HIGH-RESOLUTION COMPUTED TOMOGRAPHY IN THE EVALUATION OF FIBROSING ALVEOLITIS

0272-5231199 $8.00

IMAGING

+ .OO

HIGH-RESOLUTION COMPUTED TOMOGRAPHY IN THE EVALUATION OF FIBROSING ALVEOLITIS David M. Hansell, MD, FRCP, FRCR

The first investigations of the utility of CT scanning in diffuse lung disease were 60, lo4 and with undertaken in the mid-1980~,~, further technical refinements, the full potential of high-resolution computed tomography (HRCT) has been realized.5OS 59 In the past 12 years, numerous HRCT / pathologic correlative studies have increased understanding of the fine detail provided by HRCT images, particularly in the perplexing conditions that are included in the term fibrosing alveolitis. In the specific context of fibrosing lung disease, HRCT is increasingly used to detect early disease, when functional indices and plain chest radiography results are normal or equivocal. Comparative studies between the performance of HRCT and other noninvasive tests-particularly chest radiography-have repeatedly confirmed the superiority of HRCT for the diagnosis of fibrosing alveolitis, but caveats are necessary when considering such studies. By providing a precise assessment of disease pattern and extent, HRCT can provide information about the prognosis and reversibility in fibrosing lung disease. Further advantages of HRCT are the detection of complications and disorders associated with fibrosing alveolitis. In addition to the benefits that HRCT brings to the clinical management of patients with fibrosing alveolitis, new in-

sights about the natural history and behavior of fibrosing alveolitis have been gained from recent HRCT studies. For the purpose of this review, the term cryptogenic fibrosing alveolitis (synonymous with idiopathic pulmonary fibrosis) is used as the generic term that encompasses the various histopathologic subtypes (most frequently usual interstitial pneumonia). For clarity, occasional reference is made to the various subtypes that comprise the idiopathic interstitial pneumonias and readers wishing to acquaint themselves with the terminology that constitutes this confusing "alphabet soup" are referred to recent reviews.37,52, 56

TECHNICAL CONSIDERATIONS

The key determinants that contribute to the fine detail available on HRCT images are narrow beam collimation (section thickness) and a high spatial reconstruction algorithm.5l An understanding of these factors is useful in interpreting CT scans of patients with interstitial lung disease. A CT section, although apparently two-dimensional, has a third dimension of depth. The depth, or section thickness, is determined by the width of the slit through

From the Department of Radiology, Royal Brompton Hospital, London, England CLINICS IN CHEST MEDICINE VOLUME 20 * NUMBER 4 * DECEMBER 1999

739

740

HANSELL

Figure 1. CT sections through the upper lobes of a patient with a complex combination of chemotherapy-induced interstitial fibrosis and lymphangitis carcinomatosis (there is metastatic mediastinal disease from breast cancer). A, On the standard 10 mm-thick section there is amorphous opacification in the anterior segments of the upper lobes. 6,On the high-resolution CT (HRCT) 1.5 mm section, the morphology of the fibrotic and lymphangitic lung, in particular the small cystic air spaces and thickened interlobular septa, are shown to advantage because of less partial volume effect within the thin section.

which the x-ray beam passes (beam collimation). A CT section is made up of numerous picture elements (pixels)and, because the section has a thickness, each pixel has a volume. This three-dimensional element is referred to as a voxel. The computer calculates the average radiographic density of tissue within each voxel and the final CT image consists of a matrix of numerous voxels. The single density value of a voxel represents the average of the attenuation values of all the various structures within the voxel. The thicker the section, the greater the chance of structures of different densities being included within the voxel and the greater the chance of structures of different densities being included within

the voxel and the greater the averaging that occurs. This is known as the partial volume effect (Fig. 1).This is of relevance in the imaging of interstitial lung diseases because, for practical purposes, structures of less than approximately 0.3-mm width will not be individually identified on HRCT; in the case of fine intralobular fibrosis, the HRCT image will show an amorphous increase in lung density, rather than an obvious reticular pattern. This is an important point to remember when interpreting HRCT images that show ground-glass opacification. For routine HRCT scanning, a section thickness of 1 to 1.5 mm collimation is regarded as optimal. Greater section thickness reduces

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS

spatial resolution because of the partial volume effect. Differences between 1.5 mm and 3 mm collimation in the ability of CT to identify small structures are probably minor,59but subtle regional differences in the density of the lung parenchyma are more easily appreciated with 1 to 1.5 collimation images. The completeness with which a HRCT examination samples the lung depends primarily on the spacing between the thin sections. The size of the interval between sections is not standardized and ranges between 1 and 6 cm; a compromise of 2 cm has been adopted in many centers. In the specific context of fibrosing alveolitis, it seems that widely spaced sections reflect the severity of fibrosis demonstrated at a histopathologic level as ef-

741

fectively as closely interspaced When early or limited fibrosing alveolitis is suspected in individuals with an apparently normal chest radiograph, additional HRCT scans need to be performed in the prone position to prevent any confusion with the increased opacification seen in the dependent posterobasal segments of many normal individuals scanned in the usual supine position. The increased density seen in the posterior dependent lung in the supine position disappears in normal individuals when the scan is repeated at the same level with the patient in the prone position (Fig. 2). There is no advantage in scanning a patient in the prone position if there is obvious diffuse lung disease on a contemporaneous chest radiograph.88

Figure 2. Normal gravity-dependent phenomenon in an asbestosexposed individual with suspected interstitial fibrosis. A, In the supine position, there is subpleural opacification in the posterobasal segments, simulating fibrosing alveolitis. Note the bilateral pleural plaques indicating previous asbestos exposure. 6,In the prone position, the gravity-dependent opacification in the normal posterobasal segments clears.

742

HANSELL

In HRCT lung work, a high spatial frequency algorithm (also known as the edgeenhancing, sharp, or, formerly, bone algorithm)is used, which takes advantage of the inherently high-contrast environment of the lung. The high spatial frequency algorithm makes structures, such as linear and reticular opacities, visibly sharper. At the same time, it makes image noise, seen as a superimposed granularity, more conspicuous, but this artifact is rarely severe enough to obscure important detail. The preferred window settings at which images are viewed is a matter of individual preference and are influenced by the characteristics of a particular CT scanner. In general, a window level of between - 500 and -900 Hounsfield Units (HU) and a width of 1500 HU (range 1100-2000 HU) is recommended for patients with interstitial lung disease. The same imaging parameters applied to two different CT scanners may result in images of very different appearance. Although such discrepancies rarely cause diagnostic confusion, disease extent may appear different, and it is worth appreciating that such peculiarities exist between CT scanners. The radiation burden to patients from widely spaced HRCT sections is considerably less than conventional contiguous CT protocols, but the radiation dose is still relatively high and HRCT therefore should not be performed indiscriminately. A HRCT protocol of 1.5-mm sections at 10-mm intervals is approximately 6.5 times less than that delivered by a conventional CT examination (10-mm sections at 10-mm interval^).^^ The effective radiation dose of a standard 1.5 mm/lO mm HRCT examination is approximately 12 times that of a frontal and lateral chest radiograph. PATHOLOGIC ASPECTS OF FlBROSlNG ALVEOLlTlS

