High-Resolution CT of the Lung
3rd Edition

Because bronchiectasis is defined by the presence of bronchial dilatation, recognition of increased bronchial diameter is key to the CT diagnosis of this abnormality. Various sophisticated methods for measuring airway dimensions have been proposed. These include the use of digital image analysis programs, requiring operator-dependent definition of a “seed point” at the lumen-wall interface to obtain isocontour lines of the bronchial lumen [14], and automated thresholding to detect airway lumen area [54]. Whereas these approaches may allow precise quantitative assessment of airways and may prove particularly valuable in physiologic
studies, subjective visual criteria for establishing the presence of bronchial dilatation are most often used in the interpretation of scans [4,38,55,56,57,58].

For the purposes of the interpretation of HRCT, bronchial dilatation may be diagnosed using a comparison of bronchial size to that of adjacent pulmonary artery branches (i.e., the bronchoarterial ratio), by detecting a lack of bronchial tapering, and by identifying airways in the peripheral lung.

In most cases, bronchiectasis is considered to be present when the internal diameter of a bronchus is greater than the diameter of the adjacent pulmonary artery branch–that is, the bronchoarterial ratio is greater than 1 (Figs. 8-2, 8-3 and 8-4) [53]. The accuracy of this finding in diagnosing bronchiectasis has been validated in a number of studies comparing CT to
bronchography in patients who have bronchiectasis [59,60,61,62]. In patients who have bronchiectasis, the bronchial diameter is often much larger than the pulmonary artery diameter (i.e., greater than 1.5 times the artery diameter), a finding that not only reflects the presence of bronchial dilatation but also demonstrates some reduction in pulmonary artery size as a consequence of decreased lung perfusion in affected lung regions [56]. The association of a dilated bronchus with a much smaller adjacent pulmonary artery branch has been termed the signet ring sign (see Figs. 3-135 and 3-137; Figs. 8-2, 8-3 and 8-4). This sign is valuable in recognizing bronchiectasis and in distinguishing it from other cystic lung lesions.

Although an increased bronchoarterial ratio is typical of bronchiectasis, it should be kept in mind that a bronchoarterial ratio of more than 1 need not always indicate the presence of bronchiectasis [4,38,56]. There are a number of instances in which airways may appear dilated in the absence of wall destruction, although this dilatation is usually minimal; this has been reported in asthmatic patients [36,38,63], in patients living at high altitudes [38,56], and in a small percentage of normals [64] (Fig. 8-5). For example, in an HRCT evaluation of 14 normal subjects [65], although the bronchoarterial ratio averaged 0.65 ± 0.16, 7% of interpretations showed some evidence of bronchial dilatation. Kim et al. [56] found that nine of 17 (53%) normal subjects living at an altitude of 1,600 meters had evidence of at least one bronchus equal to or larger in size than the adjacent pulmonary artery. These authors also found that two of 16 (12.5%) individuals living at sea level showed evidence of at least one abnormally dilated airway. Similarly, Lynch et al. [38] compared the internal diameters of lobular, segmental, subsegmental, and smaller bronchi to those of adjacent pulmonary artery branches in 27 normal subjects living in Denver. The authors found that 37 of 142 (26%) bronchi evaluated and 59% of individuals had increased bronchoarterial ratios. Evaluation of the distribution of these abnormal airways has failed to identify any significant difference in the likelihood of abnormal bronchoarterial ratios in airways either by lobe or anteroposterior location within the lungs [56]. A similar lack of variation by segments, lobes, or lungs has been reported by Kim et al. [58].

It must be emphasized that bronchiectasis should not be diagnosed on the basis of an increased bronchoarterial ratio alone unless it is significant; in a study by Lynch et al. [38] of normal subjects and patients who had asthma, although bronchoarterial ratios of greater than 1 were frequently seen, in none of these was the bronchoarterial ratio greater than 1.5 [66]. Also, bronchial wall thickening is almost always seen in association with bronchial dilatation in patients who have bronchiectasis, as are irregularities in bronchial diameter and lack of bronchial tapering. In the normal subjects studied by Lynch et al. [38] who demonstrated an increased bronchoarterial ratio, bronchial wall thickening was relatively uncommon, and it is unlikely that any of the subjects in this study would have been diagnosed on clinical HRCT studies as having true bronchiectasis.

The definition of an abnormal bronchoarterial ratio varies widely among reported series [4,53,55,57,58,65,67], a fact that limits their comparison. In addition to variations in size criteria for considering a bronchus to be dilated (ranging from one to two times the diameter of the adjacent pulmonary artery) and the use of internal or external bronchial diameters for comparison to pulmonary artery diameter, important differences may also be attributable to the use of either visual inspection or objective measurements of bronchial and arterial dimensions. As emphasized by Kim et al. [56], visual inspection alone may lead to an overestimation of bronchoarterial ratios due to a subtle optical illusion, in which the diameter of hollow circles appears larger than solid circles despite their being identical in size. Considerable variation in bronchoarterial ratio may also be seen in normals. In a study by Kim et al. [58], the arterobronchus ratio, defined as the outer diameter of the pulmonary artery divided by the outer diameter of its accompanying bronchus measured at the subsegmental level, averaged 0.98 ± 0.14 but ranged from 0.53 to 1.39.

Despite variability in normal bronchial size and the methods by which bronchial diameter has been assessed, several studies have shown that measurements of bronchial diameter may be reliably made from HRCT [68,69]. Desai et al. [68] evaluated both inter- and intraobserver variation in CT measurements of bronchial wall circumference in 61 subsegmental bronchi and found the reproducibility of these measurements to be sufficiently clinically useful in demonstrating the progression of bronchiectasis. As emphasized by Diederich et al. [57], visual inspection remains the mainstay for assessing bronchial dilatation, as obtaining objective measurements (with or without the use of calipers) is time consuming and often clinically impractical. In this regard, use of visual inspection has been shown to
have acceptable interobserver variability. Using visual inspection only, Diederich et al. [57] found close agreement among three readers both in the detection (κ = 0.78) and assessment of the severity (κ = 0.68) of bronchiectasis.

A potential limitation of the use of bronchoarterial ratios is the necessity of identifying both airways and accompanying arteries. This may not always be possible in patients who have coexisting parenchymal consolidation [4]. Kang et al. was unable to determine the bronchoarterial ratios in three cases in a study of 47 resected lobes with documented bronchiectasis due to the presence of parenchymal consolidation [4].

Lack of bronchial tapering has come to be recognized as an important finding in the diagnosis of bronchiectasis, and subtle cylindrical bronchiectasis in particular (Table 8-2) (Fig. 8-6). It has been suggested that for this finding to be present, the diameter of an airway should remain unchanged for at least 2 cm distal to a branching point [56]. First emphasized by Lynch et al. [38] as a necessary finding for diagnosis, lack of bronchial tapering has been reported by some to be the most sensitive means for diagnosing bronchiectasis. Kang et al. [4], for example, in an assessment of 47 lobes with pathologically proved bronchiectasis, found lack of tapering of bronchial lumina in 37 cases (79%) as compared with increased bronchoarterial ratios seen in only 28 cases (60%). In another study [64], lack of tapering of bronchi was seen in 10% of HRCT interpretations in healthy subjects, compared with 95% of reviews in patients who had bronchiectasis. It should be emphasized that the accurate detection of this finding is difficult in the absence of contiguous HRCT sections, especially for vertically or obliquely oriented airways. The value of this sign is doubtful when HRCT scans are obtained
at spaced intervals in a noncontiguous fashion. As discussed in this book, this has clear implications for optimizing scan technique in patients who have suspected airways disease.

Another frequently cited manifestation of bronchiectasis is the visibility of airways in the lung periphery (Table 8-2) [4,64]. The smallest airways normally visible using HRCT techniques have a diameter of approximately 2 mm and a wall thickness of 0.2 to 0.3 mm [70]; in normal subjects, airways in the peripheral 2 cm of lung are uncommonly seen because their walls are too thin [71]. Peribronchial fibrosis and bronchial wall thickening in patients who have bronchiectasis, in combination with dilatation of the bronchial lumen, allow the visualization of small airways in the lung periphery, and this finding can be very helpful in diagnosing the presence of an airway abnormality. In a study by Kang et al. [4], bronchi visualized within 1 cm of pleura were seen in 21 of 47 (45%) bronchiectatic lobes.

Kim et al. have further assessed the value of this sign in the diagnosis of bronchiectasis [64], pointing out that although normal bronchi are not visible within 1 cm of the costal pleural surfaces, bronchi may be seen within 1 cm of the mediastinal pleural surfaces in normal subjects; visible bronchi were identified within 1 cm of the mediastinal pleura in 40% of normal subjects [64]. These authors found that visible bronchi within 1 cm of the pleural surface or bronchi touching the mediastinal pleural surfaces were visible in 81% and 53%, respectively, of HRCT interpretations in patients who had clinical or pathologic evidence of cylindrical bronchiectasis (Figs. 8-4A, 8-6, and 8-7).

Although bronchial wall thickening is a nonspecific finding, it is usually present in patients who have bronchiectasis (Table 8-2) (Figs. 8-3, 8-4, and 8-7, 8-8 and 8-9). Generally accepted criteria for determining abnormal bronchial wall thickening have yet to be established.

Airways divide by asymmetric dichotomous branching, with approximately 23 generations of branches from the trachea to the alveoli. Anatomically, second- to fourth-generation segmental airways have mean diameters of between 5 and 8 mm, corresponding to a wall thickness of approximately 1.5 mm; sixth- to eighth-generation airways have mean diameters measuring between 1.5 and 3 mm with walls of approximately 0.3 mm; and eleventh- to thirteenth-generation airways have diameters measuring 0.7 to 1 mm, with walls of 0.1 to 0.15 mm [72]. The wall thickness of conducting airways distal to the segmental level is approximately proportional to their diameter. In general, the thickness of the wall of a bronchus or bronchiole smaller than 5 mm in diameter should measure from one-sixth to one-tenth of its diameter [73]; however, precise measurement of the wall thickness of small bronchi or bronchioles is difficult, as wall thickness approximates pixel size [70].

