| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Thoracic Imaging |
1 From the Departments of Radiology (M.A.), Medicine (H.I.), and Pathology (S.Y.), National Kinki Chuo Hospital for Chest Disease, 1180 Nagasone-cho, Sakai City, Osaka 591-8555, Japan. Received July 20, 2001; revision requested September 24; final revision received March 11, 2002; accepted March 25. Address correspondence to M.A. (e-mail: akira@kinchu.hosp.go.jp).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Thin-section CT scans obtained in 60 patients with drug-induced pneumonitis were evaluated retrospectively. The patients had 31 cases of antineoplastic agent–induced pneumonitis and 29 cases of nonneoplastic agent–induced pneumonitis (antibiotic agent, 20 cases; herbal medicine [sho-saiko-to], four cases; antirheumatic agent, three cases; phenytoin, one case; disodium cromoglycate, one case). CT scans were reviewed by two chest radiologists in consensus. Correlation between arterial oxygen tension level and the extent of disease at CT was available in 21 patients. These two factors were compared by using the Spearman rank correlation coefficient.
RESULTS: The predominant findings in antineoplastic agent–induced pneumonitis were diffuse or multifocal ground-glass opacities with intralobular interstitial thickening. The predominant CT findings in antibiotic agent–induced pneumonitis were patchy ground-glass opacities with centrilobular opacities and interlobular septal lines. The predominant CT findings in herbal medicine–induced pneumonitis were diffuse ground-glass opacities with patchy consolidation. Interlobular septal lines and centrilobular opacities were observed more frequently in antibiotic agent–induced pneumonitis, and intralobular interstitial thickening was observed more frequently in antineoplastic agent–induced pneumonitis. A significant correlation was established between arterial oxygen tension level and extent of disease at CT (r = -0.84, P < .05).
CONCLUSION: In addition to ground-glass opacities and interlobular septal lines, the most common thin-section CT findings were intralobular interstitial thickening, observed in antineoplastic agent–induced pneumonitis, and centrilobular opacities, observed in antibiotic-induced pneumonitis.
© RSNA, 2002
Index terms: Computed tomography (CT), thin-section, 60.1211 • Drugs, toxicity, 60.64 • Lung, CT, 60.1211 • Pneumonitis, 60.64
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The radiographic and computed tomographic (CT) appearances of various drug-induced lung diseases have been reported. The CT findings associated with the use of some drugs, such as bleomycin (3,4) and amiodarone (5), have been well described; however, those associated with the use of others have not been well described.
The purpose of this study was to describe thin-section CT findings in patients with drug-induced pneumonitis, to compare these CT findings, and to correlate them with arterial oxygen tension level.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
When there were, among associated drugs, substances known to induce lung damage, the resolution of pneumonitis was observed after withdrawal of treatment with the drug implicated, despite continuation of treatment with the associated drugs. The patients who fulfilled all previously mentioned criteria were included in this study. Patients who were thought to have acute respiratory distress syndrome caused by the drug, who did not have resolution of abnormalities after cessation of treatment with the drug, and who died were excluded from the study according to the previously mentioned criteria. In these patients, the cases included pneumonitis induced with gemcitabine (Gemzar; Eli Lilly, Japan) in two cases, that induced with paclitaxel (Taxol; Bristol Myers Squibb, USA) in one case, and that induced with 7-ethyl-10-(4-[1-piperidino]-1-piperidino) carbonyloxycamptothecin, CPT-11 (irinotecan hydrochloride; Topotecin, Daiichi, Japan), in one case. Sixty patients who met our criteria were included. Informed consent was provided by the patients, and the study was approved by the hospital’s internal review board. The patients were referred for CT because of unexplained symptoms and pulmonary abnormalities while they were receiving the medications.
