High-Resolution Computed Tomography of Interstitial and Occupational Lung Disease

Reticular Pattern

A reticular pattern on computed tomography (CT) almost always represents significant ILD. Morphologically, a reticular pattern may be caused by thickened intralobular or interlobular septa or honeycomb (fibrotic) destruction. Although intralobular and interlobular septal thickening often coexist, there is a morphological distinction: intralobular thickening is mostly below the spatial resolution of HRCT, thus appearing as faint ground-glass opacification rather than distinct fine reticulation. Intralobular septal thickening is seen in all ILDs, but it is most common in fibrosing lung disorders.

Numerous thickened interlobular septa indicate an extensive interstitial abnormality; causes include infiltration by fluid (e.g. pulmonary oedema) and/or abnormal cells (e.g. lymphangitis carcinomatosa). Although thickened interlobular septa can be a consequence of infiltration by fibrosis, this feature is not a frequent finding in fibrosing lung disorders (e.g. idiopathic pulmonary fibrosis [IPF]). Interlobular septal thickening is usually described as smooth (seen in pulmonary oedema and alveolar proteinosis) or irregular/beaded (e.g. lymphangitic spread of tumour, and seldom sarcoidosis), but the distinction is not always easily made. In some diseases, a perilobular distribution may give the spurious impression of thickening of the interlobular septa. However, such a pattern reflects a pathological process that is ‘smeared’ around the internal lobular surface and is most frequently associated with organising pneumonia.

Severe pulmonary fibrosis usually results in a coarse reticular pattern made up of interlacing irregular linear opacities. The reticular pattern of end-stage fibrotic lung is characterised by cystic air spaces surrounded by irregular walls, also known as honeycombing. The certain identification of honeycombing, as opposed to other forms of reticulation, is not always straightforward but is of particular relevance, as it may influence patient care. Honeycomb is a distinctive feature for the radiological definition of usual interstitial pneumonitis (UIP). The extensive fibrosis accompanying honeycomb is oftentimes associated with distortion of normal lung morphology, resulting in irregular dilatation of segmental and subsegmental airways (traction bronchiectasis/bronchiolectasis); in the periphery of the lung, it can be difficult to distinguish dilated airways from true honeycomb destruction.

Nodular Pattern

A nodular pattern is a feature of both interstitial and airspace diseases. The distribution and density of nodules may help refine what can be a lengthy differential diagnosis. Nodules within the lung interstitium, especially those related to the lymphatic vessels, are seen in the interlobular septa and in the subpleural and peribronchovascular regions; this distribution is seen most frequently in sarcoidosis but also in lymphangitis carcinomatosa. Centrilobular nodules are seen in several acute or chronic conditions affecting terminal bronchioles, including infection, hypersensitivity, or cigarette smoke ( Table 9.1 ). The distinction between subacute hypersensitivity pneumonitis (HP) and respiratory bronchiolitis–interstitial lung disease (RB–ILD) can be difficult, because both cause relatively low-density, poorly defined centrilobular nodules which may look identical on HRCT. A random distribution of very small well-defined nodules is seen in patients with haematogenous spread of tuberculosis, pulmonary metastases, pneumoconiosis and, rarely, sarcoidosis.