Understanding of fibrosing lung disease has been hindered by confusing pathologic terminology and imprecision in clinical diagnosis. Idiopathic pulmonary fibrosis and c y p togenic fibrosing alveolitis can be regarded as synonymous and these umbrella terms encompass the many subtypes of the so-called interstitial pneumonias, which are in a constant state of reclassification. UIP is the most common subtype. Specific histopathologic criteria need to be satisfied for the term UIP to be appropriately used, and there is increasing usage of the term nonspecific or nonclassifable

interstitial pneumonitis (NSIP or NCIP) for cases that do not fulfill the histopathologic criteria for UIP or the other interstitial pneumonias. It is possible that, with changes in selection of patients for lung biopsy (in part modulated by an increasing reliance on HRCT), UIP may no longer be the most frequent or “usual” diagnosis made by pathologists. The fibrosing alveolitides are the most common category of nongranulomatous and nonneoplastic chronic diffuse lung diseases found at open lung biopsy.14 At a pathologic level, these disorders are characterized by progressive alveolar wall fibrosis and variable intraalveolar and interstitial exudate.R Extensive pulmonary fibrosis represents a final common pathway of inflammation and repair that may be idiopathic or may result from any of a wide variety of insults. An early report of a disease resembling fibrosing alveolitis was made by HammanRich in 194425but it now seems probable that the patients described by Hamman-Rich had an acute interstitial pneumonitis, similar to acute respiratory distress ~yndrome,~ which is characterized by diffuse alveolar damage, hyaline membrane deposition, and a more 70 An important rapid clinical progre~sion.~~, clinical and histopathologic distinction became apparent in early studies of CFA: At open lung biopsy, appearances could be categorized as predominantly inflammatory (with cellular infiltration in the air spaces or interstitium) or predominantly fibrotic.’O, 7x, x6 The term UlP was introduced by Liebow& to denote the much more common fibrotic form of disease, characterized by progressive derangement of distal lung architecture and the formation of large cystic air spaces lined by bronchiolar epithelium (honeycomb lung). A key characteristic of UIP, which has recently been reiterated, is the apparent spatial and temporal heterogeneity of the lesions, such that foci of ”active” disease, mature fibrosis, and normal lung all coexist in an individual case.37 The much smaller subgroup of patients with DIP have an abundance of macrophages within the air spaces and interstitial infiltration by lymphocytes and monocytes; crucially, fibrosis is a relatively minor feature and the alveolar architecture is preserved in cases categorized as DIP/5 A clear histopathologic distinction between UIP and DIP may be absent in some patients, and there is frequently a lack of histopathologic uniformity between

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS

biopsies taken from different lobes in the same patient. Although most authorities now regard UIP and DIP as distinct diseases, with different natural histories,'O>44 others have argued that DIP represents an early stage of UIP rather than a separate entity,68,84 the basic hypothesis being that early inflammatory disease precedes and leads to mature fibrosis. In the absence of in vivo longitudinal pathologic studies, however, complete understanding of the evolution of fibrosing alveolitis remains elusive. More recently, a subset of cases that do not fulfill the pathologic criteria for the previously described interstitial pneumonias has been The main feature said to separate these cases from the other interstitial pneumonias is the apparently uniform temporal homogeneity of the pathologic process. Katzen~tein~~ has proposed that these cases are labelled NSIP. The clinical utility of this further subdivision of the interstitial pneumonias is not yet fully established, but patients assigned the label NSIP appear to have a more benign clinical course compared with patients with UIP.9 The proportion of cases designated as NSIP that merely represent a sampling error (that is, a more representative biopsy would change the pathologic diagnosis to, for example, UIP or hypersensitivity pneumonitis) is unknown. HIGH-RESOLUTION COMPUTED TOMOGRAPHY FEATURES OF FIBROSING ALVEOLITIS AND INTERPRETIVE CONSIDERATIONS

The typical radiographic pattern of established fibrosing alveolitis consists of reticulonodular shadowing concentrated at the lung bases, with the reticular component becoming more pronounced as the disease progresses. The most distinctive radiographic feature of fibrosing alveolitis is its predilection for the subpleural and lower zones of the lungs. In a minority of patients, the predominant pattern at presentation is ground-glass opacification, which may reflect a more cellular (DIP end of the spectrum) pathology. The radiographic pattern alone, however, is not a reliable predictor of the histopathologic subtype (Fig. 3). Moreover, there is considerable variation between observers in assigning a radiographic pattern in cases of fibrosing alve01itis.l~ Because of the basal distribution of early fibrosing alveolitis, CT is able to show rela-

743

Figure 3. Frontal chest radiograph of a patient with the clinical features of fibrosing alveolitis. There is a fine reticular pattern at the lung bases that extends into the periphery of the midzones. Open lung biopsy specimen showed nonspecific interstitial pneumonitis (NSIP).

tively limited disease in the radiographically inaccessible costophrenic recesses. The earliest CT sign of fibrosing alveolitis is faint subpleural opacification in the posterobasal segments of the lower lobes (Fig. 4). The peripheral subpleural disposition of disease is a striking characteristic of fibrosing alveolitis on CT and it is this distribution that is often regarded as diagn~stic.~~, 85 As the interstitial fibrosis progresses, a reticular pattern containing small cystic air spaces becomes evidenP7, (Fig. 5). In fibrosing alveolitis, thickened interlobular septa are not a pronounced feature, probably because of the severe accompanying architectural distortion. The initially small cystic air spaces later form a more obvious honeycomb patternz and, in some patients, may enlarge considerably, possibly because of a check valve mechanism54 (Fig. 6 ) . This pattern is irreversible and tends to progress, even in patients receiving lowdose steroid therapy.2 Advanced fibrotic destruction (honeycomb lung) may be so widespread that the typical peripheral distribution is lost. Nevertheless, even in advanced disease, observers are frequently successful in distinguishing cases of end-stage fibrosing alveolitis from other advanced destructive lung diseases on the basis of HRCT appearance^.^^ Much attention has been given to the two CT patterns of disease that are most frequently encountered in fibrosing alveolitisnamely, a reticular pattern and ground-glass opacification. When a reticular pattern pre@

744

HANSELL

Figure 4. HRCT scan (prone position) through the lung bases of a patient with an apparently normal chest radiograph. The subpleural distribution of a fine reticular pattern is typical of established fibrosing alveolitis.

Figure 5. The typical HRCT appearances of fibrosing alveolitis (biopsy-proven usual interstitial pneumonitis). There is a reticular pattern containing small cystic air spaces in a predominantly subpleural distribution. The less-extensive component of ground-glass opacification probably reflects fine intralobular fibrosis in this case. Prolonged steroid treatment was responsible for the abundance of mediastinal fat.

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS

745

Figure 6. Large cystic air spaces within honeycomb lung in a patient with advanced fibrosing alveolitis.

dominates, histopathologic correlation has shown this to correspond to established fibrosis.58,63,92 In contrast, in cases at the cellular or active alveolitis end of the spectrum (DIPtype), ground-glass opacification is the dominant abnormality on CT.29,58 It is worth emphasizing that this cellular form of fibrosing alveolitis is seen in a minority of patients at presentation, probably rather less than 15%. In this rarer cellular form of interstittial pneumonia, the typical subpleural distribution is often rendering the CT appearances nonspecific (Fig. 7). In some patients, lung biopsies from different sites may show fea-

tures of pure UIP and pure DIP and this may be reflected in the HRCT appearances (Fig. 8). Although areas of ground-glass opacification can be identified in most cases of fibrosing alveolitis, interspersed among areas of a reticular pattern, this often represents fine intralobular fibrosis below the limits of resolution of HRCT. When there is dilatation and distortion of the bronchi within areas of ground-glass opacification, this should be regarded as indirect evidence of fine interstitial fibrosis (so-called traction bronchiectasis or bronchioZectasis)n (Fig. 9). If dilatation of the bronchi cannot be ascribed to traction

Figure 7. Extensive patchy ground-glass opacification in a patient with biopsy-proven desquamative interstitial pneumonia (DIP). In contrast to the typical subpleural distribution of UIP, the location of disease in this case is random with no subpleural predilection.

746

HANSELL

Figure 8. HRCT section through the midzones of a patient before lung biopsy. A specimen from the right upper lobe showed DIP on histopathologic examination, whereas a sample from the right lower lobe showed a UIP pattern (note the fine honeycomb pattern posteriorly in the apical segment of the right lower lobe).

Figure 9. Close-up view of HRCT of the right lung (patient prone) to show dilatation of bronchi (arrowheads) within lung that is of abnormally increased attenuation (ground-glass opacification). Open-lung biopsy specimen showed diffuse fine interstitial fibrosis with no disruption of the alveolar architecture.

