tors of disease progression in patients with CPFE. It is preferable that patients with a clinical syndrome of CPFE be excluded from IPF trials [12, 134] and trials of ILD in connective tissue disease [65] that use FVC as a primary endpoint. In clinical practice, serial changes in FVC and DLco are used to monitor disease progression in IPF [33], but they are not appropriate in CPFE patients, with unfortunately no clear-cut alternate functional parameter proposed so far. In one study, a decline in FEV1 by 10% or more at 6 or 12 months was useful to assess disease progression, and predicted a poor outcome [132]; FEV1:FVC might also be useful [151]. Whether these observations are useful in the clinic awaits further evaluation. Overall, disease progression in CPFE is usually monitored using a combination of clinical, imaging, and multiple functional parameters, with less emphasis on FVC trends than in the monitoring of ILD without concurrent emphysema.

Imaging

Chest radiograph may show hyperlucency of the upper zones of the lungs, and diffuse parenchymal in ltrates in the lower lobes (Fig. 33.3). HRCT of the chest has dramatically enhanced the recognition of CPFE and is key to the diagnosis. Patients with CPFE present with CT radiographic features suggestive of both emphysema and pulmonary brosis (mostly in the lower lobes). Emphysema generally predominates in the upper lobes, but may be present in the lower zones or be admixed with brosis. CPFE is characterized by a wide variety of appearances on chest HRCT.

Fig. 33.3  Chest radiograph of CPFE syndrome showing hyperlucency of the upper zones of the lungs, and diffuse parenchymal in ltrates in the lower lobes (male, smoker)

Computed Tomography Characteristics and Patterns

CPFE is characterized by the presence of emphysema and interstitial brosis. Emphysematous foci, identi ed on HRCT as a region of low attenuation not bounded by visible walls [152], can be categorized as centrilobular, paraseptal, or panacinar [153]. Fibrosis on chest HRCT is identi ed as regions of increased parenchymal attenuation, with reticulation and/or ground-glass opacities, variably associated with honeycombing, traction bronchiectasis and/or volume loss. A majority of patients have an HRCT pattern of UIP [8], however, other patterns have been reported on imaging and/ or histopathology [8, 154–157] (Table 33.3). Mild to moderate ground-glass attenuation is more prevalent in CPFE than in typical IPF patients [8], likely corresponding to NSIP [156] or to various smoking-related ILDs such as desquamative interstitial pneumonia [154] or respiratory bronchiolitis-­ associated ILD. Therefore, not all patients with (tobacco-related) CPFE have IPF [12], and not all patients with CPFE in the setting of connective tissue disease have a pattern of UIP [26].

In the seminal series [8], interstitial changes were characterized by honeycombing (95%), reticulation (87%), traction bronchiectasis (69%), and architectural or bronchial distortion (39%), predominating in the lower lung zones and in subpleural areas. Non-prominent ground-glass attenuation

Table 33.3  Main radiological and histopathological patterns in patients with CPFE. Several patterns or features may be combined in one individual

a These terms are used to describe overlapping patterns of brosis linked to cigarette smoking

was present in two thirds of the patients. Some degree of centrilobular emphysema was present in 97% of the patients [8, 123]. In addition, paraseptal emphysema present in 93% of patients with CPFE represented the predominant type of emphysema in more than half of patients [8]. Importantly, paraseptal emphysema seems to be more frequent in CPFE than in COPD [130], and may actually be considered a hallmark of the syndrome [12, 123]. Panacinar emphysema is seldom observed. Dilatation of pulmonary artery suggestive of PH may be present at HRCT.

Thick-Walled Large Cysts

One peculiar pattern observed in CPFE is that of thick-­ walled large cysts (or “air spaces with brotic walls”) of the lower zones of the lungs [26, 65, 158]. Thick-walled large cysts are frequently observed in areas where reticulation is present. They are larger than 2.5 cm in diameter and delimitated by a wall at least 1 mm thick. They may be associated or not with typical honeycombing, from which they differ by the larger size of the cysts and lack of clustering, whereas honeycombing corresponds to clustered, cystic air spaces, 3–10 mm and up to 2.5 cm in diameter [152]. They may be associated with more extensive emphysema at imaging [158]. Thick-walled large cysts likely result from the development of pulmonary brosis in the setting of emphysematous lung, with enlargement of the cysts due to retraction forces in brotic lung [147, 159]. Enlargement over time of thick-walled large cysts has been described [158]. They are considered as one feature of the CPFE syndrome [26], and may often correspond to histopathological lesions of smoking-­related interstitial brosis (SRIF) [65, 113, 158]. The evolution of these thick-walled large cysts is yet to be described.

Imaging Phenotypes

Due to the high heterogeneity of imaging in patients with CPFE, attempts were made to identify distinct imaging phenotypes [147] including:

\1.\ emphysema in the upper zones and brosis in the lung bases, with no or little overlap of emphysema and brosis in between (separate processes) (Fig. 33.4);

\2.\ progressive transition from emphysema lesions to brosis, with signi cant overlap or admixture in mid areas (Fig. 33.5);

\3.\ conspicuous paraseptal emphysema with predominant subpleural bullae/cysts with thickened walls (Fig. 33.6).

