High-resolution CT imaging of the lung for patients with primary Sjogren's syndrome

PURPOSE: To assess pulmonary abnormalities in patients with primary Sjogren's syndrome (PSS) using high-resolution computed tomography (HRCT). MATERIAL AND METHODS: The HRCT scans of 24 patients with the diagnosis PSS were retrospectively reviewed regarding the presence, extension and distribution of 16 pathological findings. RESULTS: Nineteen patients (79.2%) showed pathological findings and in five patients (21.8%) the HRCT scan was judged to be normal. A predominance of abnormalities in the lower lobes and subpleural areas was detected. The following pathologies were found: bronchiectasis, thin-walled cysts and small pulmonary nodules (46.2%), ground-glass attenuation and emphysema (37.8%), interlobular-septal thickening (29.4%), honeycombing (25.2%), bronchial wall thickening, tree-in-bud pattern (21.0%), mosaic perfusion (16.8%), architectural distortion (12.6%). Airspace consolidation, air trapping, large nodules (10-30mm) and masses (>30mm), mediastinal lymph node enlargement (>15mm) and free pleural fluid were seen each in 4.2%. In 7 of the 11 patients with thin-walled cysts areas of ground-glass attenuation were detected. CONCLUSION: HRCT seems is contributive to the characterization of the wide variety of lung abnormalities in PSS. Airway disease alone or in association with the presence of varying degrees of interstitial disease represents the main findings in accordance with earlier reports. Unexpectedly, almost half of the patients had thin-walled cysts on the HRCT scans, which etiology is unclear but could be associated with areas of ground-glass attenuation indicating LIP.     

Pulmonary fibrosis and emphysema: Is the emphysema type associated with the pattern of fibrosis?

AIM: To investigate whether the predominant emphysema type is associated with the high resolution computed tomography (HRCT) pattern of fibrosis in combined pulmonary fibrosis and emphysema (CPFE). METHODS: Fifty-three smokers with upper lobe emphysema and lower lobe pulmonary fibrosis on - HRCT - were retrospectively evaluated. Patients were stratified into 3 groups according to the predominant type of emphysema: Centrilobular (CLE), paraseptal (PSE), CLE = PSE. Patients were also stratified into 3 other groups according to the predominant type of fibrosis on HRCT: Typical usual interstitial pneumonia (UIP), probable UIP and nonspecific interstitial pneumonia (NSIP). HRCTs were scored at 5 predetermined levels for the coarseness of fibrosis (Coarseness), extent of emphysema (emphysema), extent of interstitial lung disease (TotExtILD), extent of reticular pattern not otherwise specified (RetNOS), extent of ground glass opacity with traction bronchiectasis (extGGOBx), extent of pure ground glass opacity and extent of honeycombing. HRCT mean scores, pulmonary function tests, diffusion capacity (DLCO) and systolic pulmonary arterial pressure were compared among the groups. RESULTS: The predominant type of emphysema was strongly correlated with the predominant type of fibrosis. The centrilobular emphysema group exhibited a significantly higher extent of emphysema (P < 0.001) and a lower extent of interstitial lung disease (P < 0.002), reticular pattern not otherwise specified (P < 0.023), extent of ground glass opacity with traction bronchiectasis (P < 0.002), extent of honeycombing (P < 0.001) and coarseness of fibrosis (P < 0.001) than the paraseptal group. The NSIP group exhibited a significantly higher extent of emphysema (P < 0.05), total lung capacity (P < 0.01) and diffusion capacity (DLCO) (P < 0.05) than the typical UIP group. The typical UIP group exhibited a significantly higher extent of interstitial lung disease, extent of reticular pattern not otherwise specified, extent of ground glass opacity with traction bronchiectasis, extent of honeycombing and coarseness of fibrosis (0.039 > P > 0.000). Although the pulmonary arterial pressure was higher in typical UIP group relative to the NSIP group, the difference was not statistically significant. CONCLUSION: In CPFE patients, paraseptal emphysema is associated more with UIP-HRCT pattern and higher extent of fibrosis than centrilobular emphysema.     

