| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Service de Pneumologie et Allergologie Pédiatriques, Hôpital
Necker Enfants Malades, Assistance Publique des Hôpitaux de Paris, 149
rue de Sèvres, 75015 Paris, France.
2 Service de Radiologie Pédiatriques, Hôpital Necker Enfants
Malades, Assistance Publique des Hôpitaux de Paris, Paris, France.
Received June 10, 2007;
accepted after revision March 11, 2008.
Address correspondence to J. de Blic
(j.deblic{at}nck.aphp.fr).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Fifty-nine patients selected over a 14-year period (29 girls, 30 boys; mean age, 6 ± 4.9 years; range, 2 months–18 years) had nine disorders. CT scans were evaluated independently by two experienced chest radiologists, who were unaware of pathologic or clinical data. The radiologists recorded specific CT findings of infiltrative lung disease and were asked to give the most likely diagnosis and up to two differential diagnoses. Descriptive statistic analysis was followed by logistic regression analysis for each elementary lesion on the grid of abnormalities.
RESULTS. A correct first-choice diagnosis was made in 38% of CT observations. The correct diagnosis was among the three main choices in 59% of CT observations. Pulmonary alveolar proteinosis (n = 18) was most frequently correctly diagnosed; it was the first-choice diagnosis 47% of the time and among the three main choices 72% of the time. The correct first-choice diagnosis of idiopathic pulmonary fibrosis (n = 16) was made 43% of the time; of hypersensitivity pneumonitis (n = 4), 37% of the time; of sarcoidosis (n = 7), 28% of the time; of idiopathic pulmonary hemosiderosis (n = 6), 16% of the time; and of connective tissue diseases (n = 5), 10% of the time. All single cases of pulmonary fibrosis with calcification, lymphangiectasia, and Langerhans' cell histiocytosis were correctly diagnosed.
CONCLUSION. Our results showed there are limitations to diagnosing chronic infiltrative lung disease in children on the basis of CT data alone. We suppose that these differences are explained by the technical difficulties of high-resolution CT in children, the insufficient number of cases of and data on high-resolution CT of children, and the heterogeneity of lesions of a given cause.
Keywords: chronic infiltrative lung disease • high-resolution CT • pediatric
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Since its introduction in the 1980s, high-resolution CT, which can depict fine detail within the lung, has become the reference standard of imaging of chronic infiltrative lung disease in adults. High-resolution CT findings contribute to detection, evaluation of pulmonary damage, prognosis, and screening for complications [4]. It has proved useful in the diagnosis of chronic infiltrative lung disease with increased diagnostic accuracy, depending on the pathologic mechanism and the series considered [5–8]. When findings are characteristic and associated with clinical findings, use of high-resolution CT has led to a decrease in the number of lung biopsies [4, 9, 10]. Owing to the relative rarity of chronic infiltrative lung disease among children, few data are available on high-resolution CT findings in this population [11, 12]. The aim of this study was to evaluate the accuracy of diagnoses made with CT in a series of 59 children, for whom a diagnosis had already been ascertained. To our knowledge, this series is the largest of its type.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The cases of 59 patients (29 girls, 30 boys; mean age, 6 ± 4.9 [SD] years; range, 2 months–18 years) were identified over a 14-year period (January 1990–February 2004). The following nine disorders were evaluated: pulmonary alveolar proteinosis (n = 18), idiopathic pulmonary fibrosis (n = 16), sarcoidosis (n = 7), idiopathic pulmonary hemosiderosis (n = 6), connective tissue disease (n = 5), hypersensitivity pneumonitis (n = 4), Langerhans' cell histiocytosis (n = 1), pulmonary lymphangiectasia (n = 1), and pulmonary fibrosis with calcification (n = 1).
Imaging
CT was performed with several types of equipment. In eight cases, electron
beam CT scans were obtained with the following parameters: 130 kV; 630 mA;
acquisition time, 0.1 second; section thickness, 3 mm. In 51 cases,
conventional high-resolution CT scans were obtained with the fol lowing
parameters: 120–140 kV; 80–160 mA; section thick ness, 1–1.5
mm; acquisition time, 1 se cond. A high-spiral-resolution reconstruction algo
rithm was used, and scans were obtained with windows appropriate for viewing
the lung parench yma (width, 1,600 HU; level, –600 HU). In younger
children, who needed sedation, images were ob tained with the patient supine
during quiet res piration; older child ren were able to perform controlled
ventilation. If needed, sedation was given in the form of oral chloral hydrate
syrup (50 mg/kg). Contrast agent was injected in seven patients because the
chest radio graphic findings suggested adenopathy.
Image Interpretation
CT scans were evaluated independently on separate occasions by two
experienced chest radiologists from the pediatric radiology department of our
hospital. These radiologists did not participate in selection of the images.
They were unaware of the pathologic and clinical data, patient name, sex, age,
and examination date.