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS

747

Figure 10. HRCT scan through the lower lobes of a patient with rheumatoid arthritis. There is a faint but definite subpleural reticular pattern, indicating limited fibrosing alveolitis. Many of the subsegmental bronchi are dilated and thick walled. The bronchial dilatation is not a consequence of “traction bronchiectasis.” Airflow limitation was the dominant functional abnormality in this patient.

bronchiectasis-that is, abnormally dilated airways can be identified in normal lung, distant from a reticular pattern-a diagnosis of rheumatoid arthritis should be entertained%,73 (Fig. 10). In biopsy-proven cases of NSIP, no characteristic HRCT pattern or distribution has been described so far. The first descriptive CT study of patients with NSIP reported that ground-glass opacification, either alone or in combination with other parenchymal patterns (most frequently patchy air space consolidation), was the most common fh1ding.6~Honeycombing was not a feature and six of seven patients showed complete or partial resolution of the CT abnormalities at just over 1 year. A more recent multicenter study of 50 patients with NSIP confirmed that ground-glass opacification was the most prevalent (75%) CT abnormality and this was subpleural and basal in half the cases.3oIn contrast to the earlier smaller honeycombing was present in just over one quarter of patients; the majority of these cases showed honeycombing in the same distribution as UIP. Other CT features included irregular linear opacities (34%), air space consolidation (16%), and poorly defined nodules (14%). Interestingly, when observers were asked to assign a likely diagnosis on the basis of the CT appearance alone, UIP was the diagnosis chosen in 32% of cases, 22% of cases were thought to

be consistent with the existing CT descriptions of NSIP, 20% were assigned a diagnosis of hypersensitivity pneumonitis, and 14% were thought to be cases of bronchiolitis obliterans organizing pneumonia (BOOP).The observers were aware of the purpose of the study and the composition of the study group, so these results need to be interpreted with caution; however, the heterogeneity of the suggested diagnoses indicates that NSIP does not appear to have a consistent CT pattern (Fig. ll), which is at variance with the suggestion that NSIP represents a distinctive clinicopathologic entity.9,37 Moderately enlarged mediastinal lymph nodes are a frequent finding on CT in patients with fibrosing lung disease,6l whether lone cryptogenic fibrosing alveolitis5or in association with systemic sclerosis,*, 90 asbest~sis,~~ or rheumatoid arthritis4s(Fig. 12). The extent of pulmonary involvement, not the pattern of parenchymal disease (ground-glass opacification versus a reticular pattern) seems to be the main determinant of the prevalence of enlarged mediastinal lymph nodes.22, There may also be some connection between the state of disease activity and mediastinal lymphadenopathy; patients with fibrosing alveolitis on steroid therapy have a significantly lower prevalence of enlarged mediastinal lymph nodes than patients not receiving treatment.19 Because most individuals who develop

748

HANSELL

Figure 11. Two examples of biopsy-proven NSIP. Neither case shows the typical subpleural distribution of UP, or any other characteristic HRCT features. A, Patchy focal areas of ground-glass opacification with a reticular pattern and some thickened interlobular septa. B, In this case there is more extensive ground-glass opacification. There is mild dilatation of some of the bronchi within the abnormal lung and loss of volume in both lower lobes, reflecting interstitial fibrosis.

cryptogenic fibrosing alveolitis are smokers, centrilobular emphysema frequently coexists. As a result, loss of volume in the lower lobes caused by fibrosing alveolitis may be counteracted, both functionally and radiographically, by coexisting emphysema in the upper lobes. Such individuals are characterized by the absence of airflow obstruction but a greatly reduced transfer factor and gross hypoxia attributable to summation of the effects of lo* emphysema and fibrosing al~eolitis.'~, Pleural disease is not a feature of lone cryptogenic fibrosing alveolitis. Dense subpleural pulmonary fibrosis occasionally mimics pleural thickening on chest radiography. When there is clear evidence of pleural effu-

sion or thickening, an explanation must be sought. Asbestos-induced pleuropulmonary disease, a connective tissue disease, a subpleural bronchial carcinoma, pulmonary embolism, or supervening heart failure are some of many possible explanations. Some individuals have an abundance of intrathoracic fat that may be related to steroid therapy; deposition of extrapleural fat can produce a convincing radiographic appearance of bilateral pleural disease that is usually symmetrically disposed along the lateral chest walls (Fig. 13). Because of the potential significance of pleural disease in the context of fibrosing alveolitis, CT may be needed to clarify the nature of suspected pleural disease.

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS

Figure 12. Several moderately enlarged large mediastinal lymph nodes (arrowheads) (the largest measuring > I cm in short axis) in a patient with UIP. These lymph nodes are thought to be reactive and are a common finding on CT in patients with fibrosing alveolitis.

Figure 13. A, Close-up view of the chest radiograph of a patient with a right middle lobe bronchogenic carcinoma on biopsy and biopsy-proven UIP. Note the apparent pleural thickening (arrows). B, On CT there is right hilar lymphadenopathy and an excess of extrapleural fat (arrowheads) that is responsible for the radiographic appearance of pleural thickening.

749

750

HANSELL

The ability of observers to reproducibly categorize HRCT patterns is important, particularly in investigations of disease activity that rely on the distinction between reticular and ground-glass patterns. Observer variation in grading patterns in fibrosing alveolitis was evaluated in a study that included two experienced radiologists and two inexperienced trainee^.'^ Patterns were assigned according to the three-point system (predominant reticular pattern, mixed pattern, predominant ground-glass pattern) in 126 CT scans and 108 concurrent chest radiographs. At least three of the observers agreed on the CT pattern in 81% of CT examinations but in only 54% of chest radiographs; the K coefficient of agreement was barely clinically acceptable for CT (K = 0.47), but unacceptable for chest radiography (K = 0.16). Observer agreement is probably understated in this study because of the inclusion of two inexperienced observers (later studies reported higher observer agreement with K values ranging from 0.51 to 0.8340,97). The confidence of the observers in assigning a pattern was higher in cases with more extensive disease (P
ROLE OF HIGH-RESOLUTION COMPUTED TOMOGRAPHY IN THE DIAGNOSIS OF FlBROSlNG ALVEOLITIS

In terms of sensitivitv and specificity, HRCT represents a significant advance over chest radiography for the diagnosis of diffuse interstitial lung disease.24,49, 58, 85, 97 In the specific context of fibrosing alveolitis, however, the diagnostic advantages of HRCT over chest radiography may not be as great as some of these studies suggest. The overall diagnostic sensitivity of CT for the diagnosis of fibrosing alveolitis can be extracted from five studies7,24, 49, 64, 85 that comprise a total of 501 patients with a variety of interstitial lung diseases, 145 of whom had a final diagnosis of cryptogenic fibrosing alveolitis (CFA). A correct first-choice diagnosis of CFA was made in 84% of cases on CT versus 73% on

chest radiography. A formal meta-analysis of various, superficially similar, studies is not possible because of differences in CT protocol, patient selection, and observer experience (one study62deliberately included an inexperienced observer). Nevertheless, in all the relevant studies to date, the trend has been toward increased diagnostic accuracy with HRCT compared with chest radiography. The relatively small diagnostic advantage of HRCT is occasionally overlooked when interpreting these studies, however; a recent fact sheet circulated to chest physicians in the United Kingdom stated that ” . . . the appearances on high-resolution CT scanning, which is much more accurate diagnostically than the chest X-ray, may avoid the need for invasive procedures in many cases” (italics this aut h o r ’ ~ ) : ~The ~ basic radiographic pattern of reticulonodular shadowing at the lung bases remains, in the appropriate clinical setting, highly suggestive of the diagnosis of fibrosing alveolitis. The value of HRCT is largely one of added confidence, and this factor has not been evaluated in detail, partly because of the relatively crude statistical tools that are available for evaluating the subtle shades of an observer’s level of confidence. In the few series in which observer confidence was assessed,24, 49, 64 a confident diagnosis was made more often on CT than on chest radiography and, even more importantly, a confident CT diagnosis was almost always correct. The landmark study of Mathieson et a149showed that a confident diagnosis of various diffuse lung diseases was made by three observers at least twice as often on CT as on chest radiography. Furthermore, the accuracy of a confident firstchoice diagnosis on CT was !%%, compared with 77% for chest radiography. When the first-choice diagnosis was not made with high confidence, however, the advantage of CT (60% accuracy in 51% of cases) over chest radiography (51% accuracy in 77% of cases) was greatly diminished. The confidence of a radiologist or clinician, which is largely determined by experience (that is, case throughput), is a factor that should not be underestimated. Given that the study by Mathieson et a1 was performed 10 years ago, it seems probable that the results of this widely cited study understate the diagnostic accuracy of HRCT in contemporary practice. Investigations of the diagnostic accuracy of HRCT have mostly been confined to comparisons of the sensitivity of HRCT versus chest radiography. Because of this type of design,