Computed tomography in pulmonary sarcoidosis

We studied the high resolution CT (HRCT) scans of 15 patients with biopsy-proven sarcoidosis and correlated the findings with pulmonary function tests (12 patients), 67Ga scans (10 patients), bronchoalveolar lavage (five patients), recent transbronchial biopsy (six patients), and recent open lung biopsy (three patients). The HRCT features included small nodules, thickened interlobular septa, patchy focal increase in lung density, honeycombing, and central conglomeration of vessels and bronchi. Active alveolitis was present by gallium scanning criteria in 5 of 10 cases. By bronchoalveolar lavage criteria, activity was present in three of five cases. Patchy increase in density may correlate with active alveolitis as seen on 67Ga scanning. High resolution CT was better than chest X-radiography for demonstration of patchy increase in density and for distinguishing nodules from septal thickening. Both nodules and patchy density were partly reversible following therapy. Nodular densities seen on CT correlated with the presence of granulomata on histology. Resting pulmonary function tests correlated poorly with presence and extent of lung disease on HRCT. The presence on HRCT of focal fine nodules, patchy focal increase in lung density, and central crowding of bronchi and vessels should suggest the diagnosis of sarcoidosis. In some patients, HRCT can identify unsuspected parenchymal lung disease and document the reversible components of sarcoid lung disease.     

Idiopathic pulmonary fibrosis: spectrum of high-resolution CT findings

OBJECTIVE: Characteristic high-resolution CT (HRCT) findings of idiopathic pulmonary fibrosis (IPF) include reticulation, architectural distortion, and honeycombing involving mainly the lung periphery and the lower lobes. In 50% of IPF patients, HRCT is nonspecific. This article illustrates the HRCT findings of IPF correlating with the pathology. CONCLUSION: The spectrum of HRCT manifestations varies from typical findings that allow confident diagnosis to atypical patterns mimicking other diseases, including predominance of ground-glass opacity, consolidation, nodules, and atypical distribution of lesions.     

Unilateral proximal interruption of the pulmonary artery in adults: CT findings in eight patients

PURPOSE: The authors retrospectively reviewed CT findings of unilateral proximal interruption of the pulmonary artery in eight adults. METHODS: Patient age ranged from 23 to 65 years (mean, 44 years), and three men and five women (six left-side affected and two right-side affected) comprised the study group. High resolution CT was obtained in five patients. Associated anomalies of great vessels included five cases of right aortic arch and one tetralogy of Fallot. RESULTS: On the affected side, complete defects of pulmonary arteries were seen in all patients. CT revealed serrated pleural thickening in six patients (75%), subpleural parenchymal bands in five (63%), and mosaic attenuation in three (38%) on the affected lung. Dilatation of bronchial arteries in five patients (71%), internal thoracic arteries in four (57%), and intercostal arteries in three (43%) were also seen. As for other parenchymal changes, cystic bronchiectasis and honeycombing were seen in two (25%) respective cases. On the unaffected lung, mosaic attenuation was seen in four (50%) patients. CONCLUSIONS: The main manifestations of unilateral proximal interruption of the pulmonary artery were complete defects of pulmonary arteries, enlarged bronchial arteries, and formed pleural thickening or some parenchymal changes of the lung.     

Honeycombing on CT; its definition,pathologic correlation,and future direction of its diagnosis

Honeycombing on CT is the clue for the diagnosis of usual interstitial pneumonia (UIP) and its hallmark. According to the ATS-ERS-JRS-ALAT 2010 guideline, the patients with honeycombing on CT can be diagnosed as UIP without surgical biopsy. On CT scans, it is defined as clustered cystic airspaces, typically of comparable diameters of the order of 3–10 mm, which are usually subpleural and have well-defined walls. Pathologically, honeycombing consists of both collapsing of multiple fibrotic alveoli and dilation of alveolar duct and lumen Although the definition of honeycombing seems to be strict, recognition of honeycombing on CT is various among each observer Because typical honeycombing is frequently observed in the patients with UIP, we should judge clustered cysts as honeycombing when a diagnosis of UIP is suspected.     

Early asbestosis: evaluation with high-resolution CT

To determine the earliest stage at which lesions in asbestosis can be diagnosed and to assess their progression, 23 asbestos-exposed patients with minimal or no abnormalities at plain radiography were examined with high-resolution computed tomography (HRCT) twice, with an interval of 12-37 months between examinations. In 21 of the patients, parenchymal abnormalities were found. Major parenchymal features seen at CT included thickened intralobular and interlobular lines, subpleural curvilinear lines, pleural-based nodular irregularities, hazy patches of increased attenuation, small cystic spaces, and small areas of low attenuation. At paired serial CT, subpleural isolated dots or branching structures connected with the most peripheral branch of the pulmonary artery started to appear in lower subpleural zones and then became confluent to create pleural-based nodular irregularities. CT-pathologic correlation led to the conclusion that the confluence of subpleural peribronchiolar fibrosis creates subpleural fibrosis.     

