Occupational and Entvronmental Medicine 1994;51:239-243

Findings from high resolution computedtomography of the lung and pleura of symptomfree workers exposed to amosite who had normalchest radiographs and pulmonary function tests

Silvano Neri, Alessandro Antonelli, Fabio Falaschi, Piero Boraschi, Lidio Baschieri

AbstractThe lungs of 50 symptom free workersexposed to amosite and with normal pul-monary function tests were examined byhigh resolution computed tomography(HRCT). Twenty five had normal stan-dard chest radiographs whereas theother 25 had radiographs interpretedas near normal (International LabourOffice profusion score <01 or suspectedpleural plaques). In 13 of the workers theresults of HRCT were negative; in 22pleural plaques were found, in five therewas only parenchymal involvement, and10 had both pleural and parenchymalchanges. The mean duration of exposureto amosite was significantly longer forthe subjects with parenchymal signs thanfor those with normal parenchyma andfor the workers with pleural plaques thanfor those with normal pleura and lungparenchyma. The prevalence of identi-fied pleural and parenchymal abnormali-ties in the 50 workers was alsosignificantly higher than in a referencegroup without exposure to asbestos. It isconcluded that HRCT may detect initiallung and pleural involvement in symp-tom free workers exposed to amosite andthe mean duration of exposure is longerfor subjects with parenchymal or pleuralinvolvement.

(Occup Environ Med 1994;51:239-243)

PostgraduateOccupationalMedicine SchoolS NeriA AntonelliF FalaschiL BaschieriInstitute of Radiology,University ofPisa,ItalyF FalaschiP BoraschiRequests for reprints to:Alessandro Antonelli MD,Postgraduate OccupationalMedicine School, Instituteof Clinical Medicine H,University of Pisa, ViaRoma 67, 56100 Pisa, Italy.Accepted 13 September1993

Pulmonary diseases related to asbestos are

important medical concerns and legal issues.In the absence of histological examination theclinical diagnosis of asbestosis in workersexposed to asbestos is currently based on thefollowing criteria': (1) chest radiographic evi-dence of small irregular opacities with a pro-fusion of at least 1/1 according to theInternational Labour Office classification2;(2) a restrictive pattern of lung impairment;(3) a reduction of CO diffusing capacity; (4)the presence of bilateral late or paninspiratorycrackles on auscultation at the posterior lungbases not cleared by cough. The findingsfrom chest radiography are thought to be themost important and the presence of pleuralabnormalities could help to indicate that a

patient with pulmonary fibrosis has asbes-tosis. It is well known that chest radiographs

may underestimate the presence of bothinterstitial lung involvement and pleuralplaques. In fact pathological changes ofasbestosis-5 have been found in the absenceof radiographically evident fibrosis or abnor-malities of pulmonary function. The issue ofthe earliest lesions of asbestosis and theirsignificance remains under debate.78 Highresolution computed tomography (HRCT)has now provided a sensitive tool to showinitial lung parenchymal fibrosis and pleuralasbestos related lesions in symptomaticworkers exposed to asbestos,9 even whenstandard chest radiographs are negative. Thespecificity of the HRCT findings in asbestosisis, however, still a matter of controversy.The aim of our perspective study was to

investigate the possibility of detecting earlyasbestos related lung and pleural lesionswithin a cohort of symptom free workersexposed to amosite who had negative chestradiographs, and CO diffusing capacity andpulmonary function tests in the normal range.Furthermore we evaluated the relationbetween the duration of exposure to amositeand the prevalence of abnormalities detectedby HRCT in the study group and thespecificity of HRCT findings, by comparisonof the exposed group with a control groupcomposed of patients without any knownprofessional exposure to asbestos.

MethodsAll the subjects of the study group worked ina factory producing naval furniture where onecase of mesothelioma'° had been previouslyfound by the sentinel health event method. Inthis factory a material containing about 40%amosite was used, from 1958 to 1980. Thiswas especially in the carpentry departmentwhere panels were built consisting of asbestosinsulation between two external layers ofwood. The study group comprised 127 work-ers whose chest radiographs, performedaccording to ILO criteria, were previouslyjudged as negative by outside readers. Threeradiologists experienced in the reading ofradiographs according to the ILO classifica-tion have re-examined the 127 standard chestradiographs.Among the 127 workers we selected the

subjects who fitted the following criteria: (1)documented occupational exposure to asbestos;(2) absence of any clinical symptomatology

