Thorax 1992;47:645-650 645 High resolution computed tomographic assessment of asbestosis and cryptogenic fibrosing alveolitis: a comparative study

N Al-Jarad, B Strickland, M C Pearson, M B Rubens, R M Rudd

Abstract High resolution computed tomography is Background The aim of this study was particularly valuable in the early detection of to compare the distribution and con- lung fibrosis in asbestos workers in whom the figuration of lung opacities in patients chest radiograph appears normal or shows with cryptogenic fibrosing alveolitis and only pleural disease.7 Yoshimura et al'° and asbestosis by high resolution computed Akira et al" have correlated the types of tomography. shadows seen in the high resolution computed Methods Eighteen patients with tomogram in patients with asbestosis with the cryptogenic fibrosing alveolitis and 24 histological findings at necropsy. with asbestosis were studied. Two The differentiation between asbestosis and independent observers assessed the type cryptogenic fibrosing alveolitis is often and distributions of opacities in the difficult, particularly when the extent of upper, middle, and lower zones of the occupational asbestos exposure is unclear. computed tomogram. The potential value of high resolution Results Upper zone fibrosis occurred in computed tomography in differentiating these 10 of the 18 patients with cryptogenic diseases has not been assessed specifically. In fibrosing alveolitis and in six of the one study computed tomographic features 24 patients with asbestosis. A specific were investigated in 118 patients with different pattern in which fibrosis was distributed types of interstitial fibrosis, who later had open posteriorly in the lower zones, laterally lung biopsy. High resolution computed in the middle zones, and anteriorly in the tomography gave the correct diagnosis in 76% upper zones was seen in 11 patients with of cases compared with 57% of cases by plain cryptogenic fibrosing alveolitis and in chest radiography."2 The computed tomo- four with asbestosis. Band like intra- graphic changes in patients with asbestosis in pulmonary opacities, often merging with that study were described as being the same as the pleura, were seen in 19 patients with those in patients with cryptogenic fibrosing asbestosis but in only two with crypto- alveolitis with the addition of bilateral pleural genic fibrosing alveolitis. Areas with thickening,'2 but only two patients with a reticular pattern and a confluent asbestosis were included. or ground glass pattern were the We compared the findings of high resolution commonest features of cryptogenic computed tomography in patients with fibrosing alveolitis (15 and 14 patients cryptogenic fibrosing alveolitis and asbestosis respectively) but were uncommon in to determine whether there are differences asbestosis (four and three patients). other than frequency of associated pleural Pleural thickening or plaques were seen changes. in 21 patients with asbestosis and in none with cryptogenic fibrosing alveolitis. Conclusion Apart from showing Methods pleural disease high resolution com- PATIENTS puted tomography showed that confluent We studied 18 patients with cryptogenic (ground glass) opacities are common in fibrosing alveolitis (four women and 14 men, cryptogenic fibrosing alveolitis and rare mean age 57 (range 33-75) years) and 24 men in asbestosis whereas thick, band like with asbestosis (mean age 59 (34-75) years). opacities are common in asbestosis and Asbestosis was diagnosed when fine mid to late rare in London Chest cryptogenic fibrosing alveolitis. inspiratory crackles and pulmonary opacities of Hospital, London E2 9JX N Al Jarad High resolution computed tomography is Table 1 Mean (9500 confidence interval) age and pack M C Pearson years of smoking in patients with asbestosis and M B Rubens superior to chest radiography in investigating cryptogenicfibrosing alveolitis. Difference between groups R M Rudd interstitial lung disease because there is less not significant Royal Brompton and superimposition of structures, which allows a National Heart better assessment of the type, distribution, Cryptogenic Hospital, London and fibrosing SW3 6HP severity of parenchymal abnormalities Asbestos alveolitis B Strickland than is possible by chest radiography.' Several n = 24 n= 18 Reprint requests to: recent studies have described the computed Dr R M Rudd Age (years) 59 (56 to 64) 57 (51 to 63) tomographic appearance of various diffuse Pack years of smoking 26 (19 to 35) 29 (20 to 38) Accepted 1 March 1992 interstitial lung diseases.24 646 Jarad, Strickland, Pearson, Rubens, Rudd

