Hospital. All patients underwent lung function examination and radiologic evaluation with CT and chest radiography, as outlined in the following text.
Radiologic Examinations
Posteroanterior chest radiographs were obtained with x-ray tube potentials of 100 kVp and grids on the same day that CT was performed. CT scans were obtained by using a commercially available scanner (CT HiSpeed Advantage; GE Medical Systems, Milwaukee, Wis). Nonenhanced volumetric CT scans (7-mm sections) were obtained with a 3.5-mm intersection interval from the apex to the base of the lung in full inspiration, and transverse CT scans (1- and 20-mm intervals) were reconstructed with a high-spatial-resolution (bone) algorithm.
CT scans were acquired at two window levels and widths appropriate for the lung parenchyma (-500/1,500 HU and -700/1,000 HU) and the mediastinum (50/250 HU). Lungs with markedly increased volume were imaged by using a -500/1,500 HU window setting, whereas other lungs were imaged by using a -700/1,000 HU window setting. Five CT parameters—nodular profusion, PMF, nodular profusion index (NPI, which comprises emphysema and nodular profusion), emphysema index (EI), and mean lung attenuation—were evaluated visually or quantitatively according to the protocol summarized in the following text. Attenuation thresholds were applied to quantify emphysema, nodular profusion, and mean lung attenuation (16–19).
Qualitative chest radiographic evaluation.—Chest radiographs were independently reviewed by two readers: a respiratory physician (C.M.T.) with a special interest in silicosis and a thoracic radiologist (G.C.O.). Both observers read the radiographs at different times and in different areas and graded the profusion of nodules according to the 1980 ILO classification system: Grade 0 means 0/-, 0/0, and 0/1; grade 1, 1/0, 1/1, and 1/2; grade 2, 2/1, 2/2, and 2/3; and grade 3, 3/2, 3/3, and 3/+ (1). When the two readers differed by at least one grade, the two readers read the radiograph of concern together and assigned a revised grade by consensus so that correlation analysis of the CT and clinical parameters could be performed.
Qualitative CT evaluation.—Two radiologists (G.C.O. and P.L.K.) who had no prior knowledge of the chest radiographic nodular profusion grades evaluated the volumetric CT and thin-section CT scans together to assess the degree of nodular profusion. The extent of PMF at CT was evaluated by using thin-section CT data that were transferred to a free-standing workstation (Windows Advantage; GE Medical Systems). The evaluations of nodular profusion and PMF at CT were performed separately from each other—12 months apart—and separately from the qualitative evaluation of the chest radiographs—at least 6 months apart. One reader (G.C.O.) performed both CT scan and chest radiograph evaluations and was blinded to the other radiologic grades assigned at each evaluation.
Each lung was divided into upper (apex to carina), middle (carina to inferior pulmonary vein), and lower (inferior pulmonary vein to lung base) zones. The system for grading nodular profusion at CT (Fig 1) involved the use of a modification of the scale described by Bergin and colleagues (4): Grade 0 meant no nodules; grade 1, a small number of nodules without vascular obliteration; grade 2, a larger number of nodules with mild vascular obliteration; grade 3, a large number of nodules with moderate vascular obliteration; and grade 4, a large number of nodules with severe vascular obliteration, with or without coalescence (<1.5 cm). The overall CT nodular profusion grade was derived from the sum of the grades assigned to each of the six lung zones (three zones of each lung).
|
|
|
|
|
|
|
|
To quantify emphysema and nodular profusion, attenuation threshold values of less than -950 HU and greater than -100 HU, respectively, were applied. The ratios of emphysema area to total lung area and nodular profusion area to total lung area for all six lung sections yielded the EI and NPI, respectively. Two experienced radiologists (G.C.O. and I.W.T.H.), blinded to the other CT and chest radiographic grades for the patients, performed the quantitative evaluations independently. G.C.O. quantified the CT scans obtained in 50 men, and I.W.T.H. quantified the CT scans obtained in the remaining 26 men.
Clinical Parameters
All clinical parameter data, including lung function, were obtained and collated (by K.W.T.T., T.F.C., M.S.M.I., and W.K.L.). Cigarette consumption and occupational silica exposure were quantified in terms of number of pack-years and number of years of exposure, respectively. We inquired about the job each man held during his lifetime work experience to determine the relative level of silica dust exposure, weighted to duration of job. Jobs involving rock cutting, drilling, and/or crushing and/or tunneling were judged to be associated with high relative dust exposure, and jobs involving unskilled general labor at construction sites and quarries were judged to be associated with low relative dust exposure (20).