The radiologists recorded the following specific CT findings of infiltrative lung disease used in assessment of adult patients [14]: nodules (size and character), linear opacities (interlobular septal thickening or nonseptal parenchymal band), honey combing, consolidation, ground-glass opacity, peri bronchiolar thickening, cystic air space, pleural effusion, pleural or scissural thickening, paren chymal distortion, and presence of associated mediastinal adenopathy. The distribution of these findings was evaluated as follows: upper zone (apex to carina), middle zone (carina to lower pulmonary veins), lower zone (below the lower pulmonary veins), anterior, posterior, or random distribution. Each pattern was classified as absent, mild, moderate, or severe anomalies.
The radiologists were asked to give the most likely diagnosis and up to two differential diagnoses. They were supplied with a list of 11 diagnoses of chronic infiltrative lung disease found in the department in the last 14 years (the frequency of these diagnoses in the series of scans was not specified). The diagnoses were Langerhans' cell histiocytosis, sarcoidosis, hypersensitivity pneumo nitis, pulmonary alveolar protein osis, idiopathic pulmonary hemosiderosis, idio pathic pulmonary fibrosis (without histologic distinction), connective tissue disease, inherited metabolic disorder, lymph ocytic interstitial pneu monitis, pulmonary lymph angiectasia, and pul monary fibrosis with calcification.
Data Analysis
We collected 118 observations from the 59 patients because each study was
read by two radiologists. The data were entered in an Excel (Microsoft)
spreadsheet and analyzed with the SAS program (version 8.2, SAS). Descriptive
statistical analysis was followed by logistic regression analysis for each
elementary lesion on the grid of abnormalities. The kappa coefficient was used
to measure the degree of agreement (exact or not exact) between the diagnoses
given by the two radiologists.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
For all disorders considered, the radiologists gave the same first-choice
diagnosis (correct or incorrect) in 52% of cases and the same correct
first-choice diagnosis in only 27% of cases. Agreement between the
radiologists' two first diagnoses (correct or incorrect) was considered good
for pulmonary alveolar proteinosis (
= 0.66) and poor for idiopathic
pulmonary fibrosis ( = 0.21). We did not calculate the kappa
coefficient for the other diagnoses because the groups were too small. For the
most frequent elementary signs, interobserver agreement was 0.44 for
ground-glass opacity, 0.61 for linear opacity, and 0.43 for nodules.
Table 2 summarizes the CT findings in 59 children with chronic infiltrative lung disease. Because each scan was analyzed separately by the two radiologists, a total of 118 observations were made. Thus for unique disorders, the result was 50% if the radiologists gave discordant observations. Table 3 shows the predominant location of the elementary lesions for each disorder. Logistic regression analysis of each elementary sign on the grid of abnormalities showed no specific pattern of any chronic infiltrative lung disease in this study.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In only two studies [11, 12] to our knowledge have investigators collected data on imaging of chronic infiltrative lung disease in children and evaluated the diagnostic accuracy of high-resolution CT using similar methods. Copley et al. [11] reported the frequency of correct first-choice diagnoses was 61% in a series of 20 patients who had idiopathic pulmonary fibrosis (n = 8), lymphangiectasia (n = 3), follicular bronchiolitis (n = 3), pulmonary alveolar proteinosis (n = 2), lymphocytic interstitial pneumonitis (n = 2), idiopathic pulmonary hemosiderosis (n = 1), and lymphangiomatosis (n = 1). Lynch et al. [12] reported an accuracy of 40%, similar to our rate of 38%, for first-choice diagnoses, although the children in that study were very different from the children in ours. Most of the children in the study by Lynch et al. had conditions affecting the small airways, such as bronchiolitis obliterans, bronchiolitis obliterans organizing pneumonia, and bronchocentric granulomatosis. Only seven children had true infiltrative lung disease. As proposed by the Children's Interstitial Lung Disease Research Cooperative [13], we excluded children with predominant small-airways diseases.
Explanations for our results include the technical difficulties of high-resolution CT in children, the rarity of chronic infiltrative lung disease among children, and the heterogeneity of a radiologic pattern for a single disease in children. Before reading the chest high-resolution CT scans, the two radiologists came to consensus on the description of each elementary CT sign. Pediatric radiologists face several difficulties in interpreting thoracic scan data from children. It may be difficult to differentiate the diffuse ground-glass pattern from increased lung attenuation in cases of suboptimal inspiration and absence of breath-holding. These two possibilities can be differentiated by considering the position of the posterior tracheal membrane, which appears convex on inspiration and horizontal or slightly concave on expiration [15]. Controlled ventilation is a technique that should be encouraged because it considerably improves the quality of images [16].
Chronic infiltrative lung disease is much less frequent in children than in adults [2, 17]. For instance, the prevalence of idiopathic pulmonary fibrosis in children in the United Kingdom and Ireland has been estimated at 0.36/100,000 [18]. In contrast, in the adult population in the United States, the prevalence was estimated to be 14/100,000 (annual incidence, 6.8/100,000) [19]. The paucity of scan data from children with chronic infiltrative lung disease makes it necessary for pediatric radiologists to interpret children's scans on the basis of radiographic patterns in adults. Adult criteria, however, may not be the most appropriate for precise diagnosis of chronic infiltrative lung disease in children.