Table 2 Mean (95% confidence interval) results of lun' function tests expressed as percentage ofpredicted values in patients with asbestosis and cryptogenicfibrosing alveolitis Cryptogenic fibrosing Asbestosis alveolitis n = 24 n = 18 FEV, 76 (67 to 85) 82 (72 to 93) FVC 80 (71 to 88) 91 (77 to 102) FEV,/FVC 94 (85 to 103) 100 (91 to 109) Total lung capacity 89 (81 to 97)* 72 (66 to 76) Residual volume 99 (90 to 108)** 76 (67 to 85) TLCO 63 (55 to 71)** 38 (30 to 46) Alveolar volume 84 (77 to 92) 77 (70 to 84) Kco 78 (70 to 87)** 54 (44 to 63) *p < 0002,**p < 0001. FEV, forced expiratory volume in one second, FVC forced vital capacity, TLCO = carbon monoxide transfer factor, Kco = TLCO divided by alveolar volume.

a profusion grade greater than 1/0 on the International Labour Office scale'3 were seen in the chest radiograph in patients who had had substantial exposure to asbestos. In 11 patients in whom bronchoalveolar lavage was performed asbestos bodies were detected in bronchoalveolar lavage fluid. Cryptogenic fibrosing alveolitis was diagnosed on the basis of open lung biopsy in 12 patients and on clinical and plain radiographic evidence of interstitial (b)a lung disease in the other six. None of the patients with cryptogenic fibrosing alveolitis had been exposed to asbestos, had any symp- toms or signs suggesting connective tissue disease or malignancy, had positive test results for avian precipitins, or had received any drug known to induce lung fibrosis.

HIGH RESOLUITION COMPUTED TOMOGRAPHY All scans were carried out with an Elscint 2002 scanner with a scan time of 5-5 s. We took 3 mm sections, 10 mm apart, from the lung apices to the bases in full inspiration at total lung capacity using a bone algorithm reconstruction.2" Window settings for lung fields (mean window setting 1602 Hounsfield units) and for soft tissue for clear identification of the pleura (mean window setting 621 Hounsfield units) were obtained in all patients. Figure 1 Cryptogenicfibrosing alveolitis: distribution of Additional sections were obtained in the prone shadows. Fibrosis is predominantly posterior in lower sections (a), lateral in middle sections (b), and anterior position. in upper sections (c). This distribution was seen in 1I (61%) patients with cryptogenicfibrosing alveolitis and INTERPRETATION OF SCANS four (17%) patients with asbestosis. High resolution computed tomograms were assessed by two independent readers. Lung fields were divided into three equal thirds, confluent, ground glass, honeycombing), and measured from apex to base. Readers were any pleural abnormalities in each lung zone. asked to identify the presence of emphysema, Readers also noted the distribution of the opacities suggesting fibrosis (linear, reticular, abnormal shadows present in the lung fields.

LUNG FUNCTION MEASUREMENTS Table 3 Distribution ofshadowing in the upper, middle, and lower thirds of lungfield in Patients did respiratory function tests within patients with cryptogenicfibrosing alveolitis (CFA) and asbestosis three weeks of undergoing high resolution Fibrosis Emphysema computed tomography. Spirometry was done with a dry cylinder spirometer, single breath Asbestosis CFA Asbestosis CFA carbon monoxide transfer factor was measured n = 24 n= 18 p value n= 24 n = 18 p value with an Auto-link transfer factor machine, Upper third 6 10 <0-05 11 13 NS and lung volumes were measured with a com- Middlethird 19 18 <005 15 11 NS Lower third 23 18 NS 20 12 NS puterised constant volume plethysmograph. Predicted values were calculated for age, sex, High resolution computed tomographic assessment ofasbestosis and cryptogenicfibrosing alveolitis: a comparative study 647

Table 4 Types andfrequencies of opacities in asbestosis and cryptogenicfibrosing alveolitis cryptogenic fibrosing Type of opacities Asbestosis alveolitis p value Crescentic fine reticular pattern with small cysts (fig 1) 4/24 15/18 < 0-0006 Confluent or ground glass (fig 3) 3/24 14/18 <0-00008 Subpleural lines (figs 4 and 5) 8/24 4/18 NS Thick linear or band like densities in the lower zones merging with the pleura (fig 7) 18/24 2/18 <0 0002 Wedge shaped or irregular densities closely related to the pleura (figs 8 and 9) 8/24 0/18 < 0-007 Pleural plaques and diffuse pleural thickening 21/24 0/18 <0 00001 Rounded atelectasis (fig 10) 6/24 0/18 <0-03 Figure 3 Cryptogenicfibrosing alveolitis: confluent shadowing in basal segments of lower lobes with irregular and height'5 and results were expressed as subpleural areas of low attenuation, especially on right. percentages of predicted values. This "alveolar" type of change is rarely seen in asbestosis.