Dyspnea was clinically graded by using the Borg 12-point scale (ie, grades 0, 0.5, and 1–10). Within 2 weeks of the radiologic examinations, lung function examinations were conducted by using a model 2200 system (SensorMedics, Yorba Linda, Calif) according to the standard protocol recommended by the American Thoracic Society (21). The following lung function parameters were measured: FEV1, FVC, FEV1 to FVC ratio (FEV1/FVC), total lung capacity (TLC), residual volume, and diffusing capacity of lung for carbon monoxide (DLCO). TLC and residual volume were measured by using the nitrogen washout method according to standard guidelines (22). DLCO was measured by using the single-breath DLCO procedure. Spirometric and lung volume parameters were expressed as percentages of predicted values based on the prediction equations of Da Costa (23), which were derived from a study population comprising 207 Chinese individuals. Prediction equations for diffusing capacity were based on the cotton dust standard (24,25). All lung function data, except FEV1/FVC values, were expressed as percentages of predicted values. Obstructive defect was defined as a defect involving an FEV1 of less than 80% of the predicted value and an FEV1/FVC of less than 70%, whereas restrictive defect was defined as a defect involving an FEV1 or FVC of less than 80% of the predicted value and an FEV1/FVC of greater than 80%.
Statistical Analyses
Statistical analyses were performed by using a computer software package (SPSS, version 10.0; SPSS, Chicago, Ill). Interobserver agreement on chest radiographic nodular profusion grades was evaluated by determining the Cohen 
statistic for the grades assigned. Patients were divided into simple silicosis and complicated silicosis groups on the basis of CT findings. Complicated silicosis—that is, PMF—was defined as the presence of an opacity or coalescence larger than 1.5 cm in diameter. Differences in lung function, cigarette consumption, silica exposure, Borg scale grade, and CT parameters between the patients with simple silicosis and those with complicated silicosis were evaluated by using t, Mann-Whitney rank sum, and 
2 tests, as appropriate. We also stratified the patients into four different categories—never, mild (<10 pack-years), moderate (10–19 pack-years), or heavy (>20 pack-years) cigarette smokers—to evaluate the relationship between smoking and other parameters at analysis of variance. The relationship between dust exposure and cigarette consumption was evaluated with the 2 test, and Yates correction was applied whenever applicable.
Univariate analyses with Spearman correlation were performed to evaluate the relationships among the clinical parameters (ie, Borg scale dyspnea grade, age, cigarette consumption, and silica exposure), lung function parameters, nodular profusion at chest radiography, and the five CT parameters (nodular profusion, PMF, NPI, EI, and mean lung attenuation). Multiple regression analyses were performed to determine the CT parameters that best predicted lung function—in terms of FEV1, FVC, FEV1/FVC, TLC, residual volume, and DLCO—and Borg scale dyspnea grade, adjusted for cigarette consumption and duration of silica exposure. NPI and nodular profusion at CT correlated highly with the other CT parameters (Table 1) and thus were omitted as independent variables. P < .05 indicated a statistically significant difference.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Table 2 summarizes the clinical, lung function, and radiologic measurements in the 76 men with silicosis. The mean FEV1, FVC, TLC, residual volume, and DLCO values were 68.1%, 82.1%, 81.1%, 86.7%, 140.0%, and 81.1% of the predicted values, respectively. The mean FEV1/FVC was 0.60. The mean Borg scale grade for dyspnea was 2.1 ± 1.4. There were no significant differences in age, duration of silica exposure, and cigarette consumption between the men with simple silicosis and those with PMF (ie, complicated silicosis) (P > .05). However, significant differences in Borg scale grade, lung function, and radiologic parameters were observed between the two groups. Those with complicated silicosis had significantly lower FEV1, FVC, FEV1/FVC, TLC, and DLCO values than those with simple silicosis (P < .05). In addition, the complicated silicosis group consisted of a significantly higher proportion of men with abnormal FEV1, FVC, and FEV1/FVC values (<70% of predicted values) and higher grades for all radiologic parameters of silicosis, compared with the simple silicosis group. About a third of the men in the simple silicosis group also had abnormal FEV1 and FVC values.
|
|
|
value was 0.77 (P < .001) for the two readers’ grading of nodular profusion on chest radiographs, indicating good agreement. The two readers disagreed by one grade category in the reading of 10 chest radiographs. Disagreement occurred with regard to grades 1 and 2 in five cases and with regard to grades 2 and 3 in an additional five cases. After joint evaluation of these 10 cases, reader 1 upgraded the nodular profusion grade on a chest radiograph from 1 to 2, upgraded the grades on two chest radiographs from 2 to 3, and downgraded the grades on three chest radiographs from 2 to 1. Reader 2 upgraded the nodular profusion grades on two chest radiographs from 2 to 3 and downgraded the grade on one chest radiograph from 2 to 1 and the grade on another chest radiograph from 3 to 2. Ten chest radiographs (obtained in 10 patients) that were judged to be depicting simple silicosis were found at CT to actually be depicting PMF—that is, complicated silicosis (Fig 3). The CT PMF grades for these 10 patients were 1 (n = 2), 2 (n = 6), 4 (n = 1), and 5 (n = 1). In the 58 patients with PMF at CT, characteristic paracicatricial emphysema around the PMF and reduced upper lobe volumes were seen. The lower lobes were hyperlucent and enlarged and had reduced vascularity.
|
|
|
|
|
Intercorrelations between Chest Radiographic and CT Parameters
The data in Table 1 show that there were significant correlations between nodular profusion at chest radiography and various CT parameters. NPI and nodular profusion at CT correlated strongly with other CT parameters (PMF and mean lung attenuation) and were omitted as independent variables in the logistic multiple regression analyses described in the next paragraph.