Precise definitions of bronchial wall thickening have been proposed by several investigators, relying on the visual assessment or measurement of the ratio between bronchial wall thickness and the diameter of adjacent pulmonary arteries or nearby normal airways (Fig. 8-9) [57,67,74,75]. Proposed techniques include determining the ratio of airway wall thickness to the outer airway diameter, as well as the percentage wall area defined as (wall area/total airway area) × 100. Using these parameters, Awadh et al. have shown that there is a
clearly definable gradient in bronchial wall thickness between normal subjects and asthmatic patients of varying severity [76]. The relationship between wall thickness and bronchial diameter has also been assessed by Kim et al. [58], using the bronchial lumen ratio (BLR). At the subsegmental level, the BLR, defined as the inner diameter of the bronchus divided by its outer diameter, was measured at the subsegmental level on a display console. Considerable variation in the BLR was found, averaging 0.66 ± 0.06 with a range of 0.51 to 0.86 [58]. It has also been suggested that bronchial wall thickening may be diagnosed if the airway wall is at least 0.5 times the width of an adjacent vertically oriented pulmonary artery (Table 8-2).

Identification of thickened bronchial walls for the purposes of interpretation of clinical scans remains largely subjective (Fig. 8-9) [4,36]. Because bronchiectasis and bronchial wall thickening are often multifocal rather than diffuse and uniform, a comparison of one lung region to another can be helpful in making this diagnosis. It must be emphasized that using consistent window settings is very important in the diagnosis of bronchial wall thickening; bronchial walls can vary significantly in apparent thickness with different CT window settings (see Chapter 1).

Although estimates of airway wall thickening are subjective, it has been shown that visual assessment of wall thickening may be reliable when sequential scans are assessed. Using a visual estimation of wall thickness, Diederich et al. [57] found acceptable levels of agreement among three readers as to the presence or absence of bronchial wall thickening
(κ = 0.64). It should be emphasized, however, that interobserver agreement in itself might be a misleading statistic in assessing the validity of HRCT measurements. Bankier et al. [77], in a study of normal and abnormal airways evaluated by three independent observers before and after a training period, showed that although interobserver variability improved significantly on second readings, training had no effect on sensitivity. Sensitivity in detecting abnormal bronchi was 46% before and 44% after training, and specificity measured 71% and 72%, respectively [77]. Although these results are disappointing, it should be emphasized that the airway abnormalities evaluated in this study were extremely subtle, with abnormal segmental and subsegmental wall thickness measuring 1.77 mm and 0.95 mm, respectively, as compared with normal segmental and subsegmental airways measuring 1.14 mm and 0.46 mm. Furthermore, individual airways were assessed individually on targeted reconstructed images only. Nonetheless, a statistically significant difference (p = .001) between measured wall thickness of normal and pathologic bronchi was found [77].

The presence of mucus- or fluid-filled bronchi may be helpful in confirming a diagnosis of bronchiectasis (Table 8-2). The HRCT appearance of fluid- (Fig. 8-8) or mucus-filled airways is dependent both on their size and orientation relative to the scan plane. Larger mucus-filled airways result in abnormal lobular or branching structures when they lie in the same plane as the CT scan (Figs. 8-10, 8-11, 8-12, 8-13, 8-14, and 8-15). Although these may be confused with abnormally dilated vessels, in most cases, the recognition of dilated, fluid-filled airways is simplified by the identification of other areas of bronchiectasis in which the bronchi are air-filled; these are usually present if carefully sought (Figs. 8-10A and 8-11). In problematic cases, a distinction between larger fluid-filled bronchi and dilated blood vessels is easily made by rescanning patients after the bolus injection of intravenous contrast medium (Fig. 8-13). Alternatively, with the introduction of newer-generation scanners, it is now possible to obtain high quality multiplanar and maximum-intensity projection images in a variety of imaging planes as a means for further evaluation (Figs. 8-14 and 8-15).

Although commonly associated with bronchiectasis and infection, dilated mucus-filled airways in the central lung can also result from congenital bronchial abnormalities, such as bronchopulmonary sequestration or bronchial atresia (Fig. 8-15) [78,79,80,81,82,83].

It should also be emphasized that the finding of dilated mucus-filled airways, especially when central or predominantly segmental or lobular in distribution, should alert one to the possibility of central endobronchial obstruction, resulting either from tumor or foreign body aspiration. As previously discussed, the use of intravenous contrast media in this setting may allow differentiation between central tumor and fluid-filled peripheral airways.

As is discussed in greater detail in the section on Bronchiolitis, mucus- or pus-filled small airways in the lung periphery are usually identifiable either as branching structures within the center of secondary lobules, aptly described as having a tree-in-bud (TIB) appearance [27,84] or as ill-defined centrilobular nodules (Table 8-2) [70,85,84,85,86,87,88] (Figs. 8-4, 8-11, 8-16, and 8-17). These are frequently identified in association with bronchiectasis and usually indicate infection. For example, in patients who have diffuse panbronchiolitis (DPB) [89], bronchiolectasis results in small, ring-shaped, round, or branching opacities in the peripheral lung. These opacities correspond to abnormalities of distal airways, including terminal and respiratory bronchioles [86,87,88].

Over the past several years it has become increasingly apparent that in most cases, patients who have bronchiectasis also show evidence of small airway pathology. Kang et al., for example, in their study of 47 resected lobes with documented bronchiectasis, found pathologic evidence of bronchiolitis in 85% [4]. These included six lobes with obliterative bronchiolitis, 18 with inflammatory or suppurative bronchiolitis, and 16 with both obliterative and inflammatory bronchiolitis. HRCT findings consistent with bronchiolitis were identified in 30 of 47 (75%) lobes, including a pattern of mosaic perfusion (n = 21) (Fig. 8-18); bronchiolectasis (n = 17); and centrilobular nodular or branching opacities, or both (i.e., TIB) (n = 10) (Figs. 8-4 and 8-16) [4].

In particular, the findings of mosaic perfusion on inspiratory scans and focal air-trapping on expiratory scans may prove of special interest in the early diagnosis of bronchiolitis associated with large airways disease (Table 8-2) (Figs. 8-18 and 8-19). In one study of 70 patients [90] who had HRCT evidence of bronchiectasis visible in 52% of lobes evaluated, areas of decreased attenuation (i.e., mosaic perfusion) were visible on inspiratory scans in 20% of lobes, and on expiration (air-trapping) in 34%. Although areas of decreased attenuation on expiratory scans were more prevalent in lobes with severe (59%) or localized (28%) bronchiectasis, in 17% of lobes, air-trapping was identified in the absence of associated bronchiectasis. This has led to speculation that evidence of bronchiolar disease may in fact precede and even lead to the development of bronchiectasis [7,90,91].

In this same study [90], the presence of decreased attenuation on expiratory scans was also associated with mucoid impaction. This finding was seen in 73% of lobes with large mucous plugs and in 58% of those with centrilobular mucous plugs. These same authors noted a correlation (r = 0.40; p <.001) between the total extent and severity of bronchiectasis and the extent of decreased attenuation shown on expiratory CT. Not surprisingly, in 55 patients who had pulmonary function tests, the extent of expiratory attenuation abnormalities proved inversely related to measures of airway obstruction, such as forced expiratory volume in one second (FEV₁) and FEV₁/forced vital capacity (FVC) [90].

Normal bronchial arteries extend along the central airways to a level a few generations proximal to the terminal bronchioles and are the main blood supply for the bronchi. Arising directly from the proximal descending thoracic aorta, these typically measure less than 2 mm and, in addition to supplying the central airways, supply blood to the esophagus and mediastinal lymph nodes. Enlarged bronchial arteries can be identified pathologically in most cases of bronchiectasis and, when severe, usually account for the occurrence of hemoptysis in these patients. With HRCT, it has proved possible to identify both normal and abnormal
bronchial arteries in select cases, especially after a bolus of intravenous contrast [92,93].

Song et al. were able to demonstrate good correlation between non-contrast enhanced HRCT images and corresponding CT angiograms for demonstrating hypertrophied bronchial arteries in patients who had bronchiectasis [94]. Specifically, these authors showed that the finding of tubular or nodular areas of soft-tissue attenuation, distinct from blood vessels within the mediastinum and adjacent to the central airways, correctly predicted bronchial artery hypertrophy in 88% and 53% of cases, respectively. Although the diagnosis of bronchiectasis rarely is dependent on demonstrating bronchial artery hypertrophy, identification of focal bronchial wall abnormalities due to enlarged bronchial arteries is important before attempted bronchoscopy, as inadvertent biopsy may lead to significant hemorrhage [94].

Mild or cylindrical bronchiectasis is diagnosed if the dilated bronchi are of relatively uniform caliber and have roughly parallel walls (Fig. 8-2). The appearance of cylindrical bronchiectasis varies depending on whether the abnormal bronchi have a horizontal or vertical course relative to the scan plane. When horizontal, bronchi are visualized along their length and are recognizable as branching tram tracks that fail to taper as they extend peripherally and are visible more peripherally than is normal. When cylindrically dilated bronchi are oriented in a vertical direction, they are scanned in cross section and appear as thick-walled, circular lucencies.

In most cases, dilated bronchi seen in cross section can easily be distinguished from emphysematous blebs or other causes of lung cysts by identifying the signet ring sign and the continuity of the dilated bronchus on adjacent scans.

With increasingly severe abnormalities of the bronchial wall, bronchi may assume a beaded configuration referred to as varicose bronchiectasis. This diagnosis can be made consistently only when the involved bronchi course horizontally in the plane of scan (Figs. 8-3, 8-11, and 8-16). Varicose bronchiectasis is much less frequent than cylindrical bronchiectasis.