The duration of use of the medication prior to CT examination ranged from 2 days to 2 years (median, 1 month). Examination of sputum in all patients and of bronchoalveolar lavage fluid in 28 patients disclosed no viral, fungal, bacterial, or parasitic pathogens. Twenty-eight of these 60 patients underwent transbronchial lung biopsies because clinical examination findings were ambiguous. Eighteen of 32 patients who did not undergo transbronchial lung biopsy had a clinical diagnosis of drug-induced pneumonitis without pathologic examination. The remaining 14 patients did not undergo transbronchial lung biopsy because of severe dyspnea. The lung biopsy findings showed interstitial pneumonitis in all 28 patients, with marked infiltration by eosinophils in eight patients (induced with minocycline hydrochloride [Mynocine, Minocyn, and Minocin; Lederle Laboratories, Pearl River, NY] in three patients and with isoniazid [Escotin; Daiichi], cefactor [Kefnal; Sinonogi, Japan], ethambutol hydrochloride [Lederle Laboratories], itraconazole [Itrizole; Kyowa, Japan], and disodium cromoglycate [Intal; Fisons, UK] in one patient each). Results of a provocation test performed in nine patients were positive in all. Twenty-five patients were treated with steroid therapy.
The mean duration of clinical and radiographic follow-up in these 60 patients was 13 months (range, 6–48 months). Forty-two patients underwent sequential thin-section CT examinations 7–32 days apart (mean, 15 days). Eighteen of these 42 patients underwent a third thin-section CT examination 17 days to 1 year after the second examination (median, 36 days) was performed.
Thirty of 31 patients being treated with antineoplastic drugs had lung cancer. One patient treated with antineoplastic drugs had lymphoma. Areas of a mass suggestive of lung cancer were not evaluated. One of the six patients treated with antituberculous drugs had nonrespiratory tuberculosis. Five patients treated with antituberculous drugs had pulmonary tuberculosis. In these five patients, new pulmonary abnormalities on the chest radiographs were evaluated.
CT Technique
The study group consisted of 45 men and 15 women aged 20–89 years (mean age, 62 years). Thin-section CT was performed with a CT unit (Quantex Plus; Yokogawa Medical System, Hino, Japan). All CT scans were obtained at maximal inspiration by using 2-mm collimation at 20-mm intervals. Scanning extended from the lung apices to below the costophrenic angles. Images were reconstructed with a high-spatial-frequency algorithm. CT scans were obtained with the patient in the supine position. All CT scans were obtained with window settings appropriate for lung parenchyma (window width, 1,500 HU; level, -700 HU) and mediastinum (window width, 300 HU; level, 10 HU).
Image Analysis
The CT images were reviewed by two chest radiologists who analyzed all of the images together, and final decisions regarding the findings were determined by consensus. The CT images were assessed specifically for patterns, distribution, and extent of the following pulmonary abnormalities: (a) ground-glass opacity, which was defined as an area of slightly increased attenuation in which the bronchial walls and vessels remained visible; (b) consolidation, which was defined as an area of increased attenuation with obscuration of the adjacent bronchial walls and vessels; (c) areas of linear opacity, which included interlobular septal lines and intralobular interstitial thickening; (d) nodules; and (e) centrilobular nodular and branching areas of high attenuation, which were suggestive of bronchiolar involvement.
Distribution of parenchymal nodules was recorded as centrilobular when nodules were identified around peripheral pulmonary arterial branches or 3–5 mm away from the pleura, interlobular septa, or pulmonary veins. Traction bronchiectasis was defined as irregular bronchial dilatation within areas with parenchymal abnormality. Bronchiolectasis was identified on the basis of an abnormal depiction of the small airways in the peripheral portion of the lung. The distribution of each finding was classified as predominantly in the upper, middle, or lower zones of the lung; as predominantly central, peripheral, or bronchovascular; or as patchy, diffuse, or multifocal. The terms patchy, diffuse, or multifocal were used arbitrarily. Patchy was defined as existing irregularly in a number of small separate areas. Multifocal was defined as existing irregularly in a number of large separate areas.
The CT score in the upper, middle, and lower lung zones was determined by visually estimating the extent of disease in each zone. The score was based on the percentage of lung parenchyma that showed evidence of abnormality and was estimated to the nearest 5% of parenchymal involvement. The overall percentage of involvement was calculated by averaging the scores from each of the six lung zones. Also, lymph node enlargement and pleural effusion were recorded.