239

on June 17, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.4.239 on 1 April 1994. D

ownloaded from

Neri, Antneli, Falaschi, Boraschi, Baschieri

Table Results ofpulmonary function tests (% ofpredicted values) in workers withnegative or positive chest radiograph

TLC VC FEV1 FEV, IFVC % TLco sb

Negative 98-3 (9.2) 96-5 (5.2) 94 7 (7 9) 98-6 (6-3) 118-4 (11-3)Positive 101.6 (7-1) 98-4 (8 6) 92-8 (8 3) 96-2 (4 7) 123-2 (21-2)

Results are means (SD); TLC = total lung capacity; VC = vital capacity; FEV, = forced expira-tory volume in one second; FVC = forced vital capacity; TLco = carbon monoxide transferfactor.

suggestive of asbestosis or of any other lungdisease; (3) normal pulmonary function tests(TLC, VC, FVC, FEV1, FEV1/FVC), andTLCO (single breath) (see table) not morethan four weeks before HRCT; pulmonaryfunction tests were performed by HP 47120A equipment that meets American ThoracicSociety recommendations" and reference val-ues were obtained from previous Italian stud-ies'2 13; (4) workers' consent to perform theHRCT.Among this group we selected 50 workers:

25 who had a plain chest radiograph con-firmed negative, both for pleura andpulmonary parenchyma, by our three inde-pendent radiologists and 25 with radiographsfor which an ILO profusion score of 0/1 orpleural plaques were suspected by two orthree of our readers and who were considered"positive". The mean age of the 50 subjects(49 men) was 45-7 (SD 6 9) (range: 31-61)years.

Occupational history was obtained bydirect interview and by the evaluation ofevery document available from the factorymedical service before HRCT examinations.Exposure is expressed in cumulative years.

Figure I Prone HRCT scan showing a subpleural cuilinear shadow parallel to innerchest wall (see arrows).

Individual health histories were also investi-gated: none of the workers had a history ofpleurisy, empyema, haemothorax, tuberculo-sis treated with pneumothorax, or other pul-monary diseases.The interval between HRCT and standard

chest radiograph varied from 0-2 to 19months (mean interval 44 (3 8) months).The HRCT study was performed on aGE-9800 scanner. All patients were exam-ined at suspended maximum inspiration,both in the supine position and prone whenposteroinferior portions of the lungs showedobscuration dependent on gravity attenuationor any of the HRCT abnormalities associatedwith asbestosis were identified. HRCT sec-tions 1'5 mm thick were made at 20 mmintervals from the aortic arch to thediaphragm. To minimise the noise 140 KVpand 170 mAs were used, with 2-5 seconds ofacquisition time. The image was finallyobtained through a reconstruction with aconvolution filter of the bone detail algory-thm. FoV was 24/30 cm. All images werephotographed at the 500 HU window forpleura and mediastin and 1500 HU windowfor pulmonary parenchyma. HRCT scanswere interpreted by a radiologist blinded toindividual exposure data and to chest radi-ographic interpretation. Pleural alterationswere quantified applying 1980 ILO criteriato the reading of CT scans. Parenchymalabnormalities were interpreted on the basesof published data'4 and were considered diag-nostic of asbestosis when present in a combi-nation of findings (two or more).HRCT findings have been related to the

individual exposure data and compared withthose of a reference group of 20 hospitalpatients (mean age 47 1(13-4), range 26-67years; 19 men) who were given HRCT forsuspected localised pulmonary pathology (inmost of cases nodules cysts, or embolus) notconfirmed by HRCT, without occupationalor significant environmental exposure toasbestos.

Smoking histories were not different forthe subjects exposed to asbestos with negativeor positive chest radiographs and the controlgroup (pack-years respectively 12-1(9-2),13-6(8-4), and 14-8(9-7)).

Statistical analysis was by analysis of vari-ance (ANOVA-Fisher PLSD), independent ttest (two tailed), or contingency table analy-sis, as indicated. The results are expressed asmeans (SD). A statistically significant levelwas considered to be present at p <0 05.The protocol of this study was approved by

the University Ethics Committee for HumanStudies.

ResultsAmong the 50 workers examined by HRCT,22 showed pleural plaques, five exclusivelyparenchymal abnormalities, and 10 bothpleural and parenchymal abnormalities; in theother 13 workers no pathological finding wasshown.

