| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Thoracic Imaging |
1 Departments of Radiology (H.S., T.J., A.I., J.N., M.I., N.M., O.H., N.T., S.H., H.N.) and Medical Physics (T.J.), Osaka University Graduate School of Medicine, Osaka, Japan; First Department of Internal Medicine, Kumamoto University School of Medicine, Kumamoto, Japan (K.I.); Department of Respiratory Medicine, Tosei General Hospital, Aichi, Japan (H.T., Y.K.); Department of Radiology, Kurume University School of Medicine, Japan (K.F.); Department of Radiology, National Cancer Center, Tokyo, Japan (U.T.); Department of Allergy Medicine, Komaki City Hospital, Aichi, Japan (T.H.); and Department of Radiology, University of British Columbia and Vancouver General Hospital, Vancouver, British Columbia, Canada (N.L.M.). From the 2004 RSNA Annual Meeting. Received June 2, 2005; revision requested July 29; revision received August 29; accepted September 21; final version accepted December 20. Address correspondence to H.S. (e-mail: h-sumikawa{at}radiol.med.osaka-u.ac.jp).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Materials and Methods: Institutional review board approval and informed consent were not required for this retrospective review of patient records and images. Two observers working independently and without knowledge of the diagnosis evaluated the extent and distribution of various thin-section CT findings (ground-glass opacity, consolidation, reticulation, and honeycombing) in 92 patients (51 men, 41 women; mean age, 56 years; age range, 29–81 years) with a histologic diagnosis of UIP (n = 20), cellular nonspecific interstitial pneumonia (NSIP) (n = 16), fibrotic NSIP (n = 16), respiratory bronchiolitis–associated interstitial lung disease (RB-ILD) (n = 11), desquamative interstitial pneumonia (DIP) (n = 15), or lymphoid interstitial pneumonia (LIP) (n = 14). Observers used univariate and multivariate statistical analyses to compare their findings with the extent and distribution of UIP.
Results: Observers made the correct diagnosis in 145 (79%) of 184 readings. Multivariate logistic regression analysis showed that the independent findings that distinguished UIP from cellular NSIP were the extent of honeycombing and the most proximal bronchus with traction bronchiectasis (odds ratio, 5.16 and 0.37, respectively); the finding that distinguished UIP from fibrotic NSIP was the extent of honeycombing (odds ratio, 2.10). CT features that distinguished UIP from RB-ILD and DIP included extent of ground-glass opacity (odds ratio, 0.76), thickening of bronchovascular bundles (odds ratio, 1.58), the most proximal bronchus with traction bronchiectasis (odds ratio, 0.22), and the number of segments with traction bronchiectasis (odds ratio, 3.64).
Conclusion: UIP has a characteristic appearance that usually facilitates distinction from other types of chronic IIPs at thin-section CT. The most useful finding when differentiating UIP from NSIP was the extent of honeycombing.
© RSNA, 2006
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
The clinical course of patients with acute interstitial pneumonia and cryptogenic organizing pneumonia is acute or subacute, ranges from days to weeks, and typically lasts less than 3 months (2,3). In contrast, patients with NSIP, RB-ILD, DIP, LIP, IPF, or UIP typically have a chronic clinical course that lasts several months or years; most of these patients experience symptoms for more than 3 months before pneumonia is diagnosed (1,4). Cryptogenic organizing pneumonia and acute interstitial pneumonia can be distinguished from IPF or UIP by their rapid clinical course; thus, we did not include these types of pneumonia in this study.
One of the key statements regarding consensus classification is that the primary role of thin-section computed tomography (CT) is to separate patients with characteristic findings of IPF or UIP from those with other types of IIP (1). This distinction is important because the prognosis of patients with IPF or UIP is considerably worse than the prognosis of patients with other types of chronic IIP (5–8). The American Thoracic Society and the European Respiratory Society concluded that the presence of characteristic clinical and thin-section CT features of IPF or UIP allows confident diagnosis and precludes the need for surgical biopsy (1). Katzenstein and Fiorelli (9) defined NSIP in 1994; however, most of the articles about CT features of IPF or UIP were written before 1994 and probably included cases of NSIP. Some researchers have evaluated the distinguishing features between NSIP, DIP, and IPF or UIP (5,10–13); however, their studies did not include other chronic interstitial pneumonias, such as RB-ILD and LIP. Furthermore, although some studies have shown that the accuracy of thin-section CT is high when used to diagnose IPF or UIP (14–16), other studies have shown that other chronic IIPs, particularly NSIP, may mimic IPF and UIP on CT images (10,11).