无显性胸水的肺癌胸膜播种性转移:影像学检出和诊断

目的提高对无显性胸水的肺癌胸膜播种性转移的影像学检出和诊断水平。方法对12例无明显胸水征象而经手术和病理证实有胸膜播种性转移的肺癌病例的胸部平片、普通CT和高分辨率CT(HRCT)等影像学表现进行回顾性分析;并结合文献比较上述各种影像学手段对此种肺癌胸膜播种性转移的检出和诊断价值。结果本组12例肺癌胸膜播种性转移中,胸部平片、CT和HRCT对其检出率分别为8.3%、58.3%和91.7%。HRCT对肋胸膜和纵隔胸膜转移检出率为54.5%,对叶间胸膜转移的检出率为85.7%。胸壁胸膜播种性转移的HRCT表现有:(1)胸膜面毛糙和棘状小突起;(2)胸膜面凹凸不平;(3)胸膜面散在2~3 mm大小结节;(4)胸膜面5~10 mm左右圆形或半圆形结节;(5)胸膜面正常。叶间胸膜播种性转移的HRCT表现有:(1)叶间胸膜均匀增厚;(2)叶间胸膜不均匀增厚;(3)叶间胸膜上散在2~3 mm结节;(4)2~3 mm结节与增厚的叶间胸膜一起构成串珠状改变;(5)叶间裂附近成堆或散在2~3 mm结节,酷似位于肺实质内。结论HRCT是检出和诊断肺癌胸膜播种性转移的最佳影像学技术。     

Cystic Lung Disease: a Comparison of Cystic Size, as Seen on Expiratory and Inspiratory HRCT Scans

Objective

To determine the effects of respiration on the size of lung cysts by comparing inspiratory and expiratory high-resolution CT (HRCT) scans.

Materials and Methods

The authors evaluated the size of cystic lesions, as seen on paired inspiratory and expiratory HRCT scans, in 54 patients with Langerhans cell histiocytosis (n = 3), pulmonary lymphangiomyomatosis (n = 4), confluent centrilobular emphysema (n = 9), paraseptal emphysema and bullae (n = 16), cystic bronchiectasis (n = 13), and honeycombing (n = 9). Using paired inspiratory and expiratory HRCT scans obtained at the corresponding anatomic level, a total of 270 cystic lesions were selected simultaneously on the basis of five lesions per lung disease. Changes in lung cyst size observed during respiration were assessed by two radiologists. In a limited number of cases (n = 11), pathologic specimens were obtained by open lung biopsy or lobectomy.

Results

All cystic lesions in patients with Langerhans cell histiocytosis, lymphangiomyomatosis, cystic bronchiectasis, honeycombing, and confluent centrilobular emphysema became smaller on expiration, but in two cases of paraseptal emphysema and bullae there was no change.

Conclusion

In cases in which expiratory CT scans indicate that cysts have become smaller, cystic lesions may communicate with the airways. To determine whether, for cysts and cystic lesions, this connection does in fact exist, paired inspiratory and expiratory HRCT scans are necessary.     

Asbestosis: high-resolution CT-pathologic correlation

High-resolution computed tomography (HRCT) was performed in seven inflated and fixed postmortem lungs from seven asbestos-exposed patients with pathologically proved asbestosis. The parenchymal abnormalities seen at in vitro HRCT included thickened intralobular lines (n = 7), thickened interlobular lines (n = 7), pleural-based opacities (n = 7), parenchymal fibrous bands (n = 5), subpleural curvilinear shadows (n = 4), ground-glass appearance (n = 4), traction bronchiectasis (n = 4), and honeycombing (n = 2). The thickened intralobular lines were shown histologically to be due to peribronchiolar fibrosis. Thickened interlobular lines were due mainly to interlobular fibrotic thickening in four lungs and edema in three. The peribronchiolar fibrosis was most severe in the subpleural lung regions, creating curvilinear line shadows and pleural-based areas of opacity. Some subpleural fibrosis extended proximally along the bronchovascular sheath to create bandlike lesions. Areas of ground-glass appearance on HRCT scans were shown to be the result of mild alveolar wall and interlobular septal thickening due to fibrosis or edema. Postmortem HRCT findings were similar to premortem HRCT findings and correlated well with the pathologic findings of asbestosis.     