STATISTICAL ANALYSIS anteriorly in the upper zones (fig 1). This Age, pack years of smoking, and lung function distribution was seen in I 1 (61 %) patients with measurements in patients with asbestosis and cryptogenic fibrosing alveolitis but in only four cryptogenic fibrosing alveolitis were compared (17%) with asbestosis (p < 0 003). We gained by the Mann-Whitney U test. a qualitative impression that interstitial fibrosis The frequencies ofdisease in each ofthe lung and emphysema tended to be more distorting to thirds and type of shadows seen in the two lung architecture in conditions were compared by Fisher's exact cryptogenic fibrosing test or the x2 with the continuity (Yates) alveolitis than in asbestosis. Emphysema was correction when the former test was in- not found in any ofthe non-smoking patients in Results were considered either group. appropriate. sig- Table 4 summarises the types of opacities nificant when p was less than 0-05. and their frequency of occurrence in asbestosis and cryptogenic fibrosing alveolitis. In 15 ofthe 18 patients with cryptogenic fibrosing alveolitis Results. there were areas of reticular or reticulonodular There were no differences between the patients shadowing (fig 1). A fine honeycomb or cystic with asbestosis and cryptogenic fibrosing pattern with no or minimum pleural disease alveolitis in age or pack years of smoking occurred in four of the 24 patients with asbes- (table 1). There were no significant differences tosis (fig 2). Areas of confluent and sometimes between the groups for FEV, or FVC, but the ground glass pattern, characteristically residual volume, total lung capacity, carbon surrounded by a crescentic subpleural and monoxide transfer factor (TLco) and TLCO paramediastinal transradiency, were seen in 14 divided by alveolar volume (Kco) were all patients with cryptogenicfibrosingalveolitis (fig higher in the asbestosis group (table 2). 3) but in only three patients with asbestosis. Table 3 shows the distribution ofopacities in Most patients with cryptogenic fibrosing lung thirds in the two conditions. Fibrosis in alveolitis had areas of both reticular and the lower zones was present in all patients in confluent opacification. both groups; the upper and middle thirds were Discrete subpleural lines were seen in a affected more often in cryptogenic fibrosing few patients with asbestosis and cryptogenic alveolitis than in asbestosis. In patients with fibrosing alveolitis, the lines tending to be cryptogenic fibrosing alveolitis the opacification narrower in asbestosis (fig 4) than in cryp- tended to be most marked posteriorly in the togenic fibrosing alveolitis (fig 5). Subpleural lower zones, laterally in the middle zones, and

Figure 2 Asbestosis: diffuse cystic shadows in both lungs. Pleural plaque is present on lateral aspect of right lung (arrow).

Figure 4 Asbestosis: narrow subpleural curvilinear lines parallel to pleural surface on both sides. 648 Jarad, Strickland, Pearson, Rubens, Rudd