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS

such studies do not reflect the complete diagnostic process, in which imaging represents only one small part of the puzzle. The ultimate effectiveness of HRCT depends both on the expertise and confidence of the observer and upon pretest probabilities. When the clinical and radiographic picture are thought to be typical of fibrosing alveolitis, for example, a HRCT showing features consistent with fibrosing alveolitis adds little to the diagnostic process. This scenario is in contrast to the situation in which, for example, the chest radiograph shows a slightly enlarged heart and diffuse pulmonary shadowing with possible upper lobe blood diversion. In this situation, HRCT may be extremely valuable if it shows the features of fibrosing alveolitis and, in so doing, effectively excludes pulmonary edema. The goal of 100% sensitivity and specificity is clearly an inappropriate expectation for HRCT in the diagnosis of fibrosing alveolitis and most studies have encouraged more modest hopes. In this context, it is worth appreciating that an uncritical reliance on histopathologic examination of lung biopsy is also erroneous and observer variation and sampling error are not infrequent problems for pathologists. A single biopsy sample may not be representative of the predominant pathologic process elsewhere in the lungs; furthermore, two samples from different lobes may

751

yield different histopathologic diagnoses (see Fig. 8). A single study has addressed the diagnostic accuracy of CT when the appearances are regarded as typical of fibrosing alveolitis.= In a study group of 86 patients with a variety of diffuse lung diseases, including 45 patients with a final diagnosis of fibrosing alveolitis, CT features thought to be characteristic of CFA had a sensitivity of 88%, specificity of 93%, and accuracy of 91%. A shortcoming of this study was that only 69% of the patients had histologic proof of the diagnosis. Nevertheless, when the analysis was confined to patients with a final histologic diagnosis, the results were similar. Data about the specificity of CT appearances of fibrosing alveolitis are less available than for the sensitivity; in the study of Tung et al,85the false-positive rate was 13% and occurred in cases of hypersensitivity pneumonitis, BOOP, sarcoidosis, and pulmonary eosinophilia.85Arguably, in clinical practice, the pretest probability for the diagnosis of fibrosing alveolitis in at least some of these cases would have been low. Nevertheless, pulmonary sarcoidosis, in keeping with its reputation as a great mimic, may sometimes simulate the basal and peripheral pattern of fibrosing alveolitis (although thickening of the interlobular septa, rather than a honeycomb pattern, is the rule in these cases)65(Fig. 14). The overlap in HRCT ap-

Figure 14. Amorphous subpleural opacification and a reticular pattern, within which some thickened interlobular septa are visible. There are no cystic air spaces or honeycombing, but the distribution of disease is similar to that found in fibrosing alveolitis. Openlung biopsy revealed sarcoidosis. In this case, there was obvious symmetrical hilar lymphadenopathy but none of the typical parenchymal changes of pulmonary sarcoidosis in the upper zones.

752

HANSELL

pearances between BOOP and fibrosing alveolitis is consonant with the fact that both pathologic processes coexist in some individuals, particularly in patients with polymyo81 A study that sitis and derrnatomyositi~.5~, specifically investigated the ability of HRCT to differentiate between hypersensitivity pneumonitis and CFA showed HRCT to be poor at this task, although the pretest probabilities for these two diagnoses was not given.47 A major problem in interpreting the results of observer performance studies is the tertiary referral nature of the study groups. This type of bias is almost inevitable for observer studies that by definition, require a large group of subjects. For such studies to be repeated in the ”real world,” a different problem arises-that of diagnostic veracity. Even in tertiary referral institutions, the lung biopsy rate for securing a histopathologic diagnosis 35 It of fibrosing alveolitis is relatively seems unlikely that the application of goldstandard tissue diagnosis in secondary referral populations is higher. In summary, currently available CT studies suffer from a variety of shortcomings that include: (1)selection bias peculiar to a tertiary referral institution, (2) nonrepresentative observer experience and expertise, (3) underrecognition of the ability of CT to exclude, rather than confirm, specific diagnoses, and (4) the effect of pretest clinical probability. Despite the shortcomings of these observational studies, HRCT is now a widely accepted tool for the diagnosis of diffuse interstitial lung disease, to the extent that HRCT is increasingly used as a filter for the evaluation of patients with fibrosing lung disease who are considered for lung biopsy?, 71, 74 In many centers, patients are more likely to be submitted for lung biopsy if the HRCT appearances are atypical for fibrosing alveolitis or are at variance with the clinical picture. It seems that HRCT is increasingly being used to support, if not confirm, the diagnosis of fibrosing alveolitis. Although this may be a matter of consternation for purist practitioners of evidencebased medicine, history is repeating itself: HRCT replaced bronchography for the diagnosis of bronchiectasis by stealth, despite the lack of any evidence HRCT was better (or worse) at this particular task. In this context, the key question (which is at present unanswered) is: How often will lung biopsy yield a different diagnosis when the clinical and HRCT picture is that of fibrosing alveolitis?

EVALUATION OF DISEASE REVERSlBlLlTY AND PROGNOSIS WITH HIGH-RESOLUTION COMPUTED TOMOGRAPHY

As a result of the early pathologic-HRCT correlative studies, the idea that the macroscopic patterns seen on HRCT might reflect the histopathologic morphology emerged. In general terms, a reticular pattern on HRCT almost invariably reflects significant disease (in the case of fibrosing alveolitis, established fibrosis) whereas ground-glass opacification, with some important caveats, usually repre42, 58, 92, 99 Historically, sents reversible disease.28, histopathologic examination of lung biopsy specimens was regarded as the most reliable means of assessing “disease activity” and therefore, responsiveness to therapy and survival.lO,78, 86 In contrast, chest radiography,’02 pulmonary function testing,” loogallium scanning,6699mtechnetiumdiethylenetriaminepenta-acetic acid scanning,93or bronchoalveolar lavage3Ido not accurately discriminate between established fibrotic disease and reversible disease. A key study by Muller et a158showed that it was possible to assess disease activity based In this study, the on CT appearances scoring systems for both the CT and pathologic patterns were constructed to reflect the amount of increased parenchymal cellularity and therefore, disease activity. Because the main determinant of response to treatment by these patients is the proportion of active cellular disease to establish fibrosis,lQ86 some subsequent CT studies employed a scoring system to take account of the relative amounts of ground-glass opacification (reflecting increased cellularity) and reticular pattern (reflecting established fibrosis).92The interpretation of ground-glass opacification requires some care. Although the finding of widespread ground-glass opacification on HRCT usually represents reversible patholo ~ Yan, ~important ~ observation is whether there is dilatation of the airways within these areas. This phenomenon implies the presence of established fine intralobular fibrosis” (see Fig. 9). Given that many studies have confirmed the correlation between the HRCT appearance of a reticular pattern and established 63, 92 the quesfibrosis at a microscopic tion arises whether HRCT appearances in isolation can provide useful prognostic information. In a population of 142 patients (cryptogenic fibrosing alveolitis n = 76,