With severe or cystic bronchiectasis, involved airways are cystic or saccular in appearance and may extend to the pleural surface (Figs. 8-4 and 8-6, 8-7 and 8-8). On HRCT, cystic bronchiectasis may be associated with the presence of (i) air-fluid

levels caused by retained secretions in the dependent portions of the dilated bronchi (Fig. 8-8), (ii) a string of cysts, caused by sectioning irregularly dilated bronchi along their length, or (iii) a cluster of cysts (Figs. 8-7 and 8-20), caused by multiple dilated bronchi lying adjacent to each other. Clusters of cysts are most frequently seen in atelectatic lobes (Fig. 8-20), presumably as a result of chronic infection such as commonly occurs in patients who have pulmonary tuberculosis. In general, the dilated airways in patients who have cystic bronchiectasis are thick-walled; however, cystic bronchiectasis may also appear thin-walled (Fig. 8-6). Recognition of some combination of dilated bronchi, air-fluid levels, and strings or clusters of cysts should be diagnostic of cystic bronchiectasis [97].

A number of technical factors need to be considered in assessing airway pathology. As discussed in Chapter 1, these include slice thickness, slice spacing, field of view (FOV), and reconstruction algorithm. Using an FOV of 13 cm × 13 cm provides the maximum spatial resolution, with resulting pixels measuring approximately 0.25 mm × 0.25 mm [104]. If 1-mm collimation is used, voxel size is 0.25 mm × 0.25 mm × 1.00 mm, equal to 0.06 mm³ [105]. Although a small FOV is rarely used in routine clinical imaging, the ability to obtain target-reconstructed images through select areas may enhance visualization of fine parenchymal detail and be of value in select cases with airways disease.

Most important for accurate airway evaluation is the use of appropriate window levels and windows, especially in those cases for which quantitative information is sought (see Fig. 1-13). As first shown by Webb and coworkers using phantoms composed of lucite cylinders, an optimal window level for assessing the airway lumen and walls is -450 Hounsfield units (HU) [106]. A similar conclusion has been reached by McNamara et al., also by using a reference phantom [105,107]. In distinction, others have suggested that window
width is as important as or more important than window level in airway measurements. Bankier et al. [108], using inflation-fixed lungs to evaluate the effect of window width and levels on bronchial wall thickness confirmed by planimetry, concluded that an optimal window width should vary between 1,000 and 1,400 HU, whereas window levels could vary as much as between -250 and -700 HU. In our experience, for practical purposes, these windows and levels are adequate for routine visual assessment (see Fig. 1-13).

Following the report of Grenier et al., the use of HRCT sections acquired every 10 mm in deep inspiration has become standard for diagnosing bronchiectasis [109]. Using this protocol, Grenier et al. [109] have found a very high accuracy for HRCT in diagnosing bronchiectasis as compared to bronchography. Subsequent reports have further confirmed the value of these techniques for establishing the diagnosis of bronchiectasis [62,110].

Based on the results of these studies and several others, the following technique is recommended for patients who have suspected bronchiectasis (see Chapter 1). In patients in whom there are no specific clinical or radiographic signs to help localize disease, 1- to 1.5-mm high-resolution images should be obtained every 10 mm from the lung apices to bases. Despite the lack of contiguous scanning, this technique allows adequate assessment of the bronchial tree in nearly all cases. Although the routine use of thick sections is not indicated, in select cases, especially those in whom mild cylindrical bronchiectasis is suspected, selected thick sections within a limited range of interest may be of value [38].

This approach can be modified to reflect varying clinical presentations. For example, in patients presenting with hemoptysis, it is usually necessary to rule out occult central endobronchial lesions in addition to detecting bronchiectasis. This is best accomplished by obtaining 1- to 1.5-mm-thick sections every 10 mm through the upper- and lower-lung zones, and contiguous 3- to 5-mm-thick sections from the carina to the level of the inferior pulmonary veins [50]. Using this protocol, in a retrospective study of 59 patients evaluated both by CT and fiberoptic bronchoscopy, CT proved abnormal in all cases in which fiberoptic bronchoscopy depicted focal airway pathology [50]. Alternatively, when available, helical or spiral scanning can be substituted for the routine 5-mm axial images through the central airways. These have
the advantage of eliminating misregistration artifacts as well as allowing high quality three-dimensional or multiplanar reconstructions to be performed.

Additional scan techniques have also been advocated. As suggested by Remy-Jardin et al. [111], visualization of bronchiectatic segments may be enhanced by use of 20-degree cranial angulation of the CT gantry (see Fig. 1-19). Although unnecessary in most patients, this technique may be of value in equivocal cases, especially for those airways that normally course obliquely, such as the middle lobe and lingular bronchi.

The use of low-dose HRCT techniques for performing routine follow-up scans in patients who have severe chronic disease has also been suggested [74,112]. Bhalla et al. [74], evaluating scans obtained using both 70 and 20 mA, showed that high-quality HRCT images of bronchiectatic airways could be obtained in patients who had CF (Fig. 8-21).

The introduction of spiral CT scanning has led to a reconsideration of optimal scan protocols in patients who have airways disease. In a study by van der Bruggen-Bogaarts et al. [113], spiral CT was found to have a high sensitivity (91%) and specificity (99%) in the diagnosis of bronchiectasis detected using HRCT. In this study, 30 patients and 177 lobes were evaluated with spiral CT (slice thickness, 5 mm; pitch, 1; reconstructed at 2 mm) and HRCT (1.5 mm, 10 mm scan spacing). At HRCT, 14 patients showed signs of bronchiectasis in 32 lobes: Spiral CT confirmed the presence in 29 lobes. In one lobe, spiral CT was falsely positive. In a study by Lucidarme et al. [6] of 50 consecutive patients who had suspected bronchiectasis, 1.5-mm HRCT sections obtained at 10-mm intervals were compared to a single 24-second volumetric acquisition using 3-mm collimation. These authors found volumetric data acquisition to be more accurate than routine HRCT for the identification of bronchiectasis. Although helical CT failed to identify bronchiectasis in seven segments in which the disease was diagnosed with HRCT, in four cases a diagnosis of subsegmental cylindrical bronchiectasis was made only with helical CT. In comparison, there were no patients in whom a diagnosis of bronchiectasis was established by HRCT alone.

Spiral CT shows greatest promise in the diagnosis of subtle cylindrical bronchiectasis. Evaluation of the finding of lack of bronchial tapering is difficult using noncontiguous 1-mm sections, effectively reducing the value of this sign to those bronchi that course within the plane of the CT scan. As documented by Giron et al. [110], in a study of 54 patients who had bronchographic evidence of bronchiectasis, obtaining 1-mm sections every 10 mm, three cases were missed and all had mild cylindrical bronchiectasis.

The introduction of multidetector-row spiral CT scanners promises to greatly improve our ability to diagnose airway pathology (see Fig. 1-23; Figs. 8-14, 8-15, and 8-22). The ability to scan the entire thorax in a single breath hold, using 1- to 1.25-mm detector width, and enabling sections of varying thickness to be prospectively and retrospectively reconstructed at any level, is of great value in the diagnosis of airways disease. As indicated in Chapter 1, this technique may fundamentally change the manner in which HRCT is obtained, at least in some patients. The use of reconstruction techniques such as maximum- and minimum-intensity projection images and precise internal and external volumetric renderings of the airways (see Fig. 1-23; Figs. 8-14, 8-15, 8-19, and 8-22) is of potential value. Although many of these techniques are best suited to evaluating small airways disease, potential applications also extend to the larger airways. Of particular interest is the potential to generate CT bronchograms (Fig. 8-22) [8,114].

Given these options, it is apparent that choice of scan technique depends on the type of scanner available, as well as available postprocessing capabilities. It is also likely that specific protocols will continue to evolve with continued clinical experience. Our current recommendation for single-detector scanners calls for an initial series of 1-mm routine axial sections acquired every 10 mm. This may be supplemented with volumetric acquisition of 3-mm sections through the central airways with a pitch of 1.6 to 2, reconstructed every 2 mm [6,115]. Based on our initial experience, a range of potential scan techniques may be applicable to multidetector-row scanners. Using either 1-mm or 1.25-mm detector widths, we reconstruct 3- to 5-mm sections every 2 to 4 mm, as well as contiguous
1- to 1.25-mm sections or 1- to 1.25-mm sections every 10 mm.

HRCT may also be used to evaluate the presence of air-trapping in patients who have suspected bronchiectasis (see Figs. 3-148, 3-149, 3-150, 3-151, 3-152 and 3-153; Fig. 8-18) [11,17,90,91,116,117]. This may be accomplished by repeatedly acquiring scans at one preselected level during a forced expiration or as two separate acquisitions, first in deep inspiration, followed by scans obtained through the same region in expiration (see Chapters 1 and 2) [9,10,14,90,115,118,119]. This technique of paired inspiratory and expiratory images may be obtained using HRCT or spiral volumetric techniques. It should be emphasized that although a variety of strategies for acquiring expiratory scans have been proposed, including 1-mm images obtained at three levels (aortic arch, tracheal carina, and above the diaphragm) are usually sufficient to identify significant air-trapping, even when inspiratory scans are normal [117].

Regardless of the expiratory scan technique used, the resulting images allow the identification of focal areas of air-trapping, as well as changes in the appearance of the airways themselves. In a study of 100 patients who had bronchiectasis having both inspiratory and expiratory scans [91], the extent of decreased attenuation (i.e., air-trapping) on the
expiratory CT scan correlated strongly with the severity of airflow obstruction; the closest relationship (r = -0.55; p = .00005) was seen between decreased FEV₁.

It is worth noting that in select cases, minimum-intensity projection images may be more sensitive than routine images for detecting subtle regions of air-trapping on expiratory scans (Fig. 8-19) [120,121,122].

Several potential pitfalls in the diagnosis of bronchiectasis should be avoided (Table 8-4) [1]. Of particular concern are those due to respiratory and cardiac motion artifacts (Fig. 8-23). Transmitted cardiac motion artifacts frequently obscure detail in the left lower lobe and may lead to an erroneous diagnosis of subtle bronchiectasis [123]. Respiratory motion artifacts also cause ghosting that can very closely mimic the appearance of tram tracks. As previously discussed in the section Bronchial Wall Thickening, and in Chapter 1, the appearance of bronchial wall thickening is dependent on the use of appropriate window widths and levels [108]. Furthermore, on expiratory scans, bronchi can appear thicker-walled and narrower than on inspiratory scans.