240

on June 17, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.4.239 on 1 April 1994. D

ownloaded from

HRCTfindings of the lung and pleura ofsymptom free workers exposed to amosite

Figure 2 HRCT scan photographed at a lung window, showing thickened interlobularseptae in the right lung (see arrows).

Figure 3 HRCT scan, photographed at a mediastinal window, showing measure ofpleural thickening in the costovertebral space (see arrows).

Figure 4 Prevalence (byHRCT) of normalfindings, and pleural,parenchymal, and bothpleural and parenchymallesions among workersexposed to amosite andcontrols.

80- 80Exposed group

60 =.... E;Control group

8- 40

20 i

HRCT: Normal Pleural Parenchymal Pleural andlesions lesions parenchymal

lesions

The parenchymal abnormalities were: sub-pleural curvilinear lines (eight cases; fig 1);thickened interlobular (septal) lines (threecases; fig 2); hazy patches of increased atten-uation (six cases); and thickened intralobular(core) lines (three cases). In four workers twoor more of the described signs were found.No case of honeycombing was reported.

Thirty two workers had pleural plaques (fig3) (23 of them bilateral). Only one workershowed diffuse pleural thickening (the samesubject also had interstitial abnormalities).No one had benign exudative pleurisy. Innine subjects, plaques were calcified. Plaqueslocated in the tendinous part of thediaphragm were present in 21 cases (one ofthese had exclusively diaphragmatic plaques).Among the 25 subjects whose standard

radiographs were confirmed "negative" byour radiologists, HRCT showed 11 cases ofpleural plaques, two cases with both pleuraland parenchymal involvement, and two caseswith only a subpleural curvilinear shadow,which is not specific for asbestosis as an iso-lated finding. Among the workers for whompleural plaques were suggested at standardchest radiographic reading (18 cases), HRCTconfirmed these in 16 cases, in five of whomHRCT also showed parenchymal involve-ment; in two cases pleural plaques were notconfirmed and no pathological findingemerged. HRCT confirmed parenchymalinvolvement for five of the six subjects whoseradiographs showed parenchymal opacitieswith a profusion 0/1 and in two of these casesnot shown as having pleural plaques; in onecase no pathological finding was shown andin one case a standard radiograph and HRCTshowed both pleural and parenchymalinvolvement.

Sensitivity, specificity, accuracy, and posi-tive and negative predictive values of standardchest radiographs were calculated withrespect to HRCT (artificially assuming it asgold standard) and resulted in, respectively,for pleural and parenchymal findings: sensi-tivity 89% and 40%; specificity 53% and97%; accuracy 66% and 80%; negative pre-dictive value 52% and 79%; and positive pre-dictive value 89% and 86%.

In the control group (20 subjects) HRCTidentified five cases with pleural plaques (onlyin one case bilaterally; all these were 1A onthe ILO scale); parenchymal bands werefound in one subject. Lung and pleurallesions detected by HRCT in the controlgroup were significantly less than those foundamong the group exposed to amosite (contin-gency table analysis: p <0 001; fig 4).The mean duration of exposure of the

whole group of workers exposed to amositewas 6-4 (5-5) (range: 0.2-21) years and themean latency since the time of first exposureto the examination was 20-7 (5-0) years. Inthe different subgroups the mean duration ofexposure (fig 5) was 2-3 (3-0) years for sub-jects with normal HRCT (group 1); meanlatency 18-3 (5.2) years; 7-0 (5-5) years forthe workers with pleural alterations (group2); mean latency 21-1 (4-7) years; 5X2 (3-0)

241

on June 17, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.4.239 on 1 April 1994. D

ownloaded from

Ne2N, Anwnelli, Falaschi, Boraschi, Baschieri

Figure 5 Mean (bar =SD) duration of exposurefor the subgroups ofworkers (defined on thebases ofHRCT results).

20j

15

C]10

0

5

0NormalHRCT

Pleural Parenchymal Pleural andlesions lesions parenchymal

lesions

years for those with exclusively parenchymalabnormalities (group 3); mean latency 208(6.1) years; and 11-0 (5 2) years for thosewho presented both pleural and parenchymalinvolvement (group 4); mean latency 23-0(4*0) years.The duration of exposure to amosite for

the workers of group 4 and group 2 was sig-nificantly longer than for the workers ofgroup 1 (ANOVA Fisher PLSD: p < 0005and p < 005, respectively). Also, consideringparenchymal involvement separately, the sub-jects with HRCT findings suggestive of pul-monary fibrosis (group 3 + group 4) hadbeen exposed to amosite for a longer timethan workers with normal parenchyma (group1 + group 2): 9 1 years v 5-3 years (p <005; t independent test).