The best results for first-choice diagnosis are made with adult CT data on pathologic conditions rarely or not found in children. For example, Mathieson et al. [5] made a correct first-choice diagnosis in 76% of CT observations of chronic infiltrative lung disease with better accuracy than that for silicosis (93%), usual interstitial pneumonitis (89%), and lymphangitic carcinomatosis (85%). In our study, few patterns with the same cause were superimposable on patterns found in adults. The most frequent lesions in our group of patients with idiopathic pulmonary fibrosis, regardless of the histologic type of the lesion, were ground-glass opacities, septal thickening, and honeycombing. Copley et al. [11] found these lesions predominantly in the upper area of the lungs. In contrast, basal and peripheral reticular abnormalities, traction bronchiectasis, bronchiolectasis, and lower-lobe volume loss were not found in our study. The explanation may be that such signs indicate advanced fibrosis, and the scans in our study were selected 3–6 months after the initial onset of symptoms. Another reason is that features such as basal and peripheral reticular abnormalities, traction bronchiectasis, and lower-lobe volume loss typically occur in adult idiopathic interstitial fibrosis [20, 21].
Usual interstitial pneumonia is the predominant histologic type in adults, found in 90% of cases [9]. There is controversy, however, regarding the existence of this histologic pattern in children [22, 23]. Our results were consistent with the doubt. The scans of our children with idiopathic pulmonary fibrosis showed numerous extensive ground-glass opacities closely resembling those of adult patients with nonspecific interstitial pneumonia and desquamative interstitial pneumonia [24, 25].
There is a need for histopathologic classification of so-called idiopathic infiltrative lung disease in children [17]. The description of specific conditions such as neuroendocrine cell hyperplasia, pulmonary interstitial glycogenosis, and, more important, surfactant dysfunction abnormalities will widen the spectrum of conditions considered and may lead to improved correlation with CT findings [13, 26].
Another potential problem with our study was that the population did not reflect the true distribution of chronic infiltrative lung disease among children. Our patients were specifically recruited from La Réunion Island, where there is a high incidence of pulmonary alveolar proteinosis [27] compared with, for example, hypersensitivity pneumonitis [2]. The lesions visualized on the scans of patients with chronic infiltrative lung disease in this study were extremely diverse. For example, some pulmonary alveolar proteinosis CT scans may be considered typical, with adult data as the reference, because of septal thickening, diffuse ground-glass opacity, and bilateral consolidation [28, 29]. These cases were correctly diagnosed, whereas a pattern of micronodules, cysts, or honeycombing was incorrectly identified as idiopathic pulmonary fibrosis. Furthermore, childhood chronic infiltrative lung disease of various causes can progress to fibrosis in the long term. We found that the radiologists tended to overdiagnose idiopathic pulmonary fibrosis when CT showed honeycombing or cysts. We are aware that limitations of electron beam CT technique might have affected our ability to perceive findings that would have been found at slice thicknesses more commonly used in high-resolution CT.
In clinical daily practice, chronic infiltrative lung disease is diagnosed on the basis of a range of clinical, biologic, and radiologic information [30–33]. BAL data are particularly important for diagnosis [34]. For example, hypersensitivity pneumonitis can be suspected in children who have been exposed to birds such as pigeons and have chronic respiratory signs associated with restrictive lung function, high levels of specific precipitins, hyperlymphocytosis, and a prevalence of CD8-positive lymphocytes in the BAL fluid. BAL data also can be important in making a specific diagnosis of Langerhans' cell histiocytosis if the CD1a-positive cell count in the BAL fluid is greater than 5% [35], pulmonary alveolar proteinosis if the BAL fluid is milky and contains spumous macrophages associated with extracellular material and periodic acid–Schiff reagent staining of material within the macrophages [27, 36], and pulmonary hemosiderosis if the BAL fluid is hemorrhagic or contains many siderophages [37]. The diagnostic accuracy of high-resolution CT is further increased with findings at concurrent clinical evaluation and with the data on BAL fluid [38, 39]. The lack of such information made the task of our radiologists much more difficult and different from the situation they would face in daily practice.
In conclusion, the results of this study showed that there are significant limitations to the use of CT alone for diagnosing chronic infiltrative lung disease in children. It is conceivable that these limitations can be explained by technical difficulties in acquiring diagnostic images of children, by the heterogeneity of the features of particular diagnoses, and by the low number of cases of each diagnosis. Despite these difficulties, CT should always be included in the diagnostic evaluation if chronic infiltrative lung disease is suspected in children. It is helpful for the differential diagnosis based on CT pattern and facilitates assessment of the extent of disease. In the evaluation of children, systematic description of CT findings for each specific cause is needed to increase the knowledge and experience of pediatricians. Use of MDCT scanners with very short acquisition times should improve the quality of images and increase the accuracy of the etiologic diagnosis of chronic infiltrative lung disease in children.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  N. Schneebaum, H. Blau, R. Soferman, H. Mussaffi, L. Ben-Sira, M. Schwarz, and Y. Sivan Use and Yield of Chest Computed Tomography in the Diagnostic Evaluation of Pediatric Lung Disease Pediatrics, August 1, 2009; 124(2): 472 - 479. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