Multiple Regression Analysis
The results of the regression analysis are summarized in Table 6. PMF grade and EI at CT were found to be the best independent determinants of FEV1, FEV1/FVC, and TLC. Mean lung attenuation was the best determinant of FVC, DLCO, and Borg scale dyspnea grade. Neither duration of silica exposure nor cigarette consumption had an independent influence on the lung function or clinical parameters, with the exception that cigarette consumption affected DLCO.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
To measure silicosis-related lung changes, we used five CT parameters: nodular profusion, which was evaluated visually without grading of PMF; PMF, which was evaluated visually; NPI, which is a quantitative measure of nodular profusion and PMF; EI, which is a quantitative measure of emphysema; and mean lung attenuation, which is a quantitative measure of the overall effects of emphysema, nodules, and PMF. Because there is no consensus on the sizes of PMFs, in this study we defined PMF as an opacity or coalescence larger than 1.5 cm in diameter. According to the ILO radiographic classification system, PMF is characterized by the presence of a single opacity larger than 1 cm in diameter on a radiograph (1,26). According to the criteria for coal workers’ pneumoconiosis outlined by the College of American Pathologists, however, PMF is present when there is a lesion larger than 2 cm in diameter (27). In previous studies in which CT was used to grade lung nodularity in silicosis, including the present study, the ILO radiographic classification system or a modification (3,8–11,28) of the nodular profusion classification system described by Bergin et al (4) has largely been applied.
In our study, both the chest radiographic parameters and the CT parameters of nodularity were inversely related to lung function indexes. The chest radiographic and CT grades of nodular profusion had similar relationships with indexes of airflow obstruction and diffusing capacity, albeit with lower correlation coefficients, particularly with respect to FEV1 and FEV1/FVC. On the other hand, PMF at CT showed excellent correlations with these two indexes of airflow obstruction.
The two quantitative measures of nodular profusion, NPI and mean lung attenuation, were assessed directly (NPI) by using threshold values to isolate areas of attenuation greater than -100 HU and indirectly (mean lung attenuation) by using mean attenuation values. Although NPI had good correlation with lung function, it was excluded from the final regression analysis because it is a measure of both PMF and nodular profusion—PMF in particular—and because it was strongly correlated with both of these CT parameters. We believed that these two attributes warranted the exclusion of NPI in the regression analysis performed to determine the CT parameters that best predicted lung function abnormalities. Nodular profusion at CT also had to be excluded from the regression analysis because of its high colinearity with the other CT parameters.
Previous study results have suggested that there are problems associated with using mean lung attenuation as an objective measure to quantify silicosis (4). These problems are reportedly due to the unequal distribution of lung attenuation between the upper and lower lung lobes, particularly with complicated silicosis, which causes very dense upper lobes. It has been suggested that the mean attenuation of the entire region of the upper lobes be measured to avoid observer bias. Our finding that mean lung attenuation, derived from the mean of three CT lung sections, was an independent predictor of FVC and DLCO suggests that mean lung attenuation is a useful indirect measure of the overall effects of the pathologic changes associated with silicosis and a representative estimation of lung attenuation in silicosis.
Some study findings have suggested that simple silicosis without superimposed emphysema has no influence on lung function and should be considered as only a marker of exposure rather than an actual disease (4,19). Others, like us, however, have observed a more severe reduction in lung function with increasing nodular profusion and PMF and a milder reduction in lung function associated with simple silicosis (3,4,10,11,28). Although the concept of coalescence and PMF as main contributors to the functional defects in silicosis is not new (3,4,10,11,28), the present study results show PMF to be an independent determinant of airflow obstruction. The scarring associated with PMF results in the contraction of the adjacent lung and the development of cystic spaces or paracicatricial emphysema (26,29).