Mosaic Attenuation Pattern

The term ‘mosaic attenuation pattern’, or more simply mosaic pattern, refers to regional attenuation differences demonstrated on HRCT with well-defined borders corresponding to interlobular septa. The attenuation of a given area of lung depends on the amount of blood, parenchymal tissue, and air in that area, and thus the sign of a mosaic attenuation pattern is non-specific. It is the dominant abnormality in three completely different types of diffuse pulmonary disease: small airways disease, chronic occlusive vascular disease and infiltrative lung disease. In the first two processes, the decreased attenuation (‘black’) lung is abnormal; in infiltrative lung disease it is the ‘grey’ lung with increased attenuation that is abnormal. In patients in whom a mosaic attenuation pattern is the dominant abnormality, small airways disease and infiltrative lung disease are usually correctly identified but the mosaic attenuation pattern caused by occlusive vascular disease can be misinterpreted. Bronchial abnormalities and, to a lesser extent, the presence of air trapping on expiratory CT are the most useful discriminatory features in identifying small airways disease as the cause of mosaic attenuation. In chronic thromboembolic disease the mosaic pattern is caused by regional perfusion inhomogeneity. The phenomenon of hypoxic bronchodilatation causes vasoconstriction in the hypoperfused areas. Normally, expiratory CT scans help to exclude air trapping as the underlying cause. In addition, signs of pulmonary hypertension (enlarged pulmonary trunk and pathological arterio-bronchial ratio) point towards chronic thromboembolic pulmonary hypertension (CTEPH). However, in a subset of patients, coexistence of air trapping and vasoconstriction may be seen and can complicate interpretation.

Infiltrative lung disease manifesting as ground-glass opacification is the most frequent cause of the mosaic attenuation pattern. Borders between regional attenuation differences may be less well defined in infiltrative lung disease. Typically, no vascular calibre differences are present. A ground-glass pattern on HRCT is defined as a generalised increase in opacity that does not obscure pulmonary vessels and bronchial walls. At a microscopic level, the changes responsible for ground-glass opacity are complex and include partial filling of the air spaces, considerable thickening of the interstitium, or a combination of the two. Ultimately, though, the pattern of ground-glass opacity on HRCT results from subtotal displacement of air from the lungs. Indeed, ground-glass opacity may be found in many conditions, either as the predominant pattern or as an ancillary one. A predominant ground-glass pattern may be seen in subacute HP, acute respiratory distress syndrome (ARDS), pulmonary oedema, acute interstitial pneumonia (AIP), non-specific interstitial pneumonia (NSIP), and some infections, notably viral pneumonias and Pneumocystis jiroveci .

Cystic Pattern

A cyst appears as a round parenchymal area of low attenuation with well-defined interface against the normal lung. The cyst wall is usually thin (<2 mm); epithelial or fibrous wall of variable thickness are found at pathology. Cystic abnormalities are depicted both by radiography and HRCT, the latter being significantly more sensitive. They usually occur independent from pulmonary emphysema, which is one among several differential diagnoses. On HRCT, bronchiectasis is a pitfall for overcalling cysts; volumetric acquisition with multiplanar reformation clearly differentiate bronchiectasis from cysts. Cysts in the lung usually contain air but occasionally can be filled by fluid or solid material. A cavitated lesion might resemble a simple cyst, but an exceedingly thick wall should suggest a differential diagnosis (metastasis, septic emboli, pneumatocele after Staphylococcus aureus pneumonia, etc.). Lymphangioleiomyomatosis (LAM) and Langerhans cell histiocytosis (LCH) are the classical diseases with a cystic pattern. Nevertheless cysts are seen also in Pneumocystis jiroveci pneumonia (PJP), lymphoid interstitial pneumonia (LIP), Birth–Hogg–Dubé syndrome, HP and even in the normal ageing lung.

Usual Interstitial Pneumonia/Idiopathic Pulmonary Fibrosis

The term idiopathic pulmonary fibrosis is applied to patients with a histological and/or CT pattern of usual interstitial pneumonia (UIP) and compatible clinical and imaging features. However, the UIP pattern may be also secondary to other conditions such as chronic HP, asbestosis, connective tissue disease (CTD; especially rheumatoid arthritis [RA]), and rarely, drugs. Thus the interpretation of the UIP pattern requires a multidisciplinary discussion with clinicians and, sometimes, pathologists in order to obtain the final diagnosis.

The pathological features of UIP are the presence of fibroblastic foci, normal areas, dense fibrosis and honeycombing; another crucial finding is areas of fibrosis at different stages of maturity, the so-called temporal heterogeneity. The temporal heterogeneity of UIP–IPF is macroscopically depicted by CT, whilst it can be a source of low accuracy at pathology. This makes CT a pivotal player for planning pulmonary sampling.