Statistical Analysis
Arterial blood gas values determined with the patient at rest were measured with an automated blood gas analyzer. The arterial oxygen tension level that was determined on the day of the CT examination (n = 21) was used. The extent of disease demonstrated on CT scans was correlated with the arterial oxygen tension level by means of the Spearman rank correlation coefficient. The differences in the frequency of each abnormality in each subgroup were evaluated by using the Fisher exact test or the 
2 test with continuity (Yates) correction when the former test was inappropriate. Probability values of less than .05 were considered significant. Analyses were performed by using the statistical software package (SPSS-PC 7.5, version 4.0.5J; SPSS, Chicago, Ill).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
|
|
Findings with Antineoplastic Agent–induced Pneumonitis
The predominant CT finding in docetaxel-induced pneumonitis (n = 12) was a diffuse (n = 3) or multifocal (n = 7) ground-glass opacity with intralobular interstitial thickening (Fig 1). Patchy areas of consolidation were observed in six of the 10 patients. In the remaining two patients with docetaxel-induced pneumonitis, the CT finding was peripheral patchy consolidation. In two of these 12 patients, the parenchymal abnormality was unilateral, and in the remaining 10, it was bilateral. In six of these 12 patients, the distribution was predominantly peripheral. In 12 patients with toxic effects in the lung owing to docetaxel, interlobular septal lines were observed in five and nodules were observed in two. Centrilobular nodular areas of ground-glass opacity were observed in two patients. A small amount of pleural effusion was observed in two patients.
|
The CT finding with CPT-11–induced pulmonary toxic effects (n = 5) was a diffuse ground-glass opacity with both interlobular and intralobular interstitial thickening (n = 3) (Fig 2) or a patchy ground-glass opacity with intralobular interstitial thickening (n = 2). Nodules were observed in three patients. Patchy areas of consolidation were observed in three patients. A small amount of pleural effusion was observed in one patient. Traction bronchiectasis was observed in one patient.
|
|
The CT finding in one patient with a toxic reaction to bleomycin was a predominantly peripheral ground-glass opacity with intralobular interstitial thickening. A small amount of pleural effusion was observed.
The CT findings in one patient with a toxic reaction to methotrexate were bilateral peripheral ground-glass opacities with intralobular interstitial thickening, predominantly distributed in the lower lobes. Centrilobular areas of ground-glass opacity were observed in the subpleural regions.
Findings with Antibiotic Agent–induced Pneumonitis
The predominant CT finding in five patients with minocycline-induced pneumonitis (n = 9) was patchy consolidation with centrilobular opacities. The CT finding in three patients with minocycline-induced pneumonitis was a patchy ground-glass opacity. The CT findings in another patient with minocycline-induced pneumonitis were diffuse distributed centrilobular nodules with areas of ground-glass opacity (Fig 4). Interlobular septal lines and centrilobular opacities were observed in all nine patients. In these nine patients, the distribution of parenchymal abnormalities was diffuse in three, peripheral in three, and patchy in three. Nodules were observed in two of the nine patients, and intralobular interstitial thickening was observed in another two patients. Pleural effusion was observed in three patients, and hilar lymphadenopathy was observed in one.
|
|
The CT finding in three patients with isoniazid-induced pneumonitis was a diffuse patchy ground-glass opacity. Interlobular septal lines were observed in two of the three patients. Centrilobular opacities were observed in one of the three patients (Fig 5), and intralobular interstitial thickening was observed in one. The CT finding in one patient with ethambutol-induced pneumonitis (n = 2) was a diffuse ground-glass opacity with patchy areas of consolidation. Centrilobular opacities, interlobular septal lines, and small nodules were observed. A small amount of pleural effusion was observed. The CT finding in the other patient with ethambutol-induced pneumonitis was peripheral consolidation with surrounding ground-glass opacities. The CT findings in one patient with rifampin (Rifadin; Daiichi)-induced pneumonitis were generalized smooth interlobular septal lines, which appeared to represent noncardiogenic interstitial pulmonary edema (Fig 6). A small amount of ground-glass opacities and a few centrilobular opacities were observed. The patient’s symptoms were chest pain and fever but no dyspnea.