Figure 5 Cryptogenic Circumscribed pleural plaques and areas of fibrosing alveolitis: on the thickening were seen in 19 right irregular, dense more diffuse pleural crescentic subpleural lines patients with asbestosis (fig 10), although in are visible wherefibrosis is some the pleural thickening was minor in most severe; on the left thickness and extent. Two further patients irregular subpleural line is visible. The line is less well with asbestosis had pleural plaques alone, and defined than that normally three had no detectable pleural disease. Areas seen in asbestosis. of rounded atelectasis due to pleural infolding (Blesovsky syndrome, figs 8 and 10) were seen in six patients with asbestosis and in none with cryptogenic fibrosing alveolitis. No definite pleural thickening was seen in cryptogenic fibrosing alveolitis, but multiple small peaks of _.. density arising from the visceral pleura on both lateral and mediastinal surfaces were often present (fig 1). lines were the only pulmonary abnormalities seen in three patients with asbestosis but were always associated with other types ofinterstitial opacities in cryptogenic fibrosing alveolitis. Discussion Two patients with asbestosis had discrete sub- The distribution and type of radiographic pleural lines in the lower zones posteriorly shadows in asbestosis and cryptogenic fibrosing which disappeared in the prone position (fig 6). alveolitis have not been compared sys- Both patients had fine end inspiratory crackles tematically before, probably because the on auscultation and the plain radiograph was presence of pleural disease in asbestosis has interpreted as showing interstitial fibrosis. been used as the prime differentiating radiogra- Thick linear opacities perpendicular to the phic finding. Pleural plaques are not invariably pleura were seen in the lower zones in 18 present in asbestosis, however, and in their patients with asbestosis but in only two with absence these two conditions may be indistin- cryptogenic fibrosing alveolitis (fig 7). Wedge guishable by chest radiography. shaped or curvilinear intrapulmonary opacities Our study suggests that features visible on (figs 8 and 9) were seen in eight patients with high resolution computed tomograms aid asbestosis but were not identified in any patient differentiation between asbestosis and with cryptogenic fibrosing alveolitis. These cryptogenic fibrosing alveolitis. Pulmonary opacities were usually close to the pleura and fibrosis affecting the upper and middle thirds of merged with it, although in two patients they the lungs was more common in cryptogenic were wholly within the lung parenchyma (fig 9). fibrosing alveolitis than in asbestosis, in which fibrosis was often localised to the lower lobes. In cryptogenic fibrosing alveolitis a characteristic Figure 6 Asbestosis: (a) pattern of opacities in the posterior areas of the subpleural curvilinear lines lower zones, the lateral areas of the middle in supine position in 63 zones, and the anterior areas ofthe upper zones year oldformer lagger; (b) these lines disappear in was commonly seen; the posterior and lateral prone position. This may aspects of the upper zones appeared relatively represent early stage normal. We also found that when asbestosis asbestosis, in whichfibrosis is shown by summation affected the anterior segments of the upper effect of vascular zones the abnormality was usually less severe shadowing. than that seen in cryptogenic fibrosing alveolitis. Confluent or ground glass opacities were

6 a

Figure 7 Asbestosis: coarse subpleural reticular pattern in right lung which merges with thickened posterior and diaphragmatic pleura. High resolution computed tomographic assessment ofasbestosis and cryptogenicfibrosing alveolitis: a comparative study 649

Figure 8 Asbestosis: on radiographic evidence of asbestosis in smokers the left long thick irregular compared with non-smokers.17-20 linear densities lie In asbestosis the commonest features were perpendicular to and merge with pleural surface coarse linear opacities, often adjacent to pleural and are associated with thickening. This pattern, which was seen rarely soft reticular density. On in cryptogenic fibrosing alveolitis, is believed to the right there is an irregular opacity be due to interlobular and intralobular attracting vessels and fibrosis.'4 A fine regular cystic honeycomb bronchi. Appearances on pattern, as is often seen in cryptogenic fibrosing both sides may represent infolded lung in varying alveolitis,2 occurred in only a few patients with stages of evolution. asbestosis. Ventilatory capacity was slightly, although not significantly, lower in the asbestosis group than the cryptogenic fibrosing alveolitis group, more common in our patients with cryptogenic in those with asbestosis but the residual volume and total lung capacity fibrosing alveolitis than were higher, perhaps reflecting asbestos and were commonly surrounded by subpleural induced narrowing of the small airway."" or paramediastinal radiolucency, or both. shown that the The carbon monoxide gas transfer and co- Histological studies have efficient were more impaired in patients with opacities are due to mild thickening of the oedema cryptogenicfibrosing alveolitis, perhaps reflect- alveolar wall and interlobular area by ing the more widespread distribution of the or fibrosis. " The subpleural transradiency is around interstitial fibrosis. Pleural thickening may have probably due to air being trapped contributed to relative preservation of the Kco sheets of lung fibrosis.'0 1' with asbestosis. Subpleural curvilinear lines in patients with value in some patients three The question arises whether the more wide- asbestosis were discrete, and in patients spread distribution ofopacities in patients with these were the only interstitial abnormalities. the sub- cryptogenic fibrosing alveolitis than in those In cryptogenic fibrosing alveolitis, with asbestosis reflects more advanced disease pleural lines were less discrete and were always group. This is unlikely to be the associated with other interstitial opacities. In in the former discrete sub- case. Fibrosis of upper zones in cryptogenic three patients with asbestosis fibrosing alveolitis and the type of opacity were pleural lines were detected posteriorly in the in the not confined to those with more severely supine position but were not seen prone impaired lung function but were seen at an position since all three patients had fine mid to with mildly a chest early stage in some patients late inspiratory crackles and radiograph impaired lung function. appeared consistent with interstitial fibrosis. high lines are an We conclude that the features seen on We presume that these reversible computed tomography which favour early sign of asbestosis, possibly representing resolution induced a diagnosis of cryptogenic fibrosing alveolitis areas offibrosis highlighted by gravity include confluent (ground glass) shadows, vascular shadows. and cystic shadows extending Emphysema, cysts in a fibrotic area, and reticulonodular in size to the upper thirds of the lung field, and a subpleural cysts may be increased by posterior lower zone, fibrosis exerting traction.'6 These features were characteristic pattern of in with lateral middle zone, and anterior upper zone found more often patients cryptogenic opacities. Features that favour a diagnosis of fibrosing alveolitis than in those with asbestosis, or diffuse severe fibrosis. Fibrosis asbestosis include pleural plaques especially in areas of thickening, thick band like opacities in and cystic shadows were not found in the upper pleural three non-smokers the lower zones which extend to merge with the and middle zones in the isolated narrow subpleural with cryptogenic fibrosing alveolitis and the one pleura, and non-smoker with asbestosis. Smoking may curvilinear lines. induce or exacerbate the cystic disease and may also exacerbate the interstitial fibrosis. This suggestion is consistent with previous findings of increased frequency and severity of chest