Bronchiectasis is especially difficult to diagnose in patients who have concurrent parenchymal consolidation or atelectasis, as CT often discloses dilated peripheral airways that will revert to normal after resolution of the lung disease, so-called reversible bronchiectasis (Fig. 8-24) [124]. In such cases, follow-up scans are recommended, pending radiographic resolution. Another potential pitfall related to lung consolidation is the fact that consolidation may obscure vascular anatomy, rendering interpretation of bronchoarterial ratios difficult or impossible [4]. Of course, as emphasized throughout this textbook, visualization of small structures within the lung requires a high-resolution technique. This is especially true when assessing smaller airways (Fig. 8-25).

A number of cystic lung diseases may also be difficult to differentiate from bronchiectasis (Figs. 8-26 and 8-27). Included in this grouping are cavitary nodules in patients who have either widespread bronchoalveolar cell carcinoma or cavitary metastases. Rarely, cystic lesions in patients who have Pneumocystis carinii pneumonia (PCP) may superficially simulate bronchiectasis: In these cases, it should be recognized that cystic changes usually occur within areas of ground-glass opacity, simplifying differential diagnosis. In patients who have Langerhans cell histiocytosis, bizarre-shaped cysts are often seen, especially in the upper lobes. As these may appear to branch, their appearance may be suggestive of bronchiectasis, so-called pseudobronchiectasis (Figs. 8-26A and 8-27). In fact, pathologically, some of these cystic abnormalities indeed represent abnormally dilatated bronchi, presumably the result of peribronchiolar inflammation.

Bronchiectasis may occur as a component of fibrotic lung diseases or may be seen after radiation therapy. Regardless of the underlying etiology, the result is so-called traction bronchiectasis (Figs. 8-28 and 8-29). This is easily identified, as peripheral bronchi appear irregularly thick-walled or corkscrewed and are invariably found in association with either diffuse reticular changes or honeycombing [125]. Traction bronchiectasis does not represent primary airways disease and is unassociated with symptoms [125].

HRCT findings in patients who have CF have been well described (Figs. 8-31, 8-32, 8-33, 8-34 and 8-35) [74,99,100,102,103,135,140,141,142]. Bronchiectasis is present in all patients who have advanced CF who are studied using HRCT (Table 8-5) [74,99,100,102,103,135,140,141]. Proximal or perihilar bronchi are always involved when bronchiectasis is present, and bronchiectasis is limited to these central bronchi in approximately one-third of the cases, a finding that is referred to as central bronchiectasis (Figs. 8-34 and 8-35). Both the central and peripheral bronchi are abnormal in approximately two-thirds of patients [74,140]. All lobes are typically involved, although early in the disease abnormalities are often predominantly upper lobe in distribution, and a right upper lobe predominance may be present in some patients (Fig. 8-32) [74,99,100,102,103,135,140,141,142].

Cylindrical bronchiectasis is the most frequent pattern seen; it was visible in 94% of lobes in one study of patients who had severe disease [140]. Thirty-four percent of lobes in this study showed cystic bronchiectasis, whereas varicose bronchiectasis was seen in 11%. In another report, cystic lesions representing cystic bronchiectasis or abscess cavities were present in eight of 14 (57%) patients (Fig. 8-21) [74].

Bronchial wall thickening, peribronchial interstitial thickening, or both are also commonly present in patients who have CF (Figs. 8-31, 8-32, 8-33, 8-34 and 8-35) [99,102,103,143]. The thickening is generally more evident than bronchial dilatation in patients who have early disease and may be seen independent of bronchiectasis [74,142]. Thickening of the wall of proximal right upper lobe bronchi was the earliest abnormal feature visible on HRCT in one study of patients who had mild CF [74,142].

Mucous plugging is also common, reported in between one-quarter and one-half of cases [99,102,103], and may be visible in all lobes [74,141]. Collapse or consolidation can be seen in as many as 80% of cases (Figs. 8-31 and 8-33) [74,99,143]. Volume loss was visible in 20% of lobes in patients who had advanced disease [140].

Branching or nodular centrilobular opacities (i.e., tree-in-bud) which reflect the presence of bronchiolar dilatation with associated mucous impaction, infection, or peribronchiolar inflammation, can be an early sign of disease (Figs. 8-31 and 8-32) [141]. Focal areas of decreased lung opacity, representing air-trapping or mosaic perfusion, are common (Figs. 8-33, 8-34, 8-35). These can be seen to correspond to pulmonary lobules or subsegments and may appear to surround dilated, thick-walled, or mucous-plugged bronchi [141] in as many as two-thirds of patients [102]. Air-trapping can often be seen on expiratory scans (Fig. 8-35) [140].

Lung volumes may appear increased on CT, although this diagnosis is rather subjective and may be better assessed on chest radiographs [140]. Cystic or bullous lung lesions can also be visible and typically predominate in the subpleural regions of the upper lobes [74,140]. Hilar or mediastinal lymph node enlargement and pleural abnormalities can also be seen, largely reflecting chronic infection. Pulmonary artery dilatation resulting from pulmonary hypertension can also be seen in patients who have long-standing disease.

HRCT can demonstrate morphologic abnormalities in patients who have early CF who are asymptomatic, have normal pulmonary function, have normal chest radiographs, or a combination of these. In a study of 38 patients who had mild CF with normal pulmonary function [142], chest radiographs were normal in 17 (45%), showed mild bronchial wall thickening in 17, and showed mild bronchiectasis in four (10%). On HRCT in this group, features of bronchiectasis were present in 77% of all patients and in 65% of those with normal chest radiographs; only three patients had a normal HRCT [142]. In another study of HRCT findings in 12 largely asymptomatic pediatric patients who had early CF, chest radiographs were normal in seven, whereas HRCT was normal in only two. HRCT findings not visible on radiographs included bronchial wall thickening, bronchiectasis, centrilobular small airway abnormalities, and lobular or segmental inhomogeneities representing mosaic perfusion or air-trapping [141].

In patients who have more advanced disease, HRCT can also show abnormalities not visible on chest radiographs. In a study of 14 patients who had CF [74], HRCT was found to be superior to chest radiographs in detecting bronchiectasis and mucous plugging. Of a total of 162 segments assessed, bronchiectasis was detected in 124 segments using HRCT, whereas only 71 segments were considered to show this finding on chest radiographs. Mucous plugs were detected on HRCT in 38 segments, whereas they were seen on radiographs in only four segments. In a study by Hansell et al. [140], bronchiectasis was considered to be present on HRCT in 124 of 126 lobes; on chest radiographs, only 84 of 102 lung zones were considered to show this finding. Chest radiographs also underestimated the extent of bronchiectasis. Bronchiectasis was considered to be both central and peripheral in only 31% of lung zones on chest radiographs, whereas a diffuse distribution was seen in 59% of lobes using HRCT.

Despite numerous reports detailing the range of abnormalities identified in patients who have CF, few if any of these find
ings are specific. Reiff et al., in an assessment of 168 patients who had suspected bronchiectasis from a variety of etiologies, found that patients who had adult CF tended to have more widespread involvement than idiopathic bronchiectasis (p <.01) [67]. In patients who have early disease, abnormalities are often predominantly upper lobe in distribution, with a right upper lobe predominance. Despite these findings, as reported by Lee et al. and Cartier et al., a specific diagnosis of adult CF was made in only 38% and 68% of cases, respectively [101,132].

The routine clinical evaluation of CF makes use of clinical and radiograph-based scoring systems. Several authors have also suggested the use of an HRCT scoring system [74,99,144]; these are described in detail above. It is hypothesized that such a scoring system may facilitate the objective evaluation of existing and newly developed therapeutic regimens [74]. One scoring system [74], based on an assessment of the degree and extent of bronchiectasis, bronchial wall thickening, mucous plugging, atelectasis, emphysema, and other findings, showed a statistically significant correlation to the percent ratios of FEV₁/FVC (r = 0.69; p = .006) [74]. In another study based on assessment of bronchiectasis and mucous plugging [144], CT scores correlated highly with clinical (r = 0.88; p <.0001) and radiographic (r = 0.93; p <.0001) scores and several pulmonary function tests. The best correlation was with bronchiectasis.

Several reports have shown that CT offers a reliable alternative to routine radiographic and clinical methods for monitoring disease status and progression, as well as assessing
response to treatment [74,103,144,145]. These studies consistently document close correlation between HRCT findings and clinical and pulmonary functional evaluation of these patients.

HRCT may be used to closely monitor potentially reversible morphologic changes as a means for monitoring disease progression and treatment therapy. Shah et al., using a modification of the scoring system proposed by Bhalla et al. [74], reported findings in 19 symptomatic patients who had adult CF evaluated initially and after 2 weeks of therapy, compared to a control group of eight asymptomatic CF patients [100]. Reversible findings included air-fluid levels in bronchiectatic cavities, centrilobular nodules, mucous plugging, and peribronchial thickening. Significantly, whereas severity of bronchiectasis was found to correlate with FVC (p = .004) and FEV₁ (p = .02), no correlation was identified between pulmonary function test (PFT) parameters and either mucous

plugging or centrilobular nodules, suggesting that PFTs were an insensitive means for identifying potentially reversible and hence treatable disease [100].

In a related study, Helbich et al. [103] evaluated serial CT studies obtained at various time intervals of up to 48 months in 107 patients to determine both the evolution of findings and optimal time intervals for sequential CT evaluation. These authors found that 6 to 18 months of follow-up were valuable for identifying potentially reversible morphologic changes, in particular, the presence of mucous plugging. In distinction, bronchiectasis and mosaic perfusion progressed at a significantly slower rate, rendering them less useful as a means for monitoring therapeutic interventions. Of particular interest was the finding that although CT correlated significantly with PFTs and clinical scores, these same parameters by comparison with CT were relatively insensitive means for identifying either improvement or disease progression [103]. Whereas mucous plugging could be identified in 25% of patients reexamined by CT within 18 months, for example, only minor changes could be identified by PFTs.