|
|
Findings with Pneumonitis Induced with Herbal Medicine
The predominant CT finding of pneumonitis induced with herbal medicine, or sho-saiko-to (n = 4), was a diffuse ground-glass opacity with patchy consolidation (Fig 7). Intralobular interstitial thickening was observed in three patients. Interlobular septal lines and centrilobular opacities were observed in two patients, and nodules were observed in one patient. Pleural effusion was observed in one of these four patients.
|
|
Extent of Disease at CT
The overall extent of parenchymal abnormalities revealed at CT ranged from 5% to 70% (mean, 33%). Correlation between the arterial oxygen tension level and the extent of disease at CT, which was available in 21 patients, is shown in Figure 9. A significant inverse correlation was observed between arterial oxygen tension level and extent of disease at CT (r = -0.84, P < .05).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Padley et al (8) divided the CT findings of patients with drug-induced lung disease into four categories according to their dominant pattern and distribution of disease: fibrosis (irregular linear opacities with architectural distortion) with or without consolidation, ground-glass opacities, widespread bilateral consolidation, and bronchial wall thickening with areas of decreased attenuation. In the patients they studied, there were five cases of toxic reactions to bleomycin; five cases of toxic reactions to nitrofurantoin; two cases each of toxic reactions to penicillamine, busulfan (formerly busulphan), carmustine, and amiodarone; and one case each of toxic reactions to cyclophosphamide, procainamide hydrochloride, mitomycin, and methotrexate. Irregular linear opacities with architectural distortion were the most common finding.
The second common pattern was ground-glass opacity, and this pattern was believed to be a hypersensitivity phenomenon. Janzen et al (9) examined the CT scans and radiographs obtained in 45 immunocompromised patients with acute pulmonary complications who did not have acquired immunodeficiency syndrome. In these patients, there were six cases of drug-induced lung disease. In their series, drug-induced lung disease was associated with a varying CT appearance; a predominant pattern of irregular linear opacities was most typical of this diagnosis, and the confidence levels of the observers in the diagnosis were lowest for drug-induced lung disease. According to the study of Patz et al (10), the predominant CT findings in 20 patients with toxic reactions to pulmonary drugs after high-dose chemotherapy and autologous bone marrow transplantation were scattered predominantly peripheral ground-glass or consolidated opacities that occasionally looked nodular or masslike.
Docetaxel is a new anticancer drug belonging to the taxane family, with a broad spectrum of activity in multiple tumors, including carcinoma of the ovary, the breast, and the bronchus (11). Several cases of pulmonary infiltrates associated with docetaxel and paclitaxel, the other taxane analogue, were reported (11–14).
Recently, CT findings in a case of paclitaxel-induced hypersensitivity pneumonitis were reported (14); CT scanning with 7-mm collimation depicted bilateral patchy areas of increased attenuation predominating in the upper lobes. The upper lobe predominance was not observed in our cases. The CT findings were similar to those in docetaxel-induced pneumonitis, except for one case in which multiple nodular opacities were observed. Ramanathan et al (13) reported that in one of three patients with paclitaxel-induced pulmonary toxic reactions, multiple nodular opacities were observed. Bleomycin-related pulmonary nodules also have been reported. The predominant histopathologic lesion was bronchiolitis obliterans organizing pneumonia (15).
Gemcitabine is a nucleoside analog that is used to treat a variety of solid neoplasms, including malignancies of the lung, breast, pancreas, and ovary, as observed by Boiselle et al (16). These investigators reported CT features of gemcitabine-induced pulmonary toxic reactions in three patients. The CT features included ground-glass opacity, thickened septal lines, and reticular opacities. Our findings in gemcitabine-induced pulmonary toxic effects were similar to those in the study of Boiselle et al. Interlobular septal thickening was more prominent with gemcitabine-induced toxic effects than it was with other chemotherapeutic agents.