Diffuse panbronchiolitis: correlation of high-resolution CT and pathologic findings.

Diffuse panbronchiolitis (DPB) is characterized by chronic airflow limitation and airway inflammation with bronchiolar lesions. Chest radiographs of patients with DPB usually show small nodular shadows throughout both lungs. The authors investigated the nature and pathogenesis of the radiologic features of DPB by correlating high-resolution computed tomographic (HRCT) findings with histopathologic features. The HRCT images of nine patients with DPB were compared with the observations made with inflated lung specimens. The HRCT findings of DPB included centrilobularly distributed, small rounded areas of attenuation; branched linear areas of attenuation, contiguous with the small rounded areas; dilated airways with thick walls, also common outside secondary pulmonary lobules; and decreased lung attenuation in peripheral areas due to air trapping caused by bronchiolar narrowing in the subpleural zones. The authors believe that HRCT best demonstrates this characteristic location of small rounded areas of attenuation associated with dilated airways.     

High-resolution computed tomography features of nonspecific interstitial pneumonia and usual interstitial pneumonia

OBJECTIVE: To assess the accuracy of high-resolution computed tomography (HRCT) in the diagnosis of nonspecific interstitial pneumonia (NSIP). We hypothesized that the computed tomography (CT) features of NSIP could be distinguished from those of usual interstitial pneumonia (UIP). METHODS: The HRCT images of 47 patients with surgical lung biopsy-proven NSIP (n = 25) and UIP (n = 22) were independently reviewed by 2 thoracic radiologists. Predominant imaging patterns, most likely diagnosis, and diagnostic level of confidence were recorded. A confident HRCT diagnosis of NSIP was based on the presence of spatially uniform, bilateral, basal-predominant ground-glass and/or reticular opacities with little if any honeycombing, whereas UIP was confidently diagnosed if a spatially inhomogeneous, bilateral, peripheral, basal-predominant pattern of reticular opacities and honeycombing with little if any ground-glass attenuation was identified. RESULTS: A predominant pattern of ground-glass and/or reticular opacity with minimal to no honeycombing was demonstrated in 48 (96%) of 50 readings in patients with NSIP. Conversely, the presence of honeycombing as a predominant feature had a predictive value of 90% for UIP (P < 0.001). Usual interstitial pneumonia was more likely than NSIP to be subpleural and patchy (P < 0.001). A confident CT diagnosis of NSIP and UIP was correct in 73% and 88% of cases, respectively. The correctness of a CT diagnosis made at intermediate or high confidence was 68% and 88%, respectively. The kappa value for distinction of NSIP from UIP was 0.72. CONCLUSION: In contrast to previous reports, NSIP can be separated from UIP in most cases. The presence of honeycombing as a predominant imaging finding is highly specific for UIP and can be used to differentiate it from NSIP, particularly when the distribution is patchy and subpleural predominant. The presence of predominant ground-glass and reticular opacity is highly characteristic of NSIP, but there is a subset of patients with UIP who have this pattern and may require biopsy for differentiation from NSIP.     

Assessment of diffuse infiltrative lung disease: comparison of conventional CT and high-resolution CT

The diagnostic accuracies of conventional computed tomography (CT) and high-resolution CT (HRCT) in examination of the entire lungs were compared prospectively, and the diagnostic utility of a limited number of HRCT sections in the evaluation of diffuse infiltrative lung disease (DILD) was determined in 150 consecutive patients. Conventional 10-mm CT sections were equivalent to 1.2-mm HRCT sections in the recognition of nodules, masses, nodular irregularities of the interfaces, large cystic air spaces, and architectural distortion. Conventional CT was superior to HRCT in the diagnosis of micronodules and infiltrates. However, 15% of micronodules were recognized only on HRCT images because of their small size and low density. Fine bronchial and parenchymal lesions were best seen on HRCT scans. HRCT was the only technique that allowed assessment of the presence of areas of ground-glass attenuation. HRCT improves visualization of small parenchymal structures and allows a confident evaluation of DILD, providing that the entire lungs are studied. Nevertheless, an optimal CT evaluation of micronodular patterns requires additional thick-section CT scans.     