These findings lend support to the notion that HRCT should be incorporated into follow-up regimens of patients who have CF. In distinction, it has been reported that CT may play only a limited role in the preoperative assessment of patients who have CF before lung transplantation [146]. In a retrospective review of 26 patients who had CF who subsequently underwent bilateral lung transplantation, in no case was an unsuspected malignancy identified [146]. Of particular surgical interest, CT proved of little value in predicting the presence of pleural adhesions, a potential concern before transplantation.

HRCT is uncommonly indicated in the routine assessment of patients who have asthma, but it is sometimes used when complications, particularly ABPA, are suspected [36], and in documenting the presence of emphysema in smokers with asthma [155,156]. ABPA is associated with more severe bronchiectasis than that typically seen in patients who have uncomplicated asthma.

HRCT findings in patients who have uncomplicated asthma include mild bronchial dilatation. Mild bronchial dilatation has been reported in from 15% to 77% of patients who have uncomplicated asthma (Fig. 8-5) [36,38,65,152,157]. In a study by Lynch et al. [38], bronchi were defined as dilated if their internal diameters exceeded those of accompanying pulmonary arteries. Using this criteria, 77% of asthmatic patients and 153 (36%) of 429 bronchi assessed in asthmatic patients were considered dilated (Fig. 8-5). In a study by Grenier et al., bronchiectasis was found in 28.5% of the asthmatic subjects, primarily involving subsegmental and distal bronchi. As noted by Lynch and others [38,66], bronchial dilatation in asthmatic patients may partially reflect reduction in pulmonary artery diameter, due to changes in blood volume or local hypoxia, or may be physiologic; Lynch suggests caution in diagnosing mild bronchiectasis in this patient population [66].

Experimental studies in dogs [14] and asthmatic subjects [13,158] have measured bronchial luminal diameter using HRCT before and during a histamine or methacholine-induced episode of bronchospasm. These studies have found a significant reduction in the luminal diameter of small bronchi in association with acute asthma. Also, a significant decrease in lung attenuation due to air-trapping was seen in association with induced bronchospasm in these subjects [13].

Bronchial wall thickening, mucoid impaction, and centrilobular bronchiolar abnormalities such as tree-in-bud, patchy areas of lucency, and regional air-trapping on expiratory scans may also be identified on HRCT in patients who have uncomplicated asthma. Bronchial wall thickening has been reported in from 16% to 92% of patients [38,65,152,157], and there appears to be some tendency for the degree of bronchial wall thickening to correlate with the severity of disease [65,76].

Mucoid impaction has been reported in as many as 21% of cases [152]; this abnormality may clear after treatment. Branching or nodular centrilobular opacities have been
reported to be present in as many as 10% to 21% of patients, sometimes manifested as tree-in-bud. These likely reflect bronchiolar wall thickening or inflammation, with or without mucoid impaction. However, this finding is absent or tends to be inconspicuous in most patients who have asthma.

Focal or diffuse hyperlucency has been observed on inspiratory scans in from 18% to 31% of cases [38,157,159], undoubtedly due to air-trapping and mosaic perfusion. Expiratory CT can show evidence of patchy air-trapping in asthmatic patients (see Figs. 3-149 and 3-151) [160]. In a study by Park et al. [65], air-trapping involving more than a segment was seen in 50% of asthmatic patients. In some patients, air-trapping may be seen in the absence of morphologic abnormalities visible on inspiratory scans [117,161].

HRCT findings in patients who have ABPA have been well described (Table 8-6) (Fig. 8-11 and Figs. 8-36, 8-37, 8-38 and 8-39) [36,37,63,66,67,101,132,164,165,166,167,168,169]. A characteristic finding of
central bronchiectasis can be identified in nearly all cases. As documented by Panchal et al. in their study of 23 patients who had ABPA, central bronchiectasis could be identified in 85% of lobes and 52% of lung segments using 4- and 8-mm-thick sections [169]. Central bronchiectasis typically occurred in association with bronchial occlusion due to mucous plugging; air-fluid levels in dilated, cystic airways; and bronchial wall thickening.

In addition to widespread and severe central bronchiectasis, a number of ancillary findings have been reported to
occur in patients who have ABPA. As noted by Webb et al., disease involving the small airways is often present resulting in either a tree-in-bud appearance due to mucus-filled bronchioles or mosaic perfusion, and air-trapping due to bronchiolar obstruction (Figs. 8-11 and 8-39) [161].

The finding of aspergillomas in ectatic airways in patients who have ABPA has also been reported [170]. Additional parenchymal abnormalities, including consolidation, collapse, cavitation, and bullae, may be identified, especially in the upper lobes in as many as 43% of cases [169]. An identical
percentage of cases also had evidence of pleural abnormalities, especially focal pleural thickening. Masslike foci of eosinophilic pneumonia may also be seen in ABPA patients who have acute exacerbations [171,172].

Of particular interest is the finding of high-attenuation mucoid impaction (Fig. 8-40) [173,174]. First described in association with chronic fungal sinusitis, high-density mucus presumably represents the presence of calcium ions, metallic ions, or both, within viscous mucus [175]. The prevalence of this finding has been noted to be as high as 28% in one series, and when present should be considered characteristic [174].

ABPA should be distinguished from both angioinvasive and airway-invasive aspergillosis [176]. These latter entities occur almost exclusively in severely immunosuppressed individuals–for example, after bone marrow or renal transplantation or in patients who have leukemia. By definition, the airway-invasive form of the disease is associated with the presence of organisms deep to the airway basement membrane. In distinction to the angioinvasive form of disease, airway-invasive aspergillosis may occur clinically even when the degree of immunosuppression is mild. Radiologically, airway-invasive aspergillosis most often manifests as patchy areas of parenchymal consolidation. On HRCT scans, the majority of patients have been reported to have a distinct peribronchial or peribronchiolar distribution of disease, ranging from patchy areas of con
solidation to poorly defined centrilobular nodules [176]. In one study, sequential CT studies showed that airway-invasive aspergillosis resulted in bronchiectasis in cases without prior airway dilatation [176].

Rheumatoid arthritis (RA) may be associated with a variety of parenchymal abnormalities, including pulmonary fibrosis, bronchiolitis obliterans organizing pneumonia (BOOP), respiratory tract infections (including tuberculosis), and necrobiotic nodules. Airways disease, including both bronchiectasis and bronchiolectasis, is often overlooked as an association with RA [188]. Although airways disease was previously reported to occur in 5% to 10% of patients who have RA, since the introduction of HRCT, it has been estimated that bronchiectasis occurs in up to 35% of all patients who have this disease [189,190,191,192]. McDonagh et al., for example, in an evaluation of 20 patients who had clinical and radiologic evidence of RA, found that bronchiectasis could be identified in six patients, including two previously
thought to have diffuse interstitial lung disease, on the basis of chest radiographs [190]. In this same study, bronchiectasis was also identified in four of 20 asymptomatic control patients who had documented RA and normal chest radiographs. In a study of 38 patients who had documented RA reported by Remy-Jardin et al., there was evidence of either bronchiectasis or bronchiolectasis in 23 (30%) including 8% of asymptomatic patients, with additional features suggestive of small airways disease in another six patients [189]. These findings included linear or branching centrilobular opacities, or both, believed by these authors to represent BO [189]. Whereas this included seven patients who had traction bronchiectasis in areas of honeycombing, in the remaining 16 patients bronchiectasis was seen in the absence of CT evidence of lung fibrosis.

More recently, it has been shown that in a select population of RA patients who have normal radiographs, the prevalence of airways disease may be considerably higher than previously thought. In a study of 50 RA patients who did not have radiographic evidence of lung disease, Perez et al. [192] reported findings of both direct and indirect bronchial and/or bronchiolar disease in 35 (70%), including air-trapping on expiratory scans in 32%, cylindrical bronchiectasis in 30%, mosaic attenuation on inspiratory scans in 20%, and centrilobular nodules in 6%. Importantly, HRCT depicted features of small airways disease in 20 of 33 patients who had normal pulmonary function tests.

It has long been observed that airway obstruction is especially common in patients who have RA [193], leading to speculation by some to a possible etiologic relationship between bronchiectasis and RA. It has been suggested, for example, that chronic bacterial infections may trigger an immune reaction in genetically predisposed individuals, leading to autoimmune disease [188]. In this regard, it has been observed that bronchiectasis may precede the development of RA by several decades. However, in distinction, it has also been suggested that steroid therapy or related treatment, especially with immunosuppressive therapy itself, may lead to an increased incidence of respiratory infections. It has also been suggested that the association between RA and bronchiectasis may reflect a shared genetic predisposition, although this remains controversial [191]. More recently, it has been suggested that airway obstruction in RA patients is only secondarily related to bronchiectasis and in fact primarily reflects the presence of BO [189,194].

Regardless of the precise role played by bronchiectasis in the etiology of RA, it has been shown that although there is no evidence to suggest that patients who have coexisting bronchiectasis have more severe RA, these individuals appear to have decreased survival. Swinson et al. [195], in a case-control study of the 5-year survival of 32 patients who had both RA and bronchiectasis matched for age, gender, and disease duration with 32 patients who had RA alone, and an additional 31 patients who had bronchiectasis alone, found that patients who had both RA and bronchiectasis were 7.3 times as likely to die than the general population, five times more likely to die than patients who had RA alone, and 2.4 times more likely to die than patients who had bronchiectasis alone.

Similar to the unexpectedly high incidence of bronchiectasis in patients who have RA, it has been shown that bronchiectasis may be identified in up to 20% of patients who have systemic lupus erythematosus (SLE) [196]. Banker et al., in a prospective study of 45 patients who had documented SLE and normal radiographs, found abnormal CT findings in 38%, including bronchial wall thickening in 20% and bronchial dilatation in another 18% [197]; bronchiectasis has usually not been considered a manifestation of SLE [188]. A similarly unexpected high prevalence of airway pathology has also been noted in patients who have primary Sjögren’s syndrome [198].