\4.\ the pattern of thick-walled large cysts (Fig. 33.7) [26], which may overlap with that of predominant paraseptal emphysema.

In addition, a number of observations do not t into one of the above categories [147]. When adjusting for severity ofbrosis, patients with a pattern of predominant paraseptal emphysema had higher (e.g., normal) FEV1:FVC and lower FVC and TLC values, with similar DLco, as compared to those with separate processes and progressive transition between emphysema and brosis [147]. In one study, emphysema admixed with brosis was associated with preserved lung volumes, while cases of CPFE with emphysema distant from brosis on HRCT were associated with more pronounced alteration of DLco [111]. However, whether these subgroups have physiological relevance, and whether the pattern of predominant paraseptal emphysema really is associated with preserved spirometry, as shown in patients with chronic obstructive pulmonary disease [160], warrant con rmation.

A pattern of UIP on HRCT is associated with greater mortality as compared to non-UIP patterns of ILD in CPFE [16]. In contrast, a pattern of admixed brosis and emphysema, suggestive of SRIF, may be associated with a better prognosis [113]. Further study is needed to decipher the potential clinical and physiologic relevance of imaging phenotypes in CPFE, especially with regard to outcome.

Pitfalls

Interpretation of HRCT imaging is particularly dif cult in patients with ILD and concurrent emphysema, owing to dif-culties to ascertain honeycomb changes in patients with associated emphysema. In other words, association of a pattern of NSIP and emphysema [156], with small thick-walled cystic changes, may falsely resemble honeycombing at HRCT [157], a situation sometimes coined “possible honeycombing” [161]. Presence of emphysema is one of the main reasons for disagreement between radiologists in the HRCT assessment of honeycombing [162]. Because the international criteria for the diagnosis of IPF [163] are largely based on HRCT imaging and especially the presence of honeycombing to de ne the UIP pattern, particular attention must be exerted to distinguish honeycombing from non-UIP interstitial changes with admixed emphysema, especially when emphysema is visible in the upper zones. Paraseptal emphysema is constituted by a single row of subpleural cystic spaces preferentially in the upper zones, contrasting to the clustered subpleural cysts (more speci c if two rows or more) in honeycombing that predominates in the lung bases [164].

Fig. 33.4  CPFE syndrome with centrilobular emphysema and usual interstitial pneumonia (male, smoker). Upper left panel: lung biopsy in right upper lobe showing centrilobular emphysema; upper right panel: chest HRCT at the level of the trachea showing moderate centrilobular emphysema, along with reticular changes; lower left panel: lung biopsy in right lower lobe showing usual interstitial pneumonia pattern, with

interstitial brosis and architectural distortion; lower right panel: chest HRCT in lower lung zones showing usual interstitial pneumonia pattern, with reticulation, traction bronchiectasis, honeycombing, and some ground-glass attenuation. Pathology slides courtesy of Dr Lara Chalabreysse, Lyon (France)

Quantifcation of Emphysema Extent

Radiographic quanti cation of emphysema features is necessary for research and to assess whether classi cation criteria for CPFE clinical syndrome (see below) may be met. Quanti cation of emphysema extent (over whole lung volume) in patients with CPFE mostly relies on semi-­quantitative visual HRCT assessment. It may be challenging, however, with interobserver variation. Computer-based measurement of lung density is poorly suited in CPFE due to dif culties in

differentiating honeycomb cysts from emphysema. Emphysema thresholds used to characterize CPFE on imaging vary in published studies. An ongoing international taskforce proposes to use a threshold of 5% emphysema extent for a research de nition of CPFE, and a threshold of 15% emphysema extent to de ne CPFE clinical syndrome. A minimal threshold extent of lung brosis on HRCT imaging is rarely used in CPFE, however interstitial lung abnormalities [165] (in the absence of formal ILD) associated with emphysema should not be described as CPFE.

Fig. 33.5  CPFE syndrome at chest HRCT with centrilobular emphysema and usual interstitial pneumonia pattern, progressive transition phenotype (male, smoker). (a) upper lobes showing centrilobular emphysema predominantly in anterior areas, with thickening of the interlobular septa; (b) mid regions of the lungs, showing admixture of

centrilobular emphysema and brosis; (c) lower zones showing usual interstitial pneumonia pattern with reticulation, traction bronchiectasis, some honeycombing, and superimposed ground-glass attenuation, predominating in the posterior areas

Fig. 33.6  CPFE syndrome at chest HRCT with predominantly paraseptal emphysema (male, smoker). (a) upper lobes showing predominantly paraseptal emphysema, with mild thickening of the interlobular

septa; (b) lower zones showing brotic changes with honeycombing, traction bronchiectasis, and reticulation, predominating in the right lower lobe