CPT-11 is a derivative of camptothecin. CPT-11 exhibits a marked antitumor activity against not only a broad spectrum of experimental tumor models but also pleiotropically drug-resistant tumor cell lines. The major toxic reactions are myelosuppression (predominantly leukopenia), diarrhea, and pulmonary toxic effects (17).
Etoposide is a podophyllin derivative that has been extensively tested against various malignancies and is now widely used in the treatment of small cell lung cancer (18). Interstitial pneumonitis caused by etoposide is rare. A case of a fatal pulmonary toxic reaction following the initiation of the third cycle of orally administered etoposide monotherapy was reported (19).
Antimicrobial drug–induced pneumonitis is well reported in patients treated with nitrofurantoin, sulfasalazine, amphotericin B, sulfonamide drugs, p-aminosalicylic acid, and penicillin. Many cases of pneumonitis and eosinophilia associated with ingestion of minocycline have been described (7). Toyoshima et al (20) described thin-section CT findings in six patients with minocycline-induced pneumonitis. In their study, ground-glass opacity and interlobular septal thickening were observed in all six patients, consolidation was observed in three patients, and micronodules were observed in two patients. Our findings were similar to those of Toyoshima et al. Interlobular septal lines were observed in all our cases of minocycline-induced pneumonitis. In addition, we observed centrilobular nodules in all our cases of minocycline-induced pneumonitis.
The appearance of septal lines on chest radiographs obtained in patients with edema caused by an overdose of tricyclic antidepressants has been described (21). The appearance of septal lines on radiographs obtained in patients with interleukin 2 toxic reactions also have been reported (22). In our patient with rifampin-induced pneumonitis, CT findings included interlobular septal thickening, which appeared to represent noncardiogenic interstitial pulmonary edema. However, his symptoms were fever and severe chest pain but no dyspnea. His clinical condition resembled acute chest pain syndrome (23).
In the two studies about crazy paving appearance at CT by Murayama et al (24) and Johkoh et al (25), intralobular interstitial thickening was observed in drug-induced pneumonitis. What kind of drug was studied is unclear in their study. In our study, ground-glass opacity with superimposed intralobular interstitial thickening at thin-section CT was frequent in chemotherapeutic agent–induced pneumonitis. We also observed that centrilobular opacities and interlobular septal lines were frequently observed in antibiotic agent–induced pneumonitis. The CT finding of a ground-glass opacity with centrilobular nodules resembles that in extrinsic allergic alveolitis, although centrilobular nodules are considered to be less prominent in most cases of drug-induced pneumonitis than they are in cases of extrinsic allergic pneumonitis (26,27), and interlobular septal lines are more prominent in cases of drug-induced pneumonitis than they are in cases of extrinsic allergic alveolitis.
In drug-induced pneumonitis, it was reported that the changes were predominantly peripheral in nature (3–5,10). In our cases, we observed a predominantly peripheral distribution in 19 (32%) cases. In the remaining 41 patients, the distribution was diffuse or patchy in both central and peripheral areas of the lungs. Peripheral consolidation was observed in our 10 (17%) patients. In these 10 patients were three cases of chemotherapeutic agent–induced pneumonitis (three of 31, 10%), six cases of antibiotic agent–induced pneumonitis (six of 20, 30%), and one case (100%) of disodium cromoglycate–induced pneumonitis.
Our study had several limitations. Some common causes of lung disease induced with ingestion of amiodarone, nonsteroidal antiinflammatory drugs, or nitrofurantoin were not included. No pathologic correlation was provided in regard to the varying imaging appearances of these lung diseases. Moreover, we used consensus interpretation rather than independent review of the images.
Although thin-section CT findings in drug-induced pneumonitis are nonspecific and the same drug may produce more than one pattern, the radiologist should be aware of the more frequent findings associated with the type of agent involved.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