Thoracic CT findings of adult T-cell leukemia or lymphoma

OBJECTIVE: The aim of this study was to assess pulmonary CT findings in patients with adult T-cell leukemia or lymphoma. MATERIALS AND METHODS: We retrospectively reviewed CT scans of the lung in 87 patients with adult T-cell leukemia or lymphoma who had undergone chest CT between January 1996 and March 2002 at two institutions. The CT scans were interpreted by two chest radiologists working in consensus. Parenchymal abnormalities (ground-glass attenuation, consolidation, nodules, thickening of bronchovascular bundles, interlobular septal thickening, honeycombing, crazy-paving appearance, and bronchiectasis) were evaluated, as were enlarged lymph nodes, pleural effusion, and pleural thickening. In 46 patients who underwent surgical biopsy or autopsy, CT-pathologic correlation was performed with the actual specimens by a pathologist and two chest radiologists. RESULTS: On the CT scans, abnormal findings were seen in 60 patients (69.0%). CT findings consisted of ground-glass attenuation (n = 37), centrilobular nodules (n = 25), thickening of bronchovascular bundles (n = 22), and consolidation (n = 13). These abnormalities were predominantly seen in the peripheral lung parenchyma (n = 26). Pathologically, these findings corresponded with atypical lymphocyte infiltration along the interstitium and the alveolar spaces. Pleural effusion and enlarged lymph nodes were found in 22 and 27 patients, respectively. CONCLUSION: CT findings in patients with adult T-cell leukemia or lymphoma consisted mainly of ground-glass attenuation, centrilobular nodules, and thickening of the bronchovascular bundles in the peripheral lung. These findings, although nonspecific, are considered suggestive of thoracic involvement in patients with adult T-cell leukemia or lymphoma.     

Apical opacity associated with pulmonary tuberculosis: high-resolution CT findings

To elucidate the nature of the apical opacity that is commonly seen in patients with tuberculosis--usually referred to as an "apical cap" or "apical pleural thickening"--18 patients with upper lobe tuberculosis were studied with high-resolution computed tomography (HRCT). All had a homogeneous apical opacity at least 1 cm thick on chest radiographs. Fifteen of the 18 had a history of pulmonary tuberculosis of more than 5 years duration, and nine showed evidence of ipsilateral pleurisy. HRCT scans at the apex of the thorax in all nine patients scanned at this level showed that extrapleural fat with interspersed vessels accounted for most of the plain radiographic opacity. Scans obtained at a level slightly above visible aerated lung showed extrapleural fat 3-25 mm thick peripherally and atelectatic lung centrally. At more caudal levels, at which both aerated lung and "thickened pleura" were visible on plain radiographs, HRCT showed extrapleural fat (3-20 mm thick), thickened pleura (1-3 mm thick), and atelectatic lung peripherally and areas of emphysematous bullae, bronchiectasis, and atelectatic lung centrally.     

Nonspecific interstitial pneumonia versus usual interstitial pneumonia: differences in the density histogram of high-resolution CT

OBJECTIVE: To assess the capability of the density histogram of high-resolution CT (HRCT) in differentiating nonspecific interstitial pneumonia (NSIP) from usual interstitial pneumonia (UIP) without definite honeycombing. METHODS: Twenty-eight NSIP and 32 UIP patients without definite honeycombing on CT were included in this study. We evaluated a CT image at the level of 1 cm above the diaphragm. The pixels of the lung parenchyma were classified into 4 ranges. The fractions of each range of corresponding density were calculated. The skewness (the degree of asymmetry of a distribution) and kurtosis (how sharply peaked a histogram is) were obtained from the density histogram. The mean value and the mode value of the lung attenuation were also measured. RESULTS: The fraction of the range of ground-glass opacity and reticular opacity was greater in NSIP patients (32%, 12%) than in UIP patients (23%, 8%) (P < 0.001). UIP had a larger fraction of the range of normal lung. The density histogram was less skewed (P = 0.01) and had a wider peak (P = 0.02) in NSIP (skewness = 1.2879 +/- 0.5672, kurtosis = 1.2115 +/- 1.9470) than in UIP (skewness = 1.6426 +/- 0.4664, kurtosis = 2.3880 +/- 1.8183). CONCLUSION: The density histogram reflected the differences in the CT features between NSIP and UIP. Therefore, a density histogram may be helpful for differentiating NSIP from UIP without definite honeycombing.     

Imaging of small airways disease.