DiscussionHRCT of the lung is probably the most sensi-tive current radiological means to detect earlyparenchymal and pleural asbestos related dis-eases, and has proved to be more sensitivethan chest radiographs"5 or conventionalCT.16 Our data confirm that HRCT is verysensitive and suggest that the accuracy ofstandard chest radiographs is not high insymptom free workers even if expert readersare involved. The specificity of HRCT find-ings seems to be confirmed by the high preva-lence of detected lesions among workersexposed to asbestos9 and by the proof thatHRCT findings correlate well with patho-logical findings,'7 even if it remains a contro-versial issue due to the limited number ofcorrelative pathological studies. Our datafrom the symptom free amosite exposedworkers in comparison with the controls sug-gest a probable relation between HRCT find-ings and exposure to asbestos. Pleural andparenchymal findings among controls may bedue to causes other than exposure to asbestosor a possible occupational exposure toasbestos not known from individual healthhistories. The presence of low levels ofasbestos fibres in the general environmentalso seems to be related to development ofparenchymal or pleural abnormalities.Whereas exposure-response relations for themajor asbestos related diseases have beenconsidered demonstrable,' the correlation

between the early lung lesions related toasbestos as detected by HRCT and the levelsand duration of exposure has actually not yetbeen proved. Staples et a19 did not find a sig-nificant difference in the mean duration ofexposure between a group of workers withnormal or near normal HRCT for lungparenchymal findings and those with abnor-mal HRCT suggestive of asbestosis, althoughthe mean exposure was higher for the subjectswith pleural plaques compared with workerswith normal HRCT. Al Jarad et al19 alsofailed to find any significant correlationbetween lung findings from HRCT andasbestos exposure in patients with lunginvolvement. In our study duration of expo-sure to amosite has been shown to be higherfor subjects with pleural or lung parenchymalinvolvement than for subjects with normalHRCT scans. These results indicate that theprevalence of parenchymal lesions shown byHRCT and the severity of lung involvementare dose related, by contrast with the previousfindings. There are various explanations forthe discrepancy. Our subjects were clinicallysymptom free whereas many of the subjectsevaluated by Staples et al9 presented withdyspnoea or abnormalities of pulmonaryfunction and the patients evaluated by AlJarad et al'9 had asbestosis. The mean age ofthe subjects in our study, their mean durationof exposure to asbestos, and the latency timessince first exposure were also lower than inthe previous studies; these data may explainwhy in our cohort only four cases of radio-logically confirmed asbestosis were present.Furthermore our sample of workers had beenexposed to the same type of asbestos fibre(amosite) in the same occupations, so thequality and severity of exposure at anymoment was not much different from onesubject to another, and the differences in theamount of exposure were almost exclusivelydue to differences in duration of exposure.

In conclusion our study suggests thatHRCT can show pulmonary or pleuralinvolvement in asbestos exposed workersbefore the presence of a clinical symptomatol-ogy and may allow the detection of asbestosisin some clinically symptom free workers. Infact in our study HRCT showed a combina-tion of parenchymal abnormalities sufficientto be considered diagnostic for asbestosis infour symptom free subjects with normal pul-monary function tests and diffusing capacity(in a preclinical phase). Furthermore, 11cases of pleural plaques and two cases withboth pleural and parenchymal involvementwere found by means of HRCT among the25 workers with normal chest radiographs.The lung and pleural abnormalities detectedby HRCT in our sample of symptom freeworkers seem to be related to asbestos expo-sure also in the subjects with negative orequivocal chest radiographs, as in most clini-cally evident asbestos related lung diseases.20The relevance of isolated slight pleural orparenchymal HRCT findings in the clinicallysymptom free workers with normal pul-monary function tests is not yet clear and

242

on June 17, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.4.239 on 1 April 1994. D

ownloaded from

HRCTfindings of the lung and pleura ofsymptom free workers exposed to amosite

there is no evidence that isolated parenchy-mal or pleural findings may progress and pro-duce clinical disability. We suggest a carefulsurveillance of these workers, however, andwe propose to re-evaluate them after a periodlong enough to assess the progression of suchabnormalities.