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS

0--,

-4-

I

CT grade 1 ( n = 8 )

m-¤ CT grade 2 ( n = 18)

&-A

0

753

I

CT grade 3 ( n = 50)

I

Figure 15. Survival curves of patients with fibrosing alveolitis. CT grade 1 (ground-glass opacification pattern predominant); CT grade 2 (equally extensive ground-glass opacification and reticular pattern); CT grade 3 (reticular pattern predominant). (From Wells AU, Hansel1 DM, Rubens MB, et al: The predictive value of thin-section computed tomography in fibrosing alveolitis. Am Rev Respir Dis 148:1076-1082, 1993; with permission.)

fibrosing alveolitis associated with systemic sclerosis n = 66), survival at 4 years was 100% in patients with predominantly ground-glass opacification ( n = 8) on CT, and somewhat more likely in patients with a mixed pattern (45%, n=18) than in those with a predominantly reticular pattern (15%, n = 50) [P<0.001]97(Fig. 15). Response to treatment, as judged by an improvement in lung function, was seen most frequently in patients with predominantly ground-glass opacification, and more frequently with a mixed pattern than when a reticular pattern was the dominant CT feature. More recent studies have confirmed that patients with an extensive reticular pattern on CT are much less likely to respond to treatment.u, *05 Furthermore, long-term survival is reflected as accurately by the extent of a reticular pattern on CT as by the pathologic score of the severity of fibrosis.w Follow-up CT scans provide an opportunity to study the natural and treated history of fibrosing alveolitis. In patients with fibrosing alveolitis, changes in extent of disease are caused by the resolution of ground-glass opacification or the progression of a reticular pattern; a reticular pattern per se does not

diminish in extent.” 99 An important point is that the prognostic significance of groundglass opacification depends on the extent of an associated reticular pattern. The poor outcome seen in most patients with mixed CT appearances probably reflects the fact that these individuals have predominantly fibrotic disease, consisting of both fine intralobular fibrosis and coarse fibrosis (the former producing ground-glass opacification and the latter, a macroscopic reticular pattern).&,72 Many studies imply that ground-glass opacification precedes a reticular pattern in the same area of lung.” 28, 82 Descriptive CT studies of the distribution of DIP versus UIP provide evidence that the cellular form of the disease does not inevitably progress to fibrotic honey comb, however. A predominantly subpleural distribution of disease is found in approximately 60% of cases of DIP29versus virtually all cases of The detection of serial changes in disease extent and activity in patients with fibrosing alveolitis by clinical and functional measures is difficult.66The superior sensitivity of CT alone in detecting serial change, compared with either pulmonary functional tests or chest radiography, holds only when anatomi-

754

HANSELL

cally comparable sections are obtained.95Because of the difficulty in obtaining exactly comparable CT sections, it is likely that pulmonary function tests will remain the mainstay for the routine monitoring of disease extent in patients with fibrosing alveolitis. A small proportion of patients with fibrosing alveolitis (probably fewer than 10%) show a dramatic flare-up of disease activity with accelerated deterioration in their clinical status. CT reveals superimposed multifocal consolidation and increased peripheral or generalized parenchymal opacification’ (Fig. 16). These changes, which likely reflect diffuse al-

veolar damage, are readily apparent on serial chest radiographs. FUNCTIONAL AND PATHOPHYSIOLOGICINSIGHTS DERIVED FROM HIGH-RESOLUTION COMPUTED TOMOGRAPHY Many recent investigations have relied on the ability of HRCT to map accurately the extent of patterns of fibrosing lung disease. The most common technique for quantitation of disease extent is by subjective visual esti-

Figure 16. HRCT scan through the midzones of a patient in a prone position with established fibrosing alveolitis (UIP on lung biopsy). A, A coarse reticular pattern with obvious traction bronchiectasis in the middle lobe typical of severe fibrosing alveolitis. 6, Two months later, after a 3-week history of rapidly progressive dyspnea, the HRCT shows a superimposed pattern of ground-glass opacification and dense areas of parenchymal opacification. The patient died shortly afterward despite intensive immunosuppressive therapy. The autopsy examination and rapid clinical deterioration were consistent with “accelerated fibrosing alveolitis.”

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS

mation expressed either as the percentage of lung involvement or in terms of a categorical scoring system (at either a lobar or zonal 98 Although observer variation and other sources of error may be a problem, such a subjective approach is quick and simple. Nevertheless, some of the more subtle abnormalities caused by interstitial fibrosis may escape visual detection and may be detected only by analyses of CT density measurem e n t ~More . ~ ~ sophisticated ~ approaches to the quantification of interstitial lung disease include analysis of CT density values15,27 and texture of the lung parenchyma.'6 In patients with pulmonary fibrosis, the density histogram is characteristically more peaked (kurtotic) and skewed to the left than the normal distrib~tion.~~ An operator-independent technique that relies on fractal analysis of CT sections has been reported to have good accuracy for the quantification of fibrosing alveoiitis.75 At a functional level, cryptogenic fibrosing alveolitis may be considered the archetypal restrictive lung disease, characterized by reduced lung volume and a decrease in gasdiffusing capacity. It is these indices that are routinely used in clinical practice to make an assessment of the extent of disease and monitor progress. The superiority of one physiologic measure over another in accurately reflecting the true extent of lung disease remains controversial, however. A combined clinical, radiographic, and physiologic (CRP) scoring system has been devised that attempts to improve the accuracy of estimating

755

disease extent (or severity).89The choice of pulmonary function parameters used in the CRP system, and their weighting, however, were based on the strength of previously reported correlations between pulmonary function tests and measures of structural derangement at a his topathologic level. Such correlations take no account of the regional inhomogeneity of fibrosing lung disease and account for the discrepant results between studies that have sought relationships between histopathologic severity of disease and physiologic disturbance.12,2o This is in contrast to more recent work that has shown much stronger correlations between pulmonary function tests (PFT) data and global assessment of disease extent with CT.98,loo Many patients with cryptogenic fibrosing alveolitis are cigarette smokers and emphysema therefore is a common accompaniment. This combination has long been recognized as being responsible for spuriously normal PFT results, in terms of preserved lung volume^.'^, lol On HRCT, the morphologic differences between emphysematous and fibrotic lung can usually be readily identified; furthermore, smoker's centrilobular emphysema is predominantly upper zone in distribution, compared with the basal predilection of fibrosing alveolitis. Nevertheless, there is often a "hinterland in the midzones where a clear distinction between the two processes becomes difficult or impossible (Fig. 17). The relationship between emphysema and interstitial fibrosis is probably more complex than earlier exclusive definitions allowed.79 It

Figure 17. HRCT scan through the upper lobes showing the combination of fibrosing alveolitis (subpleural reticular pattern, arrowheads) and centrilobular emphysema (most obvious in the right upper lobe, arrows).

756

HANSELL

seems likely that at least some patients with "pure" emphysema have a fibrotic componenP and this may be obvious on HRCT as a reticular element within emphysematous lunga3 In addition to the confounding effect of coexisting emphysema, variable ventilation of fibrotic lung composed of cystic air spaces is also likely to weaken the relationship between disease extent and volumetric PFT indices. Although it seems likely that the majority of such cystic "dead spaces" are ventilated,*Oothers may air-t~ap.5~ For a given extent of fibrotic lung on CT, total gas diffusing capacity is depressed, reflecting its lack of perfusion (irrespective of whether or not the affected lung is ventilated). It is the differing capacity of fibrotic lung to air-trap within and between patients that prevents measures of lung volumes being reliable in estimating lung involvement in fibrosing a l v e ~ l i t i s . ~ ~ Clinical and pathologic studies usually consider lone cryptogenic fibrosing alveolitis and the fibrosing alveolitis associated with connective tissue diseases as identical. Differences between the fibrosing alveolitides do exist, however. In contrast to lone cryptogenic fibrosing alveolitis, the fibrosing alveolitis encountered in patients with polymyositis or dermatomyositis is characterized by areas of air space consolidation (organizing pneumonia at a pathologic level) in half of patients (Fig. 18) and less extensive honeycombing