The purpose of our study was to retrospectively analyze the CT findings of chronic IIP and to determine the most helpful distinguishing findings of IPF or UIP with univariate and multivariate analyses.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Our intention was to include enough cases of each IIP to allow us to evaluate the characteristic thin-section CT findings and to assess the distinguishing features between the various IIPs. Between January 1988 and July 2004, patients with a histologic diagnosis of IPF or UIP (n = 37), NSIP (n = 41), LIP (n = 3), DIP (n = 2), or RB-ILD (n = 1) were identified by reviewing records from two of the seven institutions.
Nine patients with NSIP and 17 patients with IPF or UIP were excluded because of poor image quality due to motion artifacts or inadequate window level settings or because the CT film hard copies had been destroyed. During the same time period, patients with a histologic diagnosis of DIP (n = 13), LIP (n = 11), or RB-ILD (n = 10) were identified at the remaining five institutions. These patients were added to complete the study. Thus, 92 patients with a histologic diagnosis of chronic IIP who had undergone thin-section CT at one of seven institutions were included. IPF or UIP was diagnosed in 20 patients, cellular NSIP was diagnosed in 16, fibrotic NSIP was diagnosed in 16, RB-ILD was diagnosed in 11, DIP was diagnosed in 15, and LIP was diagnosed in 14.
IIP was diagnosed on the basis of histologic findings in specimens obtained at open lung biopsy or video-assisted thoracoscopic surgery and a clinical history that excluded all other known cases of interstitial pneumonia (ie, connective tissue disease, pneumoconiosis, or hypersensitivity pneumonitis). All IIPs were histologically proved at each participating institution, and the diagnosis was confirmed by experienced lung pathologists (with 31, 25, 20, 35, and 21 years of experience) who reviewed pathologic specimens and followed the current histologic criteria for diagnosis of each disease (1). In addition, the clinical findings in all patients were subsequently reviewed by two chest physicians (H.T. and Y.K., with 25 and 20 years of experience, respectively) to ensure that all cases fulfilled the diagnostic criteria recommended by the American Thoracic Society and the European Respiratory Society (1).
Patients with cellular interstitial pneumonia and relatively little fibrosis were placed in the cellular NSIP group, whereas patients with predominant fibrosis were placed in the fibrotic NSIP group (7). To allow multivariate analysis, RB-ILD and DIP were considered part of the spectrum of the same disease process, as considerable overlap exists between these diseases (1,17,18). Patients with RB-ILD or DIP were examined as a single group.
Acquisition and Review of Thin-Section CT Images
Thin-section CT images were obtained at end inspiration and with patients in the supine position; a variety of CT scanners were used. The scanning protocol consisted of reconstruction of 1–3-mm collimation sections with a high-spatial-frequency algorithm at 1- or 2-cm intervals. Images were photographed at window settings appropriate for viewing the lung parenchyma (window level, –600 to –700 HU; window width, 1200–1500 HU).
The images were reviewed independently in random order by two chest radiologists (K.I. and T.J., with 17 and 15 years experience, respectively). The observers were blinded to clinical and histologic information.
These observers evaluated the presence, extent, and distribution of the CT findings previously reported in patients with IIPs (1,18–25). These findings included ground-glass opacity, airspace consolidation, ill-defined centrilobular nodules, interlobular septal thickening, bronchovascular bundle thickening, intralobular reticular opacities, nonseptal linear or plate-like opacities, cysts, emphysema, honeycombing, traction bronchiectasis, and architectural distortion.
Ground-glass opacity was defined as hazy increased attenuation of the lungs that did not obscure the underlying vessels (26,27). Airspace consolidation was defined as a homogeneous increase in pulmonary parenchymal attenuation that obscured the underlying vessels (26,27). Ill-defined centrilobular nodules were defined as small and ill-defined nodules in a centrilobular distribution (26). Interlobular septal thickening was defined as abnormal widening of interlobular septa (27).