High-resolution computed tomography (HRCT) is the most useful modality for imaging of small airways disease. Direct signs of small airways disease that appear on HRCT scans are the result of changes in the airway wall or lumen. Abnormal small airways can be seen as tubular, nodular, or branching linear structures on HRCT scans. Indirect signs of small airways disease result from changes in the lung parenchyma distal to the diseased small airway and include air trapping, subsegmental atelectasis, centrilobular emphysema, and air-space nodules. Diverse inflammatory and infectious processes, such as bronchiolitis obliterans (BO), bronchiolitis obliterans with organizing pneumonia (BOOP), smoking-related diseases, and asthma affect the small airways of the lungs. HRCT findings of BO include air trapping and bronchiectasis. The predominant findings of BOOP are consolidation and ground-glass attenuation. HRCT can show abnormalities such as small nodules and areas of ground-glass attenuation even in asymptomatic smokers, but emphysema predominates in smokers with moderate or severe obstructive disease. Patients with asthma can have thickened airway walls, plugged large and small airways, subsegmental atelectasis, and air trapping, but emphysema is rarely seen even in severe asthma patients. HRCT scans can often accurately depict disease processes in the small airways and can occasionally lead to a specific diagnosis from among several clinically relevant possibilities.     

Thoracic manifestation of myeloperoxidase-antineutrophil cytoplasmic antibody (MPO-ANCA)-related disease. CT findings in 51 patients

OBJECTIVE: To assess the pulmonary CT findings of patients with serum evidence of the myeloperoxidase-antineutrophil cytoplasmic antibody (MPO-ANCA). METHODS: The pulmonary CT scans of 62 patients with serum evidence of MPO-ANCA (51 with microscopic polyangiitis, 11 with Churg-Strauss syndrome) were retrospectively assessed with regard to parenchymal, pleural, and mediastinal abnormalities. RESULTS: On the CT scans, abnormal findings were seen in 51 of the patients (82%, n = 62). Of the patients, the CT findings consisted of ground-glass attenuation in 48 of the patients (94%, n = 51), consolidation in 40 (78%), and thickening of bronchovascular bundles in 26 (51%). Pathologically, these findings corresponded to alveolar hemorrhages, interstitial chronic inflammation in the alveolar septa, vasculitis, or fibrosis. These abnormalities were predominantly seen in peripheral lung parenchyma (n = 37). CONCLUSION: The CT findings in patients with MPO-ANCA consisted mainly of ground-glass attenuation and consolidation in the peripheral lung. These findings, although nonspecific, are considered as pulmonary involvement in patients with MPO-ANCA.     

Imaging of pulmonary histiocytosis X.

Pulmonary histiocytosis X is an uncommon but important cause of pulmonary fibrosis and honeycombing in young adults. This article reviews the pathologic, clinical, radiographic, and high-resolution computed tomographic (HRCT) features of pulmonary histiocytosis X, focusing on differential diagnosis and disease progression. The main pathologic feature of pulmonary histiocytosis X is peribronchiolar inflammation, leading to fibrosis and cyst formation. Although the diagnosis of pulmonary histiocytosis X may be suspected on the basis of chest radiographic findings, the HRCT findings of predominantly upper lobe nodules and cysts are virtually pathognomonic of this disorder. As pulmonary histiocytosis X progresses, the nodules decrease in number, leaving multiple thin-walled cysts. HRCT can be useful in visually estimating the proportion of lung involved and is also valuable in distinguishing histiocytosis from other disorders that produce nodules or cysts. Chest HRCT helps confirm the pattern seen at radiography and is valuable in establishing the correct diagnosis.     

High-resolution computed tomography of drug-induced lung disease.

In order to determine the potential clinical utility of high-resolution CT (HRCT) in the assessment of drug-induced lung disease, we reviewed the chest radiographs and HRCT scans of 23 patients and five normal controls. The radiographs and HRCT scans were reviewed separately in random order by two independent observers who were not aware of the relative numbers of patients and controls. Abnormal findings were detected in the affected patients in 17/23 radiographs compared to 23/23 HRCT scans. The patients included five cases of bleomycin toxicity, five cases of nitrofurantoin toxicity, two cases each of penicillamine, busulphan, BCNU and amiodarone toxicity, and one case each of cyclophosphamide, procainamide, mitomycin and methotrexate toxicity. The HRCT appearances could be grouped into four categories according to their dominant pattern and distribution of disease. These include fibrosis with or without consolidation (n = 12), ground-glass opacities (n = 7), widespread bilateral consolidation (n = 2), and bronchial wall thickening with areas of decreased attenuation (n = 2). The results of this study indicate that HRCT is more sensitive than the radiograph in the detection of drug-induced lung disease. The appearances demonstrated by HRCT reflect the pathological mechanisms of drug-induced lung disease.