than is usually found in patients with lone cryptogenic fibrosing alveoliti~.~~ There is evidence that, despite radiographic and pathologic similarities between the fibrosing alveolitis associated with connective tissue diseases and lone cryptogenic fibrosing alveolitis,26the various diseases have different biologic behavior. CT provides a means of precisely matching the disease extent and HRCT pattern between patients with these conditions. When these, and other variables, are taken into account, the risk estimate for mortality in lone cryptogenic fibrosing alveolitis is approximately four times that associated with the fibrosing alveolitis of systemic sclero~is.~~ Although not conclusive, there is a nonsignificant trend toward a coarser reticular pattern in lone cryptogenic fibrosing alveolitis compared with the fibrosing alveolitis found in patients with systemic sclerosis." Another study did not confirm differences in the quality of the reticular pattern between these two suggesting that if differences exist, they are subtle. Interestingly, the pattern of functional impairment between lone CFA and fibrosing alveolitis associated with systemic sclerosis seems to differ. There appears to be more severe anatomic shunting in lone cryptogenic fibrosing alveolitis.96Furthermore, when disease extent is controlled for, analysis of bronchoalveolar lavage reveals differences between cryptogenic fibrosing alveolitis and the fibrosing alveolitis associated

Figure 18. HRCT scan through the lower lobes of a patient with dermatomyositis. The combination of dense opacification (somewhat bronchocentric in the right lower lobe) and a reticular pattern is typical of this condition. Lung biopsy revealed intra-alveolar granulation tissue (organizing pneumonia) that was being incorporated into the alveolar walls. In addition, there were features of established interstitial fibrosis.

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS

with systemic sclerosis. The eosinophil count is significantly higher in cryptogenic fibrosing alveolitis, raising the possibility that the eosinophil may be an important effector cell.94 CONCLUSION

The introduction of HRCT has had an impact on the diagnosis and evaluation of disease reversibility in fibrosing alveolitis. The added confidence that HRCT brings to a clinical diagnosis of many diffuse lung diseases, including fibrosing alveolitis, has resulted in fewer patients being submitted for confirmatory lung biopsy in many centers. Nevertheless, HRCT should not necessarily be regarded as the final diagnostic arbiter and an understanding of the limitations of HRCT should result in its more appropriate clinical use. From relatively simple information about disease extent and patterns on HRCT much has been learned about the behavior of fibrosing alveolitis and further insights from HRCT studies can be anticipated. References 1. Akira M, Hamada H, Sakatani M, et al: CT findings during phase of accelerated deterioration in patients with idiopathic pulmonary fibrosis. AJR Am J Roentgenol 168:79433, 1997 2. Akira M, Sakatani M, Ueda E: Idiopathic pulmonary fibrosis: Progression of honeycombing at thin-section CT. Radiology 189:687-691, 1993 3. American Thoracic Society Idiopathic Pulmonary Fibrosis Statement Committee: Idiopathic pulmonary fibrosis: Guidelines for diagnosis and treatment. Am J Respir Crit Care Med, 1999, in press 4. Askin FB: Acute interstitial pneumonia: Histopathologic patterns of acute lung injury and the Hamman-Rich syndrome revisited. Radiology 188:620621, 1993 5. Bergin C, Castellino RA: Mediastinal lymph node enlargement on CT scans in patients with usual interstitial pneumonitis. AJR Am J Roentgenol 154251-254, 1990 6. Bergin CJ, Miiller NL: CT in the diagnosis of interstitial lung disease. AJR Am J Roentgenol 145:505510, 1985 7. Bergin CJ, Coblentz CL, Chiles C, et ak Chronic lung diseases: Specific diagnosis by using CT. Am J Roentgenol 152:1183-1188, 1989 8. Bhalla M, Silver RM, Shepard JA, et al: Chest CT in patients with sclerodenna: Prevalence of asymptomatic esophageal dilatation and mediastinal lymphadenopathy. AJR Am J Roentgenol 161269-272,1993 Edwin MK, et a1 Prognostic 9. Bjoraker JA, Ryu JH, sigruficance of histopathologic subsets in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 157199-203, 1998 10. Carrington CB, Gaensler EA, Coutu RE, et ak Natu-

757

ral history and treated course of usual and desquamative interstitial pneumonia. N Engl J Med 298:801-809, 1978 11. Chan TY, Hansell DM, Rubens MB, et al: Cryptogenic fibrosing alveolitis and the fibrosing alveolitis of systemic sclerosis: Morphological differences on computed tomographic scans. Thorax 52265270, 1997 12. Chinet T, Jaubert F, Dusser D, et al: Effects of inflammation and fibrosis on pulmonary function in diffuse lung fibrosis. Thorax 45:67%78, 1990 13. Collins CD, Wells AU, Hansell DM, et al: Observer variation in pattern type and extent of disease in fibrosing alveolitis on thin section computed tomography and chest radiography. Clin Radiol 49:236240, 1994 14. Corrin B: Pathology of interstitial lung disease. In Harasawa M, Fukuchi Y, Morinari H (eds): Interstitial Pneumonia of Unknown Etiology. Tokyo, University of Tokyo Press, 1989, pp 149-168 15. Coxson HO, Hogg JC, Mayo JR, et a1 Quantification of idiopathic pulmonary fibrosis using computed tomography and histology. Am J Respir Crit Care Med 155:1649-1656,1997 16. Delorme S, Keller-Reichenbecher MA, Zuna I, et al: Usual interstitial pneumonia. Quantitative assessment of high-resolution computed tomography findings by computer-assisted texture-based image analysis. Invest Radiol 32:566-574, 1997 17. Doherty MJ, Pearson MG, O’Grady EA, et al: Cryptogenic fibrosing alveolitis with preserved lung volumes. Thorax 52:998-1002, 1997 18. Du Bois RM: Diffuse lung disease: An approach to management. BMJ 309:175-179,1994 19. Franquet T, Gimenez A, Alegret X, et al: Mediastinal lymphadenopathy in cryptogenic fibrosing alveolitis: The effect of steroid therapy on the prevalence of nodal enlargement. Clin Radiol 53:435-438, 1998 20. Fulmer DG, Roberts WC, von Gal ER, et al: Morphologic-physiologic correlates of the severity of fibrosis and degree of cellularity in idiopathic pulmonary fibrosis. J Clin Invest 63:665-676,1979 21. Gamsu G, Salmon CJ, Warnock ML, et al: CT quantification of interstitial fibrosis in patients with asbestosis: A comparison of two methods. AJR Am J Roentgenol 164:63-68, 1995 22. Garber SJ, Wells AU, Du Bois RM, et al: Enlarged mediastinal lymph nodes in the fibrosing alveolitis of systemic sclerosis. Br J Radiol 66:983-986, 1992 23. Gay SE, Kazerooni EA, Toews GB, et al: Idiopathic pulmonary fibrosis: Predicting response to therapy and survival. Am J Respir Crit Care Med 15710631072, 1998 24. Grenier P, Chevret S, Beigelman C, et al: Chronic diffuse infiltrative lung disease: Determination of the diagnostic value of clinical data, chest radiography, and CT and Bayesian analysis. Radiology 191:383-390, 1994 25. Hamman L, Rich AR Acute diffuse interstitial fibrosis of the lungs. Bulletin of Johns Hopkins Hospital 74:177-212, 1944 26. Harrison NK, Myers AR, Corrin BE, et al: Histopathology of intersitial lung disease in systemic sclerosis (SS): Comparison with cryptogenic fibrosing alveolitis and implications for pathogenesis. Thorax 44:346, 1989 27. Hartley PG, Galvin JR, Hunninghake GW, et a1 High-resolution CT-derived measures of lung density are valid indexes of intersitial lung disease. J Appl Physiol 76271-277, 1994