Thickening of bronchovascular bundles was defined as an increase in both the bronchial wall thickness and the diameter of pulmonary artery branches that was caused by thickened peribronchovascular interstitium (26). Intralobular reticular opacity was considered present when interlacing line shadows were separated by a few millimeters (26,27). Nonseptal linear or platelike opacity was defined as an elongated line of soft-tissue attenuation that was distinct from interlobular septa and bronchovascular bundles (26,27). Emphysema was defined as a focal region of low attenuation without visible walls (27).
Cysts were defined as round airspaces with a well-defined wall (27). Honeycombing was considered present when clustered cystic airspaces that ranged in size from 2 mm to 1 cm and had well-defined thick walls were seen in the subpleural regions (26,27). Traction bronchiectasis was defined as irregular bronchial dilatation within or around areas with a parenchymal abnormality.
Architectural distortion was considered present when bronchi, pulmonary vessels, or interlobar fissures or septa were abnormally displaced (27). The lungs were divided into six zones (upper, middle, and lower zones in both lungs), and each zone was evaluated separately. The upper zone was defined as the part of the lung above the level of the tracheal carina; the lower zone, as the part of the lung below the level of the inferior pulmonary vein; and the middle zone, as the portion of the lung between the upper and lower zones. Each CT finding was assessed and considered present, and the extent of involvement of the findings was evaluated visually and independently for each lung zone. A score was assigned on the basis of the percentage of lung parenchyma that showed evidence of an abnormality and was estimated to the nearest 10% of parenchymal involvement. Overall percentage of involvement was calculated by averaging the scores of the six lung zones.
The extent of traction bronchiectasis was evaluated by counting the number of segments that showed evidence of traction bronchiectasis. The following 18 segments or subsegments were evaluated: right apical upper, right anterior upper, right posterior upper, right lateral middle, right medial middle, right superior upper, right medial basal, right anterior basal, right lateral basal, right posterior basal, left apicoposterior upper, left anterior upper, left superior lingular, left inferior lingular, left superior lower, left anteromedial basal, left lateral basal, and left posterior basal. The extent of traction bronchiectasis was also quantified by assessing the generations of the most proximal bronchial branches involved. Traction bronchiectasis was scored as follows: 1, bronchial dilatation involving the trachea; 2, bronchial dilatation involving the main bronchus; 3, bronchial dilatation involving the second-generation bronchi; 4, bronchial dilatation involving the third-generation bronchi; 5, bronchial dilatation involving the fourth-generation bronchi; 6, bronchial dilatation involving the fifth-generation bronchi; 7, bronchial dilatation involving bronchi distal to the sixth-generation bronchi; and 8, no bronchial dilatation. Architectural distortion was determined to be present or absent in the whole lung.
After assessing the presence and extent of findings, the observers evaluated the predominant distribution of the findings. The distribution was classified as being predominantly in the upper lung zone, the lower lung zone, or random and as being predominantly peripheral or peribronchovascular. Distribution was considered predominantly in the upper zone when most of the findings were above the level of the tracheal carina. Distribution was considered to be predominantly in the lower zone when most of the findings were below the level of the tracheal carina. Peripheral predominance was defined as findings that involved mainly the outer third of the lung. Peribronchovascular predominance was defined as findings located mainly around the bronchus and artery.
After the observers assessed the predominant distribution of findings, they evaluated the extent of all abnormalities, except emphysema, in a whole lung to determine the percentage of lung parenchyma that showed findings consistent with IIP. This score was based on the percentage of lung parenchyma that showed evidence of an abnormality and was estimated to the nearest 5% of parenchymal involvement.
After reviewing the thin-section CT findings, the observers were asked to assign a likely diagnosis on the basis of the results of previous studies that reported the thin-section CT appearance of each entity (1,18–25,28). Differential diagnosis was limited to the four types of chronic IIP (UIP, NSIP, RB-ILD or DIP, and LIP).
Statistical Analysis
All statistical analyses were performed with statistical software (SPSS, version 12.0J; SPSS, Chicago, Ill). The interobserver variation for the extent of various abnormalities and for all abnormalities together was evaluated with the Spearman rank correlation coefficient. Interobserver variation for the existence of architectural distortion and predominant distribution was analyzed with the 
statistic. Interobserver agreement was classified as poor ( = 0.00–0.20), fair ( = 0.21–0.40), moderate ( = 0.41–0.60), good ( = 0.61–0.80), or excellent ( = 0.81–1.00). A P value of less than .05 was considered to indicate a significant difference.