After exclusion of an underlying (systemic) disease, HRCT plays a central diagnostic role for the diagnosis of UIP in the context of IPF. Guidelines, most recently updated in 2018, have defined various levels of radiological confidence.

The ‘typical UIP’ pattern on HRCT is characteristic and virtually pathognomonic; it appears as a predominantly subpleural bibasal reticular pattern within which there are areas of honeycomb destruction. As the disease progresses, it often appears to ‘creep’ around the periphery of the lung to involve the anterior aspects of the upper lobes (also known as ‘propeller blade distribution’) ( Fig. 9.1 ). The presence of ground-glass opacification is not a dominant feature and, when present, there is usually obvious traction bronchiectasis and bronchiolectasis that indicate fibrotic irreversible substrate.

A ‘probable UIP’ pattern on HRCT is characterised by predominantly subpleural bibasal reticular pattern with peripheral traction bronchiectasis or bronchiolectasis, without honeycombing ( Fig. 9.2 ). The ‘probable UIP’ pattern on HRCT reflects pathological probable or definite UIP pattern in 82%–94% of cases. The estimate of the clinical probability of IPF (e.g. age above 60 years, smoking history and no history of other potential causes of fibrosis) must be integrated to assess the likelihood of UIP in these patients.

As opposed to the two aforementioned patterns associated with IPF, the 2018 white paper from the Fleischner Society and the most recent ATS/ERS/JRS/ALAT Clinical Practice Guideline 2018 define two other HRCT patterns that demand pathological confirmation or consideration of an alternative diagnosis: namely, the ‘CT pattern indeterminate for UIP’ ( Fig. 9.3 ) and the CT features most consistent with an ‘alternative diagnosis’ ( Table 9.3 ). In an indeterminate pattern for UIP, no features ‘inconsistent with UIP’ are seen; however, distribution of fibrosis may be very heterogeneous or asymmetric and various amounts of ground glass can be seen. Features inconsistent with UIP/IPF include nodules, cysts, consolidations, air and a bronchocentric distribution. In patients whose HRCT does not demonstrate either a typical or a probable UIP pattern, a surgical lung biopsy may still demonstrate UIP pattern on histopathology. Most recent ATS/ERS/JRS/ALAT guidelines suggest the option for biopsy for all subclasses except for the ‘typical UIP’, though the recommendation is not strong and leaves room for individual decision.

Mediastinal lymphadenopathy (up to approximately 2.5 cm in diameter), unrelated to infection or malignancy, is a frequent finding in UIP–IPF.

The rapid development of a diffuse increase in the attenuation of lung parenchyma in patients with IPF should suggest the possibility of an accelerated phase (also known as acute exacerbation) of the disease ( Fig. 9.4 ), concurrent pulmonary oedema, or an atypical infection. Other complications include lung cancer and pulmonary tuberculosis ( Fig. 9.5 ); the latter usually has atypical appearances on CT caused by the presence of underlying lung fibrosis.

Non-Specific Interstitial Pneumonia

NSIP is characterised by varying degrees of interstitial inflammation and fibrosis without the specific features that allow a diagnosis of UIP. While NSIP may have significant fibrosis, it is usually temporally uniform (in comparison with UIP), and fibroblastic foci and honeycombing are absent or scanty. Although the clinical features of idiopathic NSIP resemble those of UIP, prognosis is considerably better. Non-idiopathic NSIP is most often found on lung biopsy in patients with CTD and may be the predominant histopathological pattern in some cases of drug-induced lung disease and chronic HP. On HRCT, ground glass with or without associated distortion of airways is usually the dominant pattern ( Fig. 9.6 ). Reticular abnormalities are common, but honeycombing is sparse or absent even when other signs of fibrotic changes are evident. Abnormalities are usually peribronchovascular or peripheral, although they may sometimes spare the subpleural lung. In general, NSIP may be distinguished from UIP on CT by a more prominent component of ground-glass attenuation and a finer reticular pattern in the absence of honeycombing. However, the variability of CT appearances reflects the heterogeneity of the pathological processes encompassed by NSIP and a confident diagnosis of NSIP based on CT alone is less readily made than in cases of UIP. Consolidation is reportedly a highly variable feature (0%–98%) and this discrepancy probably reflects the fact that some patients with non-idiopathic NSIP have significant amounts of histological organising pneumonia, making classification of individual cases difficult.