Figure 9 Asbestosis: on SOL-Coui-a the right thick pleural based lines penetrate deep into lungfield. On the left a thick curvilinear line is attached to pleura and extends into lungfield, forming an arch within

lungfield. .dffm-. doloomcbb---. I

Figure 10 Asbestosis: posterior pleural based opacities in both lower zones probably represent infolded pleura and lung parenchyma (pseudotumours or Blezovsky syndrome). 650 Jarad, Strickland, Pearson, Rubens, Rudd

1 Zerhouni EA, Naidich DP, Stitik FP, Khouri NF, 12 Mathieson JR, Mayo JR, Staples CA, Muller NL. Chronic Siegelman SS. Computed tomography of the pulmonary diffuseinfiltrative lung disease: A comparison ofdiagnostic parenchyma. II. Interstitial disease. J Thorac Imaging accuracy of CT and chest radiography. Radiology 1985;1:54-64. 1989;171:1 11-6. 2 Strickland B, Strickland NH. The value of high definition, 13 International Labour Office. International classification of narrow section computed tomography in fibrosing radiographs of pneumoconiosis. Geneva: ILO, 1980. alveolitis. Clin Radiol 1988;39:589-94. (Occupational safety and health series. No 22; rev 80). 3 Bergin CJ, Muller NL. CT of interstitial lung disease: a 14 Strickland B, Brennan J, Denison DM. Computed tomo- diagnostic approach. Am J Radiol 1987;148:8-15. graphy in diffuse lung disease: Improving the image. 4 Staples CA, Muller NL, Vedal S, Abboud R, Ostrow D, Clinical radiology 1986;37:335-8. Miller RR. Usual interstitial pneumonia: correlation of 15 Cotes JE. Lung function: assessment and application in CT with clinical, functional, and radiologic findings. medicine. 4th ed. Oxford: Blackwell Scientific 1979:281-7. Radiology 1987;162:377-81. 16 Wiggins J, Strickland B, Turner-Warwick M. Combined 5 Muller NL, Miller RR. State of the art: computed tomo- cryptogenic fibrosing alveolitis and emphysema; the value graphy ofchronic diffused infiltrative lung disease, part I. of high resolution computed tomography in assessment. Am Rev Respir Dis 1990;142:1206-15. Respir Med 1990;84:365-9. 6 Hansell DM, Kerr IH. The role ofhigh resolution computed 17 Weiss W. Cigarette smoking, asbestos, and pulmonary tomography in the diagnosis of infiltrative lung disease. fibrosis. Am Rev Respir Dis 1971;104:223-7. Thorax 1991;46:77-84. 18 Weiss W, Theodas PA. Pleuropulmonary disease among 7 Nakata H, Kimoto T, Nakayama T, Kido MM, Miyazaki N, asbestos workers in relation to smoking and type of Narada S. Diffuse peripheral lung disease: Evaluation by exposure. J Occup Med 1978;20:341-5. high-resolution computed tomography. Radiology 1985; 19 Lilis R, Selikoff IJ, Lerman Y, Seidman H, Gelb SK. 157:181-5. Asbestosis: interstitial pulmonary fibrosis and pleural 8 Staples CA, Gamsu G, Ray CS, Webb WR. High resolution fibrosis in a cohortofasbestos insulation workers: influence computed tomography and lung function in asbestos of cigarette smoking. Am J Ind Med 1986;10:459-70. workers with normal chest radiograph. Am Rev Respir Dis 20 Kilburn KH. Clarification of data on the etiological role of 1989;139: 1502-8. cigarette smoking in pulmonary fibrosis. Am J Ind Med 9 Aberle DR, Gamsu G, Ray CS, Feuerstein IM. Asbestos- 1984;5:421-2. related pleural and parenchymal fibrosis: detection by 21 Begin R, Cantin A, Berthiaume Y, Boileau R, Peloquin S, high-resolution CT. Radiology 1988;166:729-34. Masse S. Airway function in lifelong nonsmoking old 10 Yoshimura H, Hatakeyama M, Otsuji H, Maeda M, Ohishi asbestos workers. Am J Med 1983;75:631-8. H, Uchida H, et al. Pulmonary asbestosis: CT study of 22 Cohen BM, Adasczik A, Cohen EM. Small airways changes subpleural curvilinear shadow. Work in progress. in workers exposed to asbestos. Respiration 1984;45: Radiology 1986;158:653-8. 