1 American Thoracic Society-Medical Section of theAmerican Lung Association. The diagnosis of nonma-lignant diseases related to asbestos. Am Rev Respir Dis1986;134:363-8.

2 International Labour Office. Guidlines for the use of ILOinternational classification of radiographs of pneumoco-nioses: revised 1980. Geneva; ILO 1980. (Occupationalsafety and health series No 22.)

3 Rockoff SD, Schwartz A. Roentgenographic -under-estimation of early asbestosis by International LaborOrganization classification. Analysis of data and proba-bilities. Chest 1988;93:1088-91.

4 Kippen HM, Iills R, Suzuki Y, Valciukas JA, Selikoff UJ.Pulmonary fibrosis in asbestos insulation workers withlung cancer: a radiological and histopathological evalua-tion. BrJtInd Med 1987;44:96-100.

5 Rubino GF, Scansetti G, Pira E, et al. Pleural plaques andlung asbestos bodies in the general population: anautoptical and clinical radiological survey. L4RC SciPubl 1980;30:545-51.

6 Bader ME, Mader RA, Tierstein AS, Miller A, Selikoff IJ.Pulmonary function and radiographic changes in 598workers with varying duration of exposure to asbestos.Mt SinaiJIMed 1970;37:492-500.

7 Churg A. The diagnosis of asbestosis. Hum Pathol 1989;20:97-9.

8 Smith DD. What is asbestosis? Chest 1990;98:953-4.9 Staples CA, Gamsu G, Ray CS, Webb WR. High

resolution computed tomography and lung function inasbestos-exposed workers with normal chest radio-graphs. Am Rev Respir Dis 1989;139:1502-8.

10 Neri S, Iaia TE, Roselli MG, Battista G. Gli eventi sen-tinella in medicina del lavoro: l'esempio dell'USL 1della Regione Toscana. Epidemiol PJev 1989;39:29-34.

11 American Thoracic Society. Standardization of spiro-metry: 1987 update. Am Rev Respir Dis 1987;136:1285-98.

12 Paoletti P, Pistelli G, Fazzi P, et al. Reference values forvital capacity and flow-volume curves from a generalpopulation study. Bullettin Europeen de PhysiopatologieRespiratoire 1986;22:451-9.

13 Paoletti P, Viegi G, Pistelli G, et al. Reference equationsfor the single breath diffusing capacity. A cross sectionalanalysis and effect of body and size. Am Rev Respir Dis1985;132:806-13.

14 Aberle DR, Gamsu G, Ray CS, Feuerstein IM. Asbestosrelated pleural and parenchymal fibrosis. Detection withhigh-resolution CT. Radiology 1988;166:729-34.

15 Friedman AC, Fiel SB, Fisher MS, Radecki PD, Lev-Toaff AS, Caroline DF. Asbestos-related pleural diseaseand asbestosis: a comparison of CT and chest radio-graphy. American Journal of Roentgenology 1988;150:269-75.

16 Lynch DA, Gamsu G., Aberle DR. Conventional and highresolution computed tomography in the diagnosis ofasbestos related diseases. Radiographics 1989;9:523-51.

17 Akira M, Yamamoto S., Yokoyama K, et al: Asbestosis:high-resolution CT-pathologic correlation. Radiology1990;176:389-94.

18 Becklake MR. Asbestos-related diseases of the lungs andother organs: their epidemiology and implications forclinical practise. Am Rev Respir Dis 1976;114:187-227.

19 Al Jarad N, Poulakis N. Pearson MC, Rubens MB, RuddRM. Assessment of asbestos-induced pleural disease bycomputed tomography: correlation with chest radio-graph and lung function. RespirMed 1991;85:203-8.

20 Becklake MR. Asbestos-related diseases of the lung andpleura: current clinical issues. Am Rev Respir Dis 1982;126:187-94.

Correspondence and editorialsOccupaonal and Environmental Medicine absolute minimum. Letters are accepted onwelcomes correspondence relating to any of the understanding that they may be subjectthe material appearing in the journal. Results to editorial revision and shortening.from preliminary or small scale studies may The journal also publishes editorials whichalso be published in the correspondence are normally specially commissioned. Thecolumn if this seems appropriate. Letters Editor welcomes suggestions regardingshould be not more than 500 words in length suitable topics; those wishing to submit anand contain a minimum of references. editorial, however, should do so only afterTables and figures should be kept to an discussion with the Editor.

243

on June 17, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.4.239 on 1 April 1994. D

ownloaded from