A wide range of airway abnormalities has been identified in patients who have ulcerative colitis. In addition to BOOP and diffuse interstitial lung disease, these include subglottic stenosis, chronic bronchitis, and chronic suppurative inflammation of both large and small airways [188]. Similar to what is seen in patients who have RA, chronic suppurative airways disease may precede, coexist with, or follow, the development of inflammatory bowel disease. Of particular interest is the fact that, unlike other causes of bronchiectasis, chronic suppurative airways disease associated with ulcerative colitis frequently responds to treatment with inhaled steroids [188].

A number of reports have documented an increase in the prevalence of airway-related infections in HIV-positive/AIDS patients [31,34,199,200,201,202]. Paralleling a marked decrease in the incidence of PCP due to routine prophylaxis, lower respiratory tract infections, including bacterial pneumonia and bronchitis, have superseded PCP as the most common infections in the lungs of AIDS patients [203,204,205]. Wallace et al., for example, in a study of more than 1,000 HIV-positive patients who did not have an AIDS-defining illness, found a significantly higher incidence of acute bronchitis compared to HIV-negative subjects [205]. Most commonly, they result from organisms such as Haemophilus influenzae, Pseudomonas aeruginosa, Streptococcus viridans, or Streptococcus pneumoniae, and in fact a wide range of infectious agents has been described affecting the airways, including both mycobacterial and fungal infections [199].

First reported by Holmes et al. [201], and later confirmed by McGuinness et al. [31], an accelerated form of bronchiectasis appears to occur in HIV-positive patients. Although the etiology is unclear, it is likely that bronchiectasis results from recurrent bacterial infections affecting airways, possibly made more susceptible due to the
direct effects of HIV infection on the pulmonary immune system. A correlation has also been shown between airway dilatation identified by CT and the presence of elevated levels of neutrophils on bronchoalveolar lavage (BAL). In a study comparing BAL findings in 50 HIV-positive subjects with 11 HIV-negative individuals, King et al. showed that patients who had bronchial dilatation on CT had significantly higher BAL neutrophil counts (p = .014) as well as significantly lower diffusing capacity (p = .003) [34]. As noted by these authors, neutrophils are an important mediator of pulmonary damage, possibly due to the action of human neutrophil elastase.

AIDS-related airways disease is manifested by a variety of HRCT abnormalities, a fact that reflects the number of organisms that may be involved. Findings include bronchial wall thickening, bronchiectasis, bronchial or bronchiolar impaction with tree-in-bud, endoluminal masses or nodules, and consolidation [199]. A lower lobe predominance is typical. The common finding of air-trapping in HIV-positive patients has also been stressed [206], suggesting that small airways disease may be a significant contributor to pulmonary function decline and may precede more obvious findings of airways disease. Gelman et al. [206] evaluated the HRCT scans of 59 subjects using inspiratory and expiratory HRCT, 48 of whom were HIV-positive and 11 of whom were HIV-negative. Expiratory CT revealed focal air-trapping in 33 subjects, 30 of whom were HIV-positive and three of whom were HIV-negative (p = .0338). The mean values of FEV₁, forced mid-expiratory flow, and diffusion capacity were significantly lower for subjects with focal air-trapping than for those with normal findings on CT (p = .001, p = .021, and p = .003, respectively).

It should also be noted that although less common than bacterial or viral infections, the airways may also be the site of fungal infections, especially due to Aspergillus (Fig. 8-43). Typically occurring late in the course of HIV infection, and usually in association with other risk factors, including corticosteroid use and granulocytopenia, approximately 10% of all reported cases of aspergillosis in AIDS patients will affect the airways [199]. Whereas several distinct subtypes of aspergillosis of the airways have been described, including necrotizing acute tracheobronchitis, obstructing bronchial aspergillosis, and chronic cavitary parenchymal aspergillosis, it is likely that these represent a single spectrum of fungal disease potentially affecting the airways [207,208,209].

On one end of the spectrum is obstructing bronchial aspergillosis. Representing a stage before frank tissue invasion, obstructing bronchial aspergillosis typically presents acutely with fever, dyspnea, and progressive cough associated with the expectoration of fungal casts and is characterized on CT by the presence of mucoid impaction typically involving the lower lobe airways [209]. It has been suggested that this form of disease is unique to AIDS patients. Rarely, aspergillosis primarily affects the small airways again in the absence of bronchial inflammation. Described in only two patients who had AIDS, this form of infection has been termed chronic cavitary parenchymal aspergillosis and results in necrotic debris and fungi filling respiratory bronchioles with extension into adjacent alveolar spaces in the absence of bronchial invasion [210]. In distinction, necrotizing tracheobronchial aspergillosis (also referred to as diffuse, ulcerative, and pseudomembranous tracheobronchitis) is associated with frank tissue invasion resulting on CT in a range of abnormalities, from subtle focal irregularity and nodularity of airway walls to airway obstruction with or without accompanying parenchymal infiltrates.

Included in the category of cellular bronchiolitis is a diverse set of diseases characterized in common by the presence of inflammatory cellular infiltrates involving both the bronchiolar lumen and wall and some degree of fibrosis. Further characterized as acute or chronic, and by the predominant cell type, this classification most commonly includes (i) infectious bronchiolitis (bacterial, viral, mycoplasma, and fungal), (ii) follicular bronchiolitis in collagen-vascular diseases, (iii) panbronchiolitis, (iv) aspiration bronchiolitis, (v) bronchiolitis associated with hypersensitivity pneumonitis, and (vi) asthma. Nonspecific cellular bronchiolitis is also frequently identified in patients who have bronchiectasis and chronic obstructive pulmonary disease.

Respiratory bronchiolitis is part of a spectrum of smoking-related diseases of the airways and lungs, characterized by the accumulation of pigmented macrophages within respiratory bronchioles and alveoli. This spectrum includes respiratory bronchiolitis (RB), respiratory bronchiolitis-interstitial lung disease (RB-ILD), and desquamative interstitial pneumonitis (DIP), also known as alveolar macrophage pneumonia [213,214,215,216,217]. Although RB is typically identified as an incidental finding in asymptomatic smokers, it has been recognized that some smokers present with both signs and symptoms of diffuse interstitial lung disease, so-called RB-ILD. Histologically, these patients are considered to have an exaggerated form of RB with evidence of inflammation and fibrosis away from respiratory bronchioles and into adjacent alveolar septa. In distinction, patients who have DIP have more diffuse involvement with a greater likelihood of developing parenchymal fibrosis and progressive lung disease. It has been suggested that RB, RB-ILD, and DIP be further classified as one of a larger group of smoking-related diseases, aptly termed smoking-related interstitial lung diseases, that also includes centrilobular emphysema and Langerhans cell histiocytosis [216]. Support for this classification stems from a possible association between respiratory bronchiolitis and the development of centrilobular emphysema [216,217].

Constrictive bronchiolitis is defined histologically by the presence of concentric fibrosis predominantly between the bronchiolar epithelium and the muscularis mucosa, resulting in marked narrowing or obliteration of bronchioles, or both, in the absence of intraluminal granulation tissue polyps or surrounding parenchymal inflammation. Clinically, constrictive bronchiolitis is associated with marked airflow obstruction that usually is not responsive to steroid therapy [218,219]. Radiographically, little if any abnormality apart from hyperinflation is apparent. This pathologic entity is associated with the clinical syndrome often referred to as BO. It may result from a variety of diseases or conditions, described below in detail.

BO with intraluminal polyps, also known as BOOP or cryptogenic organizing pneumonia (COP), is defined by the presence of granulation tissue polyps within respiratory bronchioles and alveolar ducts (Masson bodies) associated with patchy organizing pneumonia [220,221,222]. Because in the majority of cases the predominant histologic abnormality is the presence of organizing pneumonia, the terms BOOP and COP are usually used to describe this entity.

Two distinct patterns of intraluminal fibrosis or Masson bodies have been described [223]. Type 1 Masson bodies are
those that show abundant myxoid matrix, sparse fibrosis, and little or no fibrin and presumably represent immature mesenchymal cells. In distinction, Type 2 Masson bodies contain fibrin and have histochemical properties of myofibroblasts. This distinction may be of importance, as in at least one study patients who had Type 1 disease showed good response to steroid therapy, whereas those who had Type 2 disease generally failed to respond to treatment [223].

Although the majority of cases of BOOP are idiopathic, similar findings may be seen in association with a variety of clinical conditions. As described by Katzenstein [219], BOOP may be classified into three separate categories. In the first, BOOP is the primary cause of respiratory illness. This includes idiopathic BOOP as well as BOOP resulting from RA, toxic inhalants, drug toxicity, and collagen-vascular diseases; prior infection, both viral and bacterial; and acute radiation pneumonitis [219,224]. In the second category are cases in which BOOP is found as a nonspecific reaction along the periphery of unrelated pathologic processes, including neoplasms, infectious granulomas, vasculitis, and even infarcts. Finally, BOOP may also be identified as a minor component of other diseases, including hypersensitivity pneumonitis, nonspecific interstitial pneumonitis, and Langerhans cell histiocytosis, among others [219]. It is for this reason that the finding of features consistent with BOOP on transbronchial biopsy is of only limited value in the absence of detailed clinical and radiologic correlation (Fig. 8-52 and 8-53) [219].

Reflecting not only the presence of intraluminal polyps, but also more extensive inflammatory changes involving alveolar ducts and alveoli, BOOP characteristically results in predominantly restrictive lung disease manifested radiographically as ill-defined areas of parenchymal consolidation [211,218,219,225]. BOOP is described in Chapter 6.