The readings of the two observers pertaining to the extent of various abnormal findings were combined by calculating the average. Disagreement regarding the existence of architectural distortion and predominant distribution was resolved by a third radiologist (H.S.) with 5 years of experience. Univariate analysis was used to compare the abnormal CT patterns in patients with IPF or UIP with the CT patterns in patients with other IIPs. The extent of various abnormalities and of all abnormalities together was assessed with the Mann-Whitney U test.
The existence of architectural distortion and predominant distribution was analyzed with the Fisher exact test. Multivariate logistic regression analysis was used to assess the predictive value of the various thin-section CT findings in distinguishing IPF or UIP from the other types of IIP. The variables in the logistic regression model were selected with a stepwise procedure. Variables were retained in the logistic regression model if they contributed to the explanatory power of the regression equation (P < .10).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
= 0.48–0.74).
|
Disease Extent
The total extent (Table 2) of parenchymal abnormalities on CT scans in patients with IPF or UIP was significantly less than that in patients with LIP (P = .003) and similar to that in patients with cellular NSIP, fibrotic NSIP, RB-ILD, or DIP (P > .05). Centrilobular nodules were rarely found in patients with IPF or UIP. Twenty patients with IPF or UIP had predominantly lower lung zone distribution, and 15 had peripheral lung zone distribution (Table 1, Fig 1). In patients with cellular NSIP, honeycombing was rarely found (Fig 2). Univariate analysis showed that cellular NSIP was associated with a significantly greater extent of bronchovascular bundle thickening and a significantly lesser extent of honeycombing, cysts, emphysema, and bronchiectasis than IPF or UIP (P < .05 for all findings). Patients with fibrotic NSIP had a greater extent of ground-glass opacity and a smaller extent of honeycombing and cysts than did patients with IPF or UIP (P < .05 for all findings) (Fig 3).
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Univariate analysis showed that cellular NSIP resulted in a smaller extent of honeycombing and less traction bronchiectasis than did IPF or UIP. However, there was no significant difference in the degree of traction bronchiectasis between fibrotic NSIP and IPF or UIP. Honeycombing and traction bronchiectasis on CT scans reflect the presence of fibrosis (28). IPF or UIP was associated with more extensive fibrosis than was NSIP. This finding is similar to the findings of previous studies (11,13,32).
In our study, all cases that were misdiagnosed as IPF or UIP (n = 15) were diagnosed as NSIP at CT scanning. Thirteen of the 15 cases misdiagnosed as IPF or UIP at thin-section CT were thought to be associated with the peribronchovascular distribution, while only two of the 25 cases correctly diagnosed as IPF or UIP were thought to be associated with the peribronchovascular distribution. The cases with IPF or UIP that were misdiagnosed as NSIP were associated with extensive ground-glass opacities and mild or no honeycombing. Honeycombing was not seen in nine of 15 cases with IPF or UIP that were misdiagnosed as NSIP. On the other hand, five cases with histologically proved NSIP were misdiagnosed as IPF or UIP on thin-section CT images. Three of these five cases had honeycombing, and all five had predominant peripheral distribution of ground-glass opacities and reticulation.
In the regression models between RB-ILD, DIP, and UIP, CT features that distinguished UIP from RB-ILD and DIP were the extent of ground-glass opacity, the extent of bronchovascular bundle thickening, the generation of the most proximal bronchial branches with bronchiectasis, and the number of segments with bronchiectasis.
Bronchiectasis reflects the presence of fibrosis. Fibrosis was less extensive in patients with DIP or RB-ILD than in patients with UIP. However, UIP, RB-ILD, and DIP are often associated with a history of smoking (1,4,33), and in our study, emphysema was a common finding in patients with RB-ILD or DIP. Emphysema with superimposed ground-glass opacities, particularly emphysema that was present in the lower peripheral parenchyma, was often difficult to differentiate from honeycombing.