758

HANSELL

28. Hartman TE, Primack SL, Kang EY, et al: Disease progression in usual intersitial pneumonia compared with desquamative interstitial pneumonia: Assessment with serial CT Disease progression in usual interstitial pneumonia compared with desquamative interstitial pneumonia. Assessment with serial CT. Chest 110378-382, 1996 29. Hartman TE, Primack SL, Swensen SJ, et al: Desquamative interstitial pneumonia: Thin-section CT findings in 22 patients. Radiology 187787-790, 1993 30. Hartman TE, Swensen SJ, Muller NL, et al: Nonspecific interstitial pneumonitis variable appearance on high-resolution CT scans of the chest. Radiology 209:178, 1998 31. Howard P: Clinical usefulness of bronchoalveolar lavage. Eur Respir J 3:377-378, 1990 32. Ikezoe J, Johkoh T, Kohno N, et al: High-resolution CT findings of lung disease in patients with polymyositis and dermatomyositis. J Thorac Imag 11:250-259, 1996 33. Johkoh T, Ikezoe J, Kohno N, et al: High-resolution CT findings of lung disease in patients with polymyositis and dermatomyositis. J Thorac Imag 11~250-259,1996 33a. Johkoh T, Ikezoe J, Kohno N, et al: High-resolution CT and pulmonary function tests in collagen vascular disease: Comparison with idiopathic pulmonary fibrosis. Eur J Radiol 18:113-121, 1994 34. Johnston IDA, Prescott RJ, Chalmers JC, et al: British Thoracic Society study of cryptogenic fibrosing alveolitis: Current presentation and initial management. Fibrosing Alveolitis Subcommittee of the Research Committee of the British Thoracic Society. Thorax 52:38-44, 1997 35. Johnston IDA, Gomm SA, Kalra S, et al: The management of cryptogenic fibrosing alveolitis in three regions of the United Kingdom. Eur Respir J 6891893, 1993 36. Katzenstein AL, Askin FB: Idiopathic interstitial pneumonia / idiopathic pulmonary fibrosis. In Bennington JL (ed): Surgical Pathology of Non-neoplastic Lung Disease (ed 2). Philadelphia, WB Saunders, 1990, pp 58-96 37. Katzenstein AL, Myers JL: Idiopathic pulmonary fibrosis: Clinical relevance of pathologic classification. Am J Respir Crit Care Med 1571301-1315,1998 38. Katzenstein ALA, Fiorelli RF: Nonspecific interstitial pneumonia / fibrosis: Histologic features and clinical significance. Am J Surg Pathol 18:136-147, 1994 39. Katzenstein ALA, Myers JL, Mazur MT Acute interstitial pneumonia: A clinicopathologic, ultrastructural, and cell kinetic study. Am J Surg Pathol 10:256-267, 1986 40. Kazerooni EA, Martinez FJ, Flint A, et al: Thinsection CT obtained at 10-mm increments versus limited three-level thin-section CT for idiopathic pulmonary fibrosis: Correlation with pathologic scoring. AJR Am J Roentgenol 169:977-983, 1997 41. Lang MR, Fiaux GW, Gillooly M, et al: Collagen content of alveolar wall tissue in emphysematous and non-emphysematous lungs. Thorax 49319-326, 1994 42. Lee JS, Gong G, Song KS, et al: Usual interstitial pneumonia: Relationship between disease activity and the progression of honeycombing at thin-section computed tomography. J Thorac Imag 13:199203, 1998 43. Leung AN, Miller RR, Muller NL: Parenchymal opacification in chronic infiltrative lung diseases:

CT-pathologic correlation. Radiology 188:209-214, 1993 44. Liebow AA: Definition and classification of interstitial pneumonias in human pathology. Progress in Respiratory Research 81-33, 1975 45. Liebow AA, Steer A, Billingsley JG: Desquamative interstitial pneumonitis. Am J Med 393369404, 1965 45a. Lung and Asthma Agency: Factsheet. Factsheet Jan:l, 1997 46. Lynch DA: Ground glass attenuation on CT in patients with idiopathic pulmonary fibrosis. Chest 110:312-313, 1996 47. Lynch DA, Newel1 JD, Logan I'M, et al: Can CT distinguish hypersensitivity pneumonitis from idiopathic pulmonary fibrosis? AJR Am J Roentgenol 4:807-811, 1995 48. Martinez FJ, Karlinsky JB, Gale ME, et al: Intrathoracic lymphadenopathy. A rare manifestation of rheumatoid pulmonary disease. Chest 97:lOlO1012, 1990 49. Mathieson JR, Mayo JR, Staples CA, et al: Chronic diffuse infiltrative lung disease: Comparison of diagnostic accuracy of CT and chest radiography. Radiology 171:111-116, 1989 50. Mayo J R High resolution computed tomography: Technical aspects. Radiol Clin North Am 29:10431049, 1991 51. Mayo JR, Webb WR, Gould R, et al: High-resolution CT of the lungs: An optimal approach. Radiology 163:507-510, 1987 52. McAdams HP, Rosado-de-Christenson ML, Wehunt WD, et al: The alphabet soup revisited: The chronic intersitial pneumonias in the 1990s. Radiographics 161009-1033, 1996 53. McDonagh J, Greaves M, Wright AR, et al: High resolution computed tomography of the lungs in patients with rheumatoid arthritis and interstitial lung disease. Br J Rheumatol33:118-122, 1994 54. Mino M, Noma S, Kobashi Y, et al: Serial changes of cystic air spaces in fibrosing alveolitis: A CTpathological study. Clin Radiol 50:357-363, 1995 55. Mino M, Noma S, Taguchi Y, et al: Pulmonary involvement in polymyositis and dermatomyositis: Sequential evaluation with CT. AJR Am J Roentgenol169:83-87, 1997 56. Muller NL, Colby TV Idiopathic interstitial pneumonias: High-resolution CT and histologic findings. Radiographics 171016-1022, 1997 57. Muller NL, Miller RR, Webb WR, et al: Fibrosing alveolitis: CT pathologic correlation. Radiology 160~585-588,1986 58. Miiller NL, Staples CA, Miller RR, et al: Disease activity in idiopathic pulmonary fibrosis: CT and pathologic correlation. Radiology 165:731-734, 1987 59. Murata K, Khan A, Rojas KA, et al: Optimization of computed tomography technique to demonstrate the fine structure of the lung. Invest Radiol 23:170175, 1988 60. Naidich DP, Zerhouni EA, Hutchins GM, et al: Computed tomography of the pulmonary parenchyma. Part 1: Distal air-space disease. J Thorac Imag 1:3953, 1985 61. Niimi H, Kang EY, Kwong JS, et al: CT of chronic infiltrative lung disease: Prevalence of mediastinal lymphadenopathy. J Comput Assist Tomogr 20:305308, 1996 62. Nishimura K, Izumi T, Kitaichi M, et al: The diagnostic accuracy of high-resolution computed tomography in diffuse infiltrative lung diseases. Chest 104:1149-1155, 1993