Airways distal to those containing identifiable cartilage in their walls are termed bronchioles. Terminal bronchioles are the last purely conductive airways, varying in length from 0.8 to 2.5 mm, and are usually found between the sixth and twenty-third airway generations. Respiratory bronchioles lie distal to terminal bronchioles and are defined as airways in which the ciliated epithelial lining is interrupted by alveoli [148,238]. Along with alveolar ducts and sacs, respiratory bronchioles comprise the gas-exchanging unit of the lung. A number of different cell types can be identified in normal bronchioles. These include special columnar secretory stem cells, known as Clara cells, as well as neuroendocrine or Kulchitsky cells. In distinction, mucus-secreting goblet cells and bronchus-associated lymphoid tissue are only rarely identified in nonsmokers.

Along with their accompanying pulmonary artery branches, bronchioles lie within the center of secondary pulmonary lobules. Normally, individual airways can only be identified
when their walls are larger than 300 microns, corresponding to bronchi between 1.5 and 2 mm in size. As a consequence, normal bronchioles with luminal diameters of approximately 0.6 mm cannot be identified on HRCT.

Although the term small airways disease is usually used synonymously with bronchiolar disease, this term as originally defined was used to describe inflammatory changes in peripheral airways in smokers resulting in moderate to severe airflow obstruction [239]. Subsequently defined by Macklem and colleagues [240] to indicate an idiopathic syndrome of chronic airflow obstruction in patients having no evidence of underlying emphysema or chronic bronchitis, this concept of small airways disease is essentially physiologic and is associated with abnormalities involving airways between 2 and 3 mm in diameter. The distinction between anatomic and physiologic definitions is significant, as small airways contribute only approximately 10% of total airway resistance. As a consequence, numerous small airways can be damaged before conventional measurements of lung function, in particular measurements of airway resistance, become abnormal.

Although normal intralobular bronchioles cannot usually be identified, direct and indirect signs of bronchiolar disease have been described [18,19,161]. Direct signs result from the presence of bronchiolar secretions, peribronchiolar inflammation, or, less commonly, bronchiolar wall thickening. Characteristically, these result in either branching or Y-shaped linear densities (Figs. 8-17 and 8-44), or poorly defined centrilobular nodules (Fig. 8-45). Less commonly, bronchiolar inflammation results in the finding of small centrilobular lucencies due to bronchiolectasis [3,70,85,89].

Indirect signs have also been described, the most important of which are the findings of mosaic perfusion on inspiratory scans and areas of lucency and air-trapping, either focal or global, on scans obtained at end expiration [116,161,241,242].

Using a combination of HRCT findings, it is possible to classify bronchiolitis into one of four basic HRCT patterns based on the predominant abnormality (Table 8-9). These include (i) bronchiolar diseases associated with a tree-in-bud appearance, (ii) bronchiolar diseases associated with poorly-defined centrilobular opacities, (iii) bronchiolar diseases associated with areas of decreased lung attenuation, and (iv) bronchiolar diseases associated with focal or diffuse ground-glass opacity, consolidation, or both. Use of this classification in conjunction with clinical findings frequently allows precise diagnoses even in the absence of histologic correlation.

The finding of a TIB pattern is nearly always due to acute infectious bronchiolitis, regardless of the underlying disease (Figs. 8-17 and 8-47A). As documented by Im et al. [27], in patients who had tuberculosis, the TIB pattern correlates pathologically with the presence of secretions within dilated terminal and respiratory bronchioles (Fig. 8-44). These are characteristically centrilobular in distribution and are most easily identified in the lung periphery. Ill-defined buds associated with the branching airway reflect the presence of peribronchiolar granulomas, a finding especially common in patients who have chronic infection due to MAC (Fig. 8-30) [84] or peribronchiolar inflammation. Poorly defined centrilobular nodules or rosettes of nodules are almost always visible in patients who have TIB; these may be seen if images show distended centrilobular bronchioles in cross section. In our experience, however, these are always ancillary to the main finding of linear or branching densities. Additional findings include areas of ground-glass opacity, consolidation, or both. With healing, a distinctive pattern of branching or V-shaped densities associated with secondary lobular emphysema may be identified, presumably the result of bronchial, bronchiolar, or both types of obstruction.

Infectious bronchiolitis is usually reversible and is most often caused by Mycoplasma pneumoniae, Haemophilus influenzae, and Chlamydia species and viral infections, especially respiratory syncytial virus. Infectious bronchiolitis is common in infants and children and is being increasingly diagnosed in adults, especially those who have atypical mycobacterial infections or other causes of chronic airways disease [28,29], AIDS [34,199], or a combination of both. Although histologic examination typically shows necrosis of respiratory epithelium with a mixed cellular infiltrate usually in association with accompanying pneumonitis, it is, in fact, rarely necessary to obtain biopsies in these patients. In our experience, HRCT findings in conjunction with clinical findings are usually sufficient to allow presumptive therapy with antibiotics, pending the results of sputum culture.

It should be emphasized that an appearance similar to that of TIB has also been reported in association with noninfectious etiologies. Numerous reports, for example, have documented linear or branching centrilobular densities in patients
who had RA or Sjögren’s syndrome [189,192,198,243]. Although usually attributed to either follicular bronchiolitis [244,245,246] or BO, many of the patients have coexisting productive cough and other clinical signs indicative of possible infection. In this regard, Hayakawa et al. correlated HRCT findings with histologic findings identified on open lung biopsy in 14 RA patients who had documented bronchiolitis [247]. Whereas linear or branching centrilobular nodules could be identified in five of seven patients who had follicular bronchiolitis and three of seven who had BO, as noted by the authors, 73% of patients had chronic sinusitis, and 93% of patients had chronic cough with sputum. Furthermore, bacterial cultures were positive in 50% and 71% of patients who had follicular bronchiolitis and BO, respectively.

In Asia, especially in Japan, a form of diffuse panbronchiolitis termed DPB is common (Figs. 8-63 and 8-65) [89]. DPB has been defined as a clinicopathologic entity, characterized by symptoms of chronic cough, sputum, and dyspnea; associated with abnormal PFTs, usually indicating mild to moderate airway obstruction and radiographic evidence of ill-defined nodular infiltrates typically basilar in distribution [248]. Histologically, DPB is characterized by chronic inflammation with mononuclear cell proliferation and foamy macrophages predominantly involving the walls of respiratory bronchioles, adjacent alveolar ducts, and alveoli, constituting the so-called unit lesion of panbronchiolitis [248]. Inclusion of this entity in the

category of infectious etiologies of bronchiolitis is justified, as nearly all patients develop superinfection with Pseudomonas. On HRCT, a characteristic diffuse TIB pattern is invariably seen, predominantly affecting the lung bases, that has been shown to disappear after treatment with erythromycin (Fig. 8-64) [249]. Despite initial improvement, DPB is usually considered a progressive disease, with 5- and 10-year survival rates reported to be in the range of 60% and 30%, respectively [248]. DPB is described in further detail below.

RB is a smoking-related disease of the airways and lungs characterized by the accumulation of pigmented macrophages within respiratory bronchioles and alveoli [213,214,215,216]. RB is typically identified as an incidental finding in asymptomatic smokers; if symptoms are associated, this disease is termed RB-ILD.

HRCT findings in RB and RB-ILD have been described, and are discussed in detail in Chapter 6 (Fig. 8-49). In most patients who have documented RB, the lungs appear either normal or show evidence of poorly defined, predominantly middle and upper lobe centrilobular ground-glass opacities, with or without accompanying areas of diffuse ground-glass opacity [235]. Findings in the few reported series published to date in patients who had RB-ILD have proven more variable. First reported in five patients by Holt et al., findings ranged from normal lungs to either ill-defined centrilobular and/or diffuse ground-glass opacity, or bibasilar areas of atelectasis and/or scarring [250]. In another study of eight patients evaluated by HRCT reported by Moon et al., five of the eight had evidence of both areas of ground-glass opacity and mild reticulation, whereas one case showed evidence of only diffuse ground-glass opacity. In all, six cases were interpreted as consistent with either RB-ILD or DIP. Interestingly, in two cases there was also evidence of diffuse centrilobular emphysema [216].

Hypersensitivity pneumonitis is associated with a chronic, nonspecific, predominantly interstitial lymphocytic pneumonitis that early in its course is primarily distributed around respiratory bronchioles with relative sparing of intervening lung [219]. This pattern may persist even later in the course of disease. Additionally, in nearly two-thirds of cases, there is also evidence of nonnecrotizing granulomas, again typically localized to the peribronchiolar interstitium. Foci of BOOP may also be identified in some cases, with characteristic findings of intraluminal fibrous plugs.

Together, these findings result in a pattern of diffuse, poorly defined centrilobular nodular opacities, especially in the subacute phase of disease (Figs. 8-45 and 8-50) [251,252,253,254,255,256,257]. These nodules are typically uniform in distribution. On occasion, their proliferation is so widespread as to result in diffuse ground-glass opacity. Close inspection, however, invariably discloses
the presence of innumerable centrilobular ground-glass nodules accounting for this appearance. Findings in hypersensitivity pneumonitis are described in greater detail in Chapter 6.

Another feature described in these patients is the finding of focal air-trapping, frequently restricted to secondary pulmonary lobules. Whereas this may be the result of inhomogeneous distribution of inhaled causative agents resulting in sparing of individual pulmonary lobules, it has alternatively been speculated that air-trapping may represent associated obliterative bronchiolitis [117,257].

Follicular bronchiolitis represents nonspecific lymphoid hyperplasia of the bronchus-associated lymphoid tissue. Histologically, follicular bronchiolitis is characterized by the finding of hyperplastic lymphoid follicles with reactive germinal centers, typically characteristically distributed along bronchioles, and, to a lesser extent, bronchi. Extension into the lung interstitium with resulting pulmonary infiltrates is generally not considered a feature of follicular bronchiolitis [258]. Most cases are associated with underlying disorders, most often collagen-vascular diseases, in particular RA [189,245,247] and Sjögren’s syndrome, immunodeficiency disorders, and hypersensitivity reactions [259]. In children, the disease typically occurs in the first 6 to 8 weeks of life, results in respiratory distress and fever, and is usually unresponsive to either bronchodilator therapy or steroids [258].