Although the CT technique that we used to quantify low-attenuation areas does not enable definition of microscopic emphysema, it enables quantification of the extent of emphysema (12,19). We believe that it is interesting that although emphysema was independently associated with reduced FEV1 and FEV1/FVC values, it was not independently related to diffusing capacity. This observation suggests that the quantified emphysema corresponded more closely to distal airspace enlargement, represented by the low-attenuation areas, than to airspace destruction (17,30). Although paracicatricial emphysema adjacent to PMF is characteristic of silicosis, it may be difficult to differentiate panlobular emphysema from hyperinflation in the lower lung lobes with visual evaluation and quantitative CT (31). Panlobular emphysema has also been misclassified in the pathologic assessment of emphysema (32).
There are several reasons for airflow obstruction in patients with silicosis, including smoking, airway hyperresponsiveness, and silica exposure itself (14,33–35). It has been postulated that airflow obstruction due to silica exposure is related to airway compression and hyperinflation that are induced by peribronchiolar scarring and to hypertrophy and scarring of bronchial-associated and intrapulmonary lymph nodes (13). The attributing factors of airflow obstruction with silicosis have been a topic of considerable debate (2,3,10–12,33,36,37). This debate is compounded by the fact that the majority of patients with silicosis are smokers, and, thus, the effects of smoking on lung function are hard to distinguish from the effects of silicosis on lung function.
The results of a meta-analysis suggest that silica exposure is associated with airflow obstruction, even in nonsmokers (13–15,37). Our study results show that neither duration of silica exposure nor cigarette consumption had an effect on lung function. The never smokers in our study had lung function impairment as severe as that of the heaviest smokers (>20 pack-years). This anomaly could have been due to inaccuracies in the smoking and dust exposure history data, despite meticulous collection efforts. Other possible explanations include the increasing susceptibility of never smokers to silica dust or the substantial exposure of the never smokers in our series to dust, the majority (70%) of whom had occupations in which they were exposed to relatively high dust concentrations.
Although cigarette smoking is detrimental to the airways and the single most important cause of chronic obstructive pulmonary disease, it is reported that only 10%–20% of smokers develop clinically important chronic obstructive pulmonary disease and that half of these individuals never develop clinically important lung function impairment (38,39). Therefore, the lack of correlation between cigarette consumption and lung function impairment might suggest that silicosis had a more substantial role in lung function impairment than did smoking among the patients in our series.
In conclusion, our qualitative and quantitative CT assessment of the lungs in patients with silicosis revealed important morphologic and functional correlations. Our study findings reemphasize the clinical importance of PMF and its association with emphysema as important independent determinants of airflow obstruction in silicosis. The results of this study suggest that emphysema with silicosis represents distal airspace enlargement rather than lung destruction. The presence of PMF could be used as an indicator of the disease severity and disability with silicosis. Mean lung attenuation could also be used as an indicator of lung restriction in workers exposed to silica. Although we do not propose the replacement of either chest radiography for the diagnosis of silica exposure or lung function examination for determining the severity of disability, CT could be used to indirectly quantify morphologic disease and functional impairment.
|
ACKNOWLEDGMENTS |
|---|
Author contributions: Guarantors of integrity of entire study, W.K.L., F.L.C., H.N.; study concepts, G.C.O., K.W.T.T., W.K.L., H.N.; study design, G.C.O., K.W.T.T., T.F.C.; literature research, G.C.O., P.L.K., I.W.T.H.; clinical studies, K.W.T.T., W.K.L., T.F.C., M.S.M.I.; data acquisition, T.F.C., G.C.O., I.W.T.H., K.W.T.T; data analysis/interpretation, G.C.O., T.F.C., C.M.T., P.L.K., I.W.T.H.; statistical analysis, G.C.O., P.L.K., M.C.Y., K.W.T.T.; manuscript preparation, G.C.O., P.L.K., M.C.Y., C.M.T.; manuscript definition of intellectual content, M.C.Y., W.K.L., H.N., F.L.C.; manuscript editing, M.C.Y., G.C.O., K.W.T.T.; manuscript revision/review, G.C.O., M.C.Y., P.L.K., K.W.T.T., H.N.; manuscript final version approval, G.C.O., M.C.Y., P.L.K., K.W.T.T.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, W.K.L., F.L.C., H.N. The complete list of author contributions appears at end of article.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Raoof, A. Amchentsev, I. Vlahos, A. Goud, and D. P. Naidich Pictorial Essay: Multinodular Disease: A High-Resolution CT Scan Diagnostic Algorithm Chest, March 1, 2006; 129(3): 805 - 815. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Arakawa, P. A. Gevenois, Y. Saito, H. Shida, V. De Maertelaer, H. Morikubo, and M. Fujioka Silicosis: Expiratory Thin-Section CT Assessment of Airway Obstruction Radiology, September 1, 2005; 236(3): 1059 - 1066. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Arakawa, K. Honma, Y. Saito, H. Shida, H. Morikubo, N. Suganuma, and M. Fujioka Pleural Disease in Silicosis: Pleural Thickening, Effusion, and Invagination Radiology, August 1, 2005; 236(2): 685 - 693. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