HRCT IN THE EVALUATION OF FIBROSING ALVEOLITIS 63. Nishimura K, Kitaichi M, Izumi T, et al: Usual interstitial pneumonia: Histologic correlation with highresolution CT. Radiology 182337-342, 1992 64. Padley SPG, Hansell DM, Flower CDR, et al: Comparative accuracy of high resolution computed tomography and chest radiography in the diagnosis of chronic diffuse infiltrative lung disease. Clin Radiol M227-231, 1991 65. Padley SPG, Padhani AR, Nicholson A, et al: Pulmonary sarcoidosis mimicking cryptogenic fibrosing alveolitis on CT. Clin Radiol 51:807-810, 1996 66. Panos RJ, Moretensen RL, Niccoli SA, et al: Clinical deterioration in patients with idiopathic pulmonary fibrosis: Causes and assessment. Am J Med 88:39& 404, 1990 67. Park JS, Lee KS, Kim JS, et al: Nonspecific interstitial pneumonia with fibrosis: Radiographic and CT findings in seven patients. Radiology 195:645448, 1995 68. Patchefsky AS, Israel HL, Hoch WS, et al: Desquamative interstitial pneumonia: Relationship to interstitial fibrosis. Thorax 28:680-693, 1973 69. Primack SL, Hartman TE, Hansell DM, et al: Endstage lung disease: CT findings in 61 patients. Radiology 189:661-686, 1993 70. Primack SL, Hartman TE, Ikezoe J, et al: Acute interstitial pneumonia: Radiographic and CT findings in nine patients. Radiology 188:817-820, 1993 71. Raghu G: Interstitial lung disease: A diagnostic approach. Are CT and lung biopsy indicated in every patient? Am J Respir Crit Care Med 151:909-914, 1995 72. Remy-Jardin M, Giraud F, Remy J, et al: Importance of ground-glass attenuation in chronic diffuse infiltrative lung disease: Pathologic-CT correlation. Radiology 189:693-698, 1993 73. Remy-Jardin M, Remy J, Cortet B, et al: Lung changes in rheumatoid arthritis: CT findings. Radiology 193~375382,1994 74. Reynolds Hy: Diagnostic and management strategies for diffuse interstitial lung disease. Chest 113:192-202, 1998 75. Rodriguez LH, Vargas PF, Raff U, et al: Automated discrimination and quantification of idiopathic pulmonary fibrosis from normal lung parenchyma using generalized fractal dimensions in high-resolution computed tomography images. Acad Radiol 210-18, 1995 76. Sampson C, Hansell DM: The prevalence of enlarged mediastinal lymph nodes in asbestos-exposed individuals: A CT study. Clin Radiol 45:340342, 1992 77. Scadding JG: Chronic diffuse interstitial fibrosis of the lungs. BMJ 1:443-450, 1960 78. Scadding JG, Hinson KF: Diffuse fibrosing alveolitis (diffuse interstitial fibrosis of the lungs): Correlation of histology at biopsy with prognosis. Thorax 22291-304, 1967 79. Snider GL, Kleinerman J, Thurlbeck WM, et al: The definition of emphysema. Report of a National Heart, Lung, and Blood Institute, Division of Lung Diseases workshop. Am Rev Respir Dis 132:182185, 1985 80. Strickland NH, Hughes JM, Hart DA, et al: Cause of regional ventilation-perfusion mismatching in patients with idiopathic pulmonary fibrosis: A combined CT and scintigraphic study. AJR Am J Roentgenol161:719-725,1993 81. Tazelaar HD, Viggiano RW, Pickersgill J, et a1 Inter-

759

stitial lung disease in polymyositis and dermatomyositis: Clinical features and prognosis as correlated with histologic findings. Am Rev Respir Dis 141:727-733, 1990 82. Terriff BA, Kwan SY, Chan-Yeung MM, et al: Fibrosing alveolitis: Chest radiography and CT as predictors of clinical and functional impairment at fol1992 low-up in 26 patients. Radiology 18444-9, 83. Tonelli M, Stem EJ, Glenny RW HRCT evident fibrosis in isolated pulmonary emphysema. J Comput Assist Tomogr 21:322-323, 1997 84. Tubbs RR, Benjamin SP, Reich NE, et al: Desquamative interstitial pneumonitis: Cellular phase of fibrosing alveolitis. Chest 72159-165, 1977 85. Tung KT, Wells AU, Rubens MB, et al: Accuracy of the typical computed tomographic appearances of fibrosing alveolitis. Thorax 483336338,1993 86. Turner-Wanuick M, Burrows B, Johnson A: Cryptogenic fibrosing alveolitis: Clinical features and their influence on survival. Thorax 35:171-180, 1980 87. van der Bruggen-Bogaarts BA, Broerse JJ, Lammers JW, et al: Radiation exposure in standard and highresolution chest CT scans. Chest 107113-115, 1995 88. Volpe J, Storto ML, Lee K, et al: High-resolution CT of the lung: Determination of the usefulness of CT scans obtained with the patient prone based on plain radiographic findings. AJR Am J Roentgenol 169~369-374,1997 89. Watters LC, King TE, Schwarz MI, et al: A clinical radiographic and physiologic scoring system for the longitudinal assessment of patients with idiopathic pulmonary fibrosis. Am Rev Respir Dis 133:97-103, 1986 90. Wechsler RJ, Steiner RM, Spim PW, et al: The relationship of thoracic lymphadenopathy to pulmonary interstitial disease in diffuse and limited systemic sclerosis: CT findings. AJR Am J Roentgenol 167101-104, 1996 91. Wells AU, Cullinan P, Hansell DM, et al: Fibrosing alveolitis associated with systemic sclerosis has a better prognosis than lone cryptogenic fibrosing alveolitis. Am J Respir Crit Care Med 149:158>1590, 1994 92. Wells AU, Hansell DM, Corrin B, et al: High resolution computed tomography assessment of disease activity in the fibrosing alveolitis of systemic sclerosis: A histopathological correlation. Thorax 47738742, 1992 93. Wells AU, Hansell DM, Harrison NK, et al: Clearance of inhaled 99mTcDTPA predicts the clinical course of fibrosing alveolitis. Eur Respir J 6797802, 1993 94. Wells AU, Hansell DM, Haslam PL, et al: Bronchoalveolar lavage cellularity: Lone cryptogenic fibrosing alveolitis compared with the fibrosing alveolitis of systemic sclerosis. Am J Respir Crit Care Med 1571474-1482, 1998 95. Wells AU, Hansell DM, Rubens MB, et al: The clinical utility of thin section computed tomography in the detection of changes in disease severity in fibrosing alveolitis. Am Rev Respir Dis 147593,1993 96. Wells AU, Hansell DM, Rubens MB, et al: Functional impairment in lone cryptogenic fibrosing alveolitis and fibrosing alveolitis associated with systemic sclerosis: A comparison. Am J Respir Crit Care Med 155:1657-1664, 1997 97. Wells AU, Hansell DM, Rubens MB, et al: The predictive value of thin-section computed tomography in fibrosing alveolitis. Am Rev Respir Dis 148:1076 1082, 1993

760

HANSELL

98. Wells AU, King AD, Rubens MB, et al: Lone cryptogenic fibrosing alveolitis: A functional-morphologic correlation based on extent of disease on thin-section computed tomography. Am J Respir Crit Care Med 155:1367-1375, 1997 99. Wells AU, Rubens MB, Du Bois RM, et al: Serial CT in fibrosing alveolitis: Prognostic significance of the initial pattern. AJR Am J Roentgen01 161:11591165, 1993 100. Wells AU, Rubens MB, Du Bois RM, et al: Functional impairment in fibrosing alveolitis: Relationship to reversible disease on thin section computed tomography. Eur Respir J 10:280-285, 1997 101. Wiggins J, Strickland B, Turner-Warwick M: Combined cryptogenic fibrosing alveolitis and emphysema: The value of high resolution computed to-

102.

103.

104.

105.

mography in assessment. Respir Med 84:365-369, 1990 Winterbauer RH, Hammar SP, Hallman KO, et al: Diffuse interstitial pneumonitis: Clinicopathologic correlations in 20 patients treated with prednisone/ azathioprine. Am J Med 65:661-672, 1978 Wollmer P, Jakobsson K, Albin M, et al: Measurement of lung density by X-ray computed tomography: Relation to lung mechanics in workers exposed to asbestos cement. Chest 912365-869, 1987 Zerhouni EA, Naidich DP, Stitik FP, et al: Computed tomography of the pulmonary parenchyma. Part 2: Interstitial disease. J Thorac Imag 1:54-64, 1985 Zompatori M, Fasano L, Battista G, et al: Course of idiopathic pulmonary fibrosis of the Wells grade 111 at presentation. Study using high-resolution computerized tomography. Radio1 Med 94:611-617, 1997

Address reprint requests to: David M. Hansell, MD, FRCP, FRCR Department of Radiology Royal Brompton Hospital Sydney Street London SW 6NP England e-mail: [email protected]