CT findings in adult patients who have follicular bronchiolitis have been described in a few small series [189,244,260]. Cardinal features include bilateral centrilobular, peribronchial, or both types of ground-glass opacity nodules measuring 3 to 12 mm in diameter, corresponding histologically with bronchial and peribronchial lymphoid infiltration (see Fig. 5-14). Areas of ground-glass opacity may also be identified, but never in isolation. In the majority of cases, although characteristic, definitive diagnosis requires histologic confirmation. Interestingly, CT findings in one report of individuals with RA and documented follic

ular bronchiolitis showed that in all cases in which both branching and poorly defined centrilobular nodules were evident, the disease was either stabilized or reversed after treatment with erythromycin, reinforcing the notion that a TIB pattern likely reflects underlying infection [247].

Similar CT findings have been identified in patients who have lymphocytic interstitial pneumonitis (LIP), especially in HIV-positive and AIDS patients (Fig. 8-51) [33]. Characterized by an interstitial infiltrate of mature lymphocytes, LIP is part of the spectrum of hyperplasia of bronchus-associated lymphoid tissue that also includes follicular bronchiolitis. Although characterized by more diffuse interstitial involvement, histologic differentiation between follicular bronchiolitis and LIP may be difficult, raising the possibility that the poorly defined centrilobular nodules identified on CT in most cases of LIP diagnosed by transbronchial biopsy actually represent follicular bronchiolitis.

BOOP is characterized histologically by the presence of granulation tissue polyps within respiratory bronchioles and alveolar ducts associated with patchy organizing pneumonia [220,221,222]. Ill-defined centrilobular nodules reflecting bronchiolitis and focal organizing pneumonia may be seen. BOOP more often results in patchy consolidation.

The HRCT appearance of BO has been described in a number of studies (Table 8-10) [18,20,39,43,87,141,230,234,242,293,294], and characteristic abnormalities have been reported in patients who have both idiopathic and secondary BO [18,230,234,294]. HRCT findings are similar regardless of the cause of disease (Fig. 8-54).

The most obvious HRCT finding is often that of focal, sharply defined areas of decreased lung attenuation associated with vessels of decreased caliber (Fig. 8-54A and C). These changes represent a combination of air-trapping and oligemia, typically occurring in the absence of parenchymal consolidation, and termed mosaic perfusion [230]. Bronchiectasis, both central and peripheral, may be present as well (Fig. 8-54A) [42]. Rarely, 2- to 4-mm centrilobular branching opacities, representing inspissated secretions within distal airways or ill-defined centrilobular opacities, may be the predominant finding [85,234], but recognizable small airway abnormalities are usually inconspicuous in patients who have BO.

Air-trapping is commonly visible on expiratory HRCT in patients who have BO (Fig. 8-54 B, D, and E). In fact, the presence of air-trapping on expiratory scans may be the only abnormal HRCT finding in patients who have BO (Fig. 8-55) [117]. In a study by Arakawa and Webb of 45 patients who had air-trapping found on routine expiratory HRCT scans [117], nine patients had normal inspiratory HRCT findings; five of these nine patients had BO.

Abnormal findings are far more evident on HRCT than on chest radiographs in patients who have BO. For example, in one study of patients who had BO [294], chest radiographs were normal in one-third of patients and showed mild hyperinflation and vascular attenuation in the remaining two-thirds. CT, on the other hand, showed widespread and conspicuous abnormalities in lung attenuation in nearly 90% of the patients.

Chang et al. [261] assessed the long-term clinical, imaging, and pulmonary function sequelae of postinfectious BO in 19 children. Clinical follow-up averaging 6.8 years revealed a high incidence of continuing problems, largely relating to asthma and bronchiectasis. Fixed airway obstruction was the most common pulmonary function abnormality. Chest radiographs showed five patterns: (i) unilateral hyperlucency of increased volume, (ii) complete collapse of the affected lobe, (iii) unilateral hyperlucency of a small or normal-sized lung, (iv) bilateral hyperlucent lungs and a mixed pattern of persistent collapse, and (v) hyperlucency and peribronchial thickening.

Lynch et al. [141] reported the HRCT findings of post-infectious BO in six children. The most striking finding in these patients was the presence of focal areas of decreased lung opacity, which usually had sharp margins. These areas of decreased opacity corresponded to segments or lobules, and the pulmonary vessels within these areas appeared to be reduced in size. Four of these six had bronchiectasis visible on HRCT (Fig. 8-56), and in all four the abnormal bronchi were in areas of lung showing decreased opacity. It is likely that the areas of decreased opacity represent regions of lung that are poorly ventilated and perfused. Areas of increased lung opacity containing vessels that were normal or large in size were also seen in this series; these areas of increased opacity reflect well-perfused areas of lung or mosaic perfusion.

Zhang et al. [295] performed a prospective study to define the HRCT features of 31 pediatric patients who had postinfectious BO. All patients had chest radiographs and lung perfusion scans, and 27 of the 31 patients had HRCT of the lung. The most common abnormal features shown on CT included bronchial wall thickening (100%), bronchiectasis (85%, graded as moderate or severe in 30%), and areas of increased and decreased attenuation (82%), likely due to mosaic perfusion secondary to air-trapping. Perfusion defects on radionuclide imaging were found in all patients. Lobular, segmental, or subsegmental atelectasis was also common, being seen in 70% of cases.

In this study [295], HRCT showed a higher sensitivity than chest radiography in detecting pulmonary abnormalities. Although bronchial wall thickening was seen on radiographs in all cases showing this finding on HRCT, areas of decreased opacity were seen in only 59% of HRCT-positive cases, and bronchiectasis was seen in only 35%.

BO is a major component of Swyer-James or MacLeod syndrome [296,297]. In patients who have Swyer-James syndrome, BO is the result of lower respiratory tract infection, usually viral, occurring in infancy or early childhood. Damage to the terminal and respiratory bronchioles leads to incomplete development of their alveolar buds. The radiographic hallmark of this syndrome is unilateral hyperlucent lung with reduced lung volume on inspiration and air-trapping on expiration (Figs. 8-56 and 8-57). Although previously necessitating confirmation either by bronchography (to demonstrate extensive bronchiecta
sis), or arteriography (to demonstrate a small central pulmonary artery and decreased peripheral vascularity), these procedures have been all but obviated by HRCT.

Marti-Bonmati et al. [293] described the CT findings in nine patients who had Swyer-James syndrome. On CT, the affected lung showed decreased opacity in eight patients (Figs. 8-56 and 8-57); in the one remaining patient, the affected lung was very small but of normal attenuation. Lung volume on the affected site was reduced in six patients and normal in three; one patient showed normal lung opacity on chest radiographs but decreased opacity on CT. In all patients, the size of the affected lung did not change on CT scans obtained during inspiration and expiration.

All nine patients had CT findings of bronchiectasis [293]. In each, cylindrical bronchiectasis was present, but two also had cystic bronchiectasis and three had varicose bronchiectasis. The lower lobes were affected in eight patients, the middle lobes or lingula in seven patients, and the upper lobes in three patients. Parenchymal abnormalities were present in eight patients.

Idiopathic BO usually affects middle-aged women and has a relentless downhill course despite steroid therapy. A more benign course has been described in a limited number of patients who have more stable disease, suggesting a wider range of clinical presentations than previously appreciated [280].
Sweatman et al. [294] described the CT findings in 15 patients who had idiopathic BO. The chest radiograph was normal in five patients and showed mild hyperinflation and vascular attenuation in the remaining ten. CT showed widespread abnormalities in 13 of the 15 patients (87%), consisting of patchy irregular areas of high and low attenuation in variable proportions (Fig. 8-58). These changes were accentuated on expiration.

HRCT findings in patients who have DPB have been extensively described (Table 8-11) [88,89,248,249,321,322]. As initially shown by Akira et al. [89], the most important findings, in order of severity, are poorly defined centrilobular
nodules, centrilobular branching opacities or TIB, and branching thick-walled centrilobular lucencies (see Fig. 3-75; Fig. 8-64). As confirmed by Nishimura et al. [88], these findings correspond to, respectively, peribronchiolar inflammation and fibrosis, dilated bronchioles with inflammatory wall thickening and intraluminal secretions, and dilated, air-filled bronchioles (Fig. 8-65) [88]. The finding of small, peripheral peribronchiolar nodules is particularly characteristic, resulting in a TIB appearance.

In addition to these findings, patients who have DPB usually show evidence of air-trapping with large lung volumes and decreased attenuation of peripheral lung parenchyma [88]. As documented by Murata et al. [322], in a study comparing positron emission tomography and CT in seven patients who had DPB, CT attenuation values were considerably lower in the periphery of the lung as compared with the central portions, a finding indicative of extensive peripheral air-trapping. Such stratified distribution of ventilatory impairment may be considered characteristic of diffuse bronchiolar narrowing.

Although the course of this disease is typically monitored clinically, HRCT may play a role in select cases. As documented by Akira et al. [321], in a study of 19 patients randomly assigned either to therapy with low-dose erythromycin or to follow-up without treatment, centrilobular nodular and branched opacities decreased in number and size in treated patients, suggesting a positive response to therapy. In distinction, among untreated patients, similar densities observed initially were found to have progressed with resultant dilatation of the proximal airways. Similar results have been shown by Ichikawa et al. [249]. These findings suggest a potential role for HRCT for monitoring as well as for predicting the outcome of therapy.

It should be noted that the finding of nodular or branching centrilobular opacities may be seen in a number of different diseases in which bronchiolar abnormalities are present [70,85]. Linear or branching centrilobular opacities result from bronchiolar dilatation with intraluminal accumulation of mucus, fluid, or pus. As noted previously, this has been termed the TIB appearance and can be identified in patients who have CF [141]; endobronchial spread of tuberculosis [27] or nontuberculous mycobacteria; lobular pneumonia or bronchopneumonia, including those with PCP or cytomegalovirus pneumonia; bronchiectasis of any cause; and other airways diseases that result in the accumulation of mucus or pus in small bronchi. The differential diagnosis of centrilobular opacities is described in detail in Chapter 3.