Cryptogenic Organising Pneumonia

This is considered in the section on airspace disease.

Respiratory Bronchiolitis–Interstitial Lung Disease and Desquamative Interstitial Pneumonia

These two entities are considered together because of their strong association with cigarette smoking . All cigarette smokers have, to some degree, inflammation around their small airways (‘respiratory bronchiolitis’), but this is clinically unimportant and not considered further here . Patients with RB–ILD generally present with an insidious onset of dyspnoea and cough . Chest radiography is relatively insensitive for the detection of RB–ILD and desquamative interstitial pneumonia (DIP) and a normal chest x-ray (CXR) has been reported in up to 20% of patients with RB–ILD and 25% in DIP .

On HRCT, the features of RB–ILD include areas of patchy ground-glass opacification (resulting from macrophage accumulation within alveolar spaces and alveolar ducts) and poorly defined low-attenuation centrilobular nodules ( Fig. 9.7 ). In addition, upper lobe centrilobular emphysema, usually of very limited extent, and areas of air trapping may be present, the latter reflecting the bronchiolitic element of this entity. In RB-ILD, scattered thickening of the interlobular septa and features of interstitial fibrosis can be seen, but this is not the dominant pattern.

Ground-glass opacification is the dominant feature seen in DIP ( Fig. 9.8 ). The distribution is typically lower zone, peripheral, and may be patchy or geographic. In some patients there are HRCT features of established fibrosis (in the form of architectural distortion with dilatation of some bronchi), usually of limited extent. Most patients with DIP or RB–ILD have a relatively stable clinical course and the two entities variably coexist. Smoking cessation is remarkable in the management of patients, but the influence of smoking on the clinical course of these patients has not been fully delineated; some patients have persistent abnormalities on HRCT even with smoking cessation and corticosteroid therapy. Because of the significant overlap between the clinical, imaging and histological features of DIP and RB–ILD and to a lesser extent between these two patterns and LCH and interstitial fibrosis, the global term ‘smoking-related interstitial lung disease’ (SR-ILD) has been proposed to encompass DIP, RB–ILD, LCH and interstitial fibrosis ( Fig. 9.9 ).

Acute Interstitial Pneumonia/Diffuse Alveolar Damage

AIP can be regarded as an idiopathic form of ARDS and is histologically (and clinically) distinct from the other interstitial pneumonias. The histological pattern seen in AIP is that of diffuse alveolar damage (DAD), which is also found in infection, CTD, drug toxicity and toxic fume inhalation. DAD has an acute exudative phase and a subsequent organising and fibrotic phase. Lung biopsy shows diffuse involvement with temporal homogeneity, which implies lung injury due to a single event. The chest radiograph shows bilateral patchy airspace opacification. HRCT demonstrates a combination of ground-glass opacification, consolidation, bronchial dilatation and architectural distortion ( Fig. 9.10 ). Ground-glass opacification on HRCT is found in all three phases of AIP, but coexistent traction bronchiectasis probably reflects the early incorporation of established fibrosis. Follow-up CT shows reticular opacities consistent with residual fibrosis. Anterior non-dependent fibrotic damage in survivors secondary to barotrauma has also been reported.