The findings associated with LIP were quite distinct from those associated with IPF or UIP. Ill-defined centrilobular nodules and thickening of bronchovascular bundles were the most common findings in patients with LIP, but they were uncommon in patients with IPF or UIP. In addition, features that are common in patients with fibrosis—such as intralobular reticular opacity, honeycombing, architectural distortion, and bronchiectasis—were uncommon in patients with LIP.
In one case that was misdiagnosed as IPF or UIP, the CT images showed extensive emphysema and cyst formation. The findings in this case were indistinguishable from honeycombing in a case of IPF or UIP. In our study, we correctly diagnosed IPF or UIP in only 63% of the IPF or UIP readings. This accuracy was lower than that in other studies (11,34,35). This was presumably due to patient selection bias because patients with characteristic findings of IPF or UIP on CT images did not undergo biopsy and therefore were not included in this study.
Our study had several limitations. First, this study was retrospective, it included only patients with IIPs, and the number of patients with each type of IIP was controlled to yield approximately the same number of patients in each group. Many cases of the most common IIPs were excluded; therefore, the cases in this study do not reflect the prevalence of the various IIPs in clinical practice (8). Namely, IPF or UIP is a common disease, and DIP, RB-ILD, and LIP are less common diseases. The cases included in our study were selected; therefore, generalizability of the results is limited. Second, we included only cases with histologic diagnoses. Thus, our study included a considerably higher proportion of patients with atypical findings of IPF or UIP than would be seen in daily clinical practice. Third, our study lacked pathologic correlation with specific CT findings, such as honeycombing and emphysema; therefore, there may have been a discrepancy between the CT and pathologic findings. Fourth, histologic diagnoses may have been subject to sampling bias. The different diagnoses are sometimes found between some specimens obtained in the same patient (36). Fifth, the thin-section CT images were obtained at various institutions with various protocols. The differences in section thickness and window level settings could have led the readers to misinterpret the CT findings.
In conclusion, typical CT images of IPF or UIP have a characteristic appearance that allows IPF or UIP to be differentiated from other types of chronic IIPs at thin-section CT in the majority of cases. However, some cases of IPF or UIP are difficult to differentiate from NSIP on thin-section CT images, and some chronic IIPs may mimic the appearance of IPF or UIP on thin-section CT images because of the presence of honeycombing. The most useful finding when differentiating between NSIP and IPF or UIP was the greater extent of honeycombing in cases of IPF or UIP.
|
ADVANCES IN KNOWLEDGE |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: DIP = desquamative interstitial pneumonia • IIP = idiopathic interstitial pneumonia • IPF = idiopathic pulmonary fibrosis • LIP = lymphoid interstitial pneumonia • NSIP = nonspecific interstitial pneumonia • RB-ILD = respiratory bronchiolitis–associated interstitial lung disease • UIP = usual interstitial pneumonia
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, H.S., T.J., Y.K., S.H., H.N.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, H.S., T.J., M.I., O.H.; clinical studies, H.S., T.J., K.I., H.T., Y.K., K.F., U.T., T.H., A.I., N.M., O.H., N.T., S.H., H.N.; statistical analysis, H.S., T.J., N.L.M.; and manuscript editing, H.S., T.J., J.N., M.I., S.H., H.N., N.L.M.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Vrielynck, T. Mamou-Mani, S. Emond, P. Scheinmann, F. Brunelle, and J. de Blic Diagnostic Value of High-Resolution CT in the Evaluation of Chronic Infiltrative Lung Disease in Children Am. J. Roentgenol., September 1, 2008; 191(3): 914 - 920. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Sumikawa, T. Johkoh, T. V. Colby, K. Ichikado, M. Suga, H. Taniguchi, Y. Kondoh, T. Ogura, H. Arakawa, K. Fujimoto, et al. Computed Tomography Findings in Pathological Usual Interstitial Pneumonia: Relationship to Survival Am. J. Respir. Crit. Care Med., February 15, 2008; 177(4): 433 - 439. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Arakawa, T. Johkoh, K. Honma, Y. Saito, Y. Fukushima, H. Shida, and N. Suganuma Chronic Interstitial Pneumonia in Silicosis and Mix-Dust Pneumoconiosis: Its Prevalence and Comparison of CT Findings With Idiopathic Pulmonary Fibrosis Chest, June 1, 2007; 131(6): 1870 - 1876. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