296-302. 11 Akira M, Yamamoto S, Yokoyama K, Kita N, Morinaga 23 Wright JL, Churg A. Severe diffuse small airways K, Higashihara T, et al. Asbestosis: high-resolution CT- abnormalities in long term asbestos miners. Br J Ind Med pathologic correlation. Radiology 1990;176:389-94. 1985;42:556-9. Adventitia My only private patient Foolishly I never did private practice. Fool- would have to have a biopsy. The posh GP said ishly because if you don't do private practice he would arrange it. I assumed that this was the you are never going to see Elizabeth Schwarz- usual thing in private practice. He sent her to a koff as a patient, or for that matter Madonna, urologist! A week or two later the answer came and the money must come in handy-though back. It was follicular carcinoma ofthe thyroid, when I started a consultant's salary was a living which apparently is well known to occur in wage. teenage girls. It wasn't well known to me. It Almost 30 years ago a posh GP rang me up to was the first I had heard of it and the last. So I say that he was sending me the 14 year old rang the posh GP and told him the girl should daughter of an American film producer-this be sent to Jack Piercy, the surgeon superinten- on the advice of an American professor of dent of New End Hospital in Hampstead. New paediatrics I had met in New York. The girl End was the name of a famous thyroid clinic had been diagnosed in New York as having originally started by London County Council. tuberculous glands of the neck and I was to It had had several great men on its staff- continue her treatment with isoniazid alone. I among them Cecil Joll, who wrote a massive didn't like the sound of that. surgical textbook on the thyroid; Geoffrey I found a rather sulky girl with bilateral Keynes (the brother of Maynard, the econ- cervical gland enlargement. The glands were omist), who was the world authority on quite large, up to about an inch in diameter, not William Blake, and also Raymond Greene (the tender, firm, and mobile. Not at all like tuber- brother of Graham, the writer), who had been culous glands. What is more the girl's mother an Everest climber. Jack Piercy was the greatest told me that the girl had started menstruating of the lot. Two days later the posh GP rang me and that during the cycle the glands changed in to say that Sir Ronald Bodley Scott, physician size quite considerably; that didn't sound like to the Queen, had said that we couldn't do tuberculosis either. better. I was very chuffed. So I started by doing a tuberculin test (the Jack Piercy told me in his letter that he had New York diagnosis had been purely clinical- treated 17 similar patients by total thyroidec- no investigations had been done). It was a Heaf tomy and block dissection ofneck and that they test, the only one available in the chest clinic (a had all survived. This girl underwent a success- shop near Oxford Circus) and it is a very good ful block dissection of the thyroid. Many years test: it requires no skill, it is virtually painless, later I heard that she was well (taking thyroxine and, conveniently, it can be read at a week. The as replacement therapy, of course). response was negative. So it wasn't tuber- I sent the film producer a bill for £25, which culosis (atypical mycobacteria as a cause of he paid by return but with no word of thanks. cervical adenitis had hardly been heard of in I thought, "I will never be as lucky as that Britain at that time). again so I'd better not see any more private I rang the posh GP and told him that the girl patients," and I never did. PETER D B DAVIES