Lymphoid Interstitial Pneumonia

The term ‘lymphoid interstitial pneumonia’ was proposed by Liebow and Carrington to describe a disease entity characterised by a widespread interstitial lymphoid infiltrate of the lung, resembling lymphoma but with a clinical course more akin to a chronic interstitial pneumonia. Although in the past LIP has been considered by some to be a pulmonary lymphoproliferative disorder, evolution to frank lymphoproliferative disease is rare and thus LIP remains within the group of interstitial pneumonias. Classically, LIP occurs in association with autoimmune diseases, most often Sjögren syndrome (SjS). Other diseases associated with LIP include dysproteinaemias, autologous bone marrow transplantation and viral, mycobacterial and human immunodeficiency virus (HIV) infections. LIP is approximately twice as frequent in women and symptoms of progressive cough and dyspnoea usually predominate. Common HRCT findings are nodules of varying sizes (which may be ill-defined), areas of ground-glass opacification, thickened bronchovascular bundles, interlobular septal thickening and thin-walled cysts (1 to 30 mm) ( Fig. 9.11 ). Airspace disease, large nodules, and pleural effusions are rare in these patients. The cysts in LIP are usually discrete, sometimes clustered and tend not to be subpleural.

Idiopathic Pleuroparenchymal Fibroelastosis

Little is known regarding aetiology of IPPFE, but recurrent infections are a feature in some patients and a few cases have been reported in association with previous bone marrow or lung transplantation. IPPFE is characterised by dense established intra-alveolar fibrosis containing prominent elastosis, and dense fibrous thickening of the visceral pleura; these changes have a striking upper zone predominance. Such apical caps tend to continue leaning against the posterior visceral pleura. HRCT appearances consist of irregular pleural thickening and ‘tags’ in the upper zones that merge with fibrotic changes in the subjacent lung ( Fig. 9.12 ; Table 9.4 ). In 2012, Reddy et al. suggested radiological criteria for ‘definite’ (pleural thickening, almost exclusively confined to the upper lobes), ‘consistent with’ (not necessarily requiring upper lobe predominance) and ‘inconsistent with’ (none of the above features). Such classification was intended for integration with histopathology. Unfortunately, surgical biopsy is rarely granted because persistent postsurgical pneumothorax is a common complication. Restrictive functional impairment is seen in IPPFE as a fibrotic ILD per se; moreover, other ILD patterns might associate: UIP pattern was reported in about 5% to 10% of cases.

Lymphadenopathy

Sarcoidosis is characterised by bilateral, symmetrical hilar and paratracheal lymphadenopathy. Some degree of lymphadenopathy is evident on chest radiography in about 70%–80% of patients at some time during the course of the condition. Hilar lymph node enlargement ranges from the barely detectable to the massive and gives the hila a dense lobulated and usually well-demarcated outline. Occasionally hilar lymphadenopathy appears to be asymmetrical or, in 1%–5% of cases, may even be strictly unilateral although this is distinctly unusual.

Paratracheal lymphadenopathy may be bilateral or unilateral and in the latter instance is usually right-sided. The most common manifestation of left-sided lymphadenopathy is enlargement of the aortopulmonary window nodes—a common and characteristic feature on the chest radiograph. Other mediastinal nodes (anterior prevascular, posterior and subcarinal) are often not identified as being enlarged on the chest radiography but on CT are seen to be affected in about half of patients. In 90% of patients with lymphadenopathy, nodal enlargement is maximal on the first radiograph and usually disappears within 6–12 months. Recurrence of lymphadenopathy is exceedingly rare.

The affected lymph nodes may calcify, sometimes in a characteristic eggshell fashion. This latter feature is shared by only a few conditions, such as silicosis and histoplasmosis. The calcification is of variable intensity and may even be relatively light with homogeneous representation over the whole lymph node, into the so-called ‘icing sugar’ pattern. Such a pattern of calcification is common in relatively small lymph nodes with even distribution throughout the mediastinum and hila (very different from calcified nodes due to tuberculous infection, which usually follow a drainage path). About 40% of patients presenting with nodal enlargement will develop parenchymal opacities, usually within a year, and of these about one-third will go on to have persistent (fibrotic) changes. Usually, nodal enlargement does not develop after parenchymal opacities have appeared.