UIP is the most common histopathologic pattern seen in patients who present with the clinical syndrome of IIP.3.^{12.13.14.15. and 41.} Under the ATS/ERS classification, the term idiopathic pulmonary fibrosis is applied exclusively to the idiopathic clinical syndrome associated with the morphologic pattern of UIP.^{4. and 18.} The key histopathologic features of UIP (Box 10.4) are: ⁴²

The areas of young connective tissue seen on biopsy in UIP consist of small foci of actively proliferating fibroblasts and myofibroblasts.43. ^{and 44.} These fibroblastic foci, representing ongoing sites of lung injury and repair, are thought to represent the earliest lesion of UIP, and are very important for the histologic diagnosis of UIP. ⁴ They may represent areas of lung response to an injury that is as yet undefined.^{45. and 46.} They often occur at the interface between fibrotic and normal lung. The extent of fibroblastic foci on lung biopsy is an important predictor of survival in patients with UIP. ⁴⁷

The recognition of fibroblastic foci as the important pathogenic lesions of UIP represents an important paradigm shift from the previous concept of alveolar inflammation or alveolitis as the precursor of UIP. ⁴¹ A chronic inflammatory cell infiltrate may indeed be seen in UIP, but is usually mild and consists mainly of lymphocytes with occasional plasma cells and eosinophils. If extensive cellularity is found, the diagnosis is probably not UIP. The extent of alveolar cellularity does not predict survival. ⁴⁷ The paucity of inflammatory infiltration of the lung, in contrast to the other idiopathic pneumonias, may explain the lack of steroid responsiveness in UIP/IPF.

Causes of a UIP pattern on histology and CT (Box 10.5) include chronic hypersensitivity pneumonitis, collagen vascular disease, familial pulmonary fibrosis, and asbestos exposure.^{4. and 48.} Drug toxicity is a rare cause of a UIP pattern, best described with nitrofurantoin. ⁴⁹

The pathogenesis of UIP remains unclear. The fibroblastic foci are thought to represent an abnormal wound healing response to lung injury, resulting from release of multiple fibrogenic cytokines. ⁴¹ The inciting agent remains unclear. There has been much speculation about the possible role of viruses, including the Epstein–Barr virus, hepatitis C virus, herpesvirus, and adenoviruses, in the initiation of IPF, but no specific relationship has been proved.^{50.51.52.53. and 54.}

UIP is more common in men than in women (1.5–1.7 : 1).^{21.38. and 55.} Most patients are over 50 at the time of diagnosis, and about two-thirds are over 60, ⁵⁵ but the disease may occur in people as young as 28.^{3. and 56.} There have been descriptions of UIP occurring in children and adolescents, but these entities would probably not meet the current criteria for diagnosis of UIP, ³ and no cases of UIP were recorded in two series of 51 and 38 pediatric patients with diffuse lung disease that used current diagnostic criteria.^{57. and 58.}

Patients with UIP usually complain of a dry cough and progressive exertional dyspnea of insidious onset. Systemic symptoms such as fever, weight loss, and arthralgia occur with variable frequency, ²⁹ although fever itself is unusual and probably occurs in less than 15% of patients.^{43. and 59.} Some patients date the onset of symptoms to an influenzalike illness. ²⁸ A small proportion of patients present without symptoms, because of an abnormal radiograph or CT scan. ⁶⁰ On physical examination, late, fine inspiratory crackles at the lung bases are an almost universal finding, ²⁶ and two-thirds to three-fourths of the patients show clubbing of the fingers.^{26.27.60. and 61.} Occasionally, there is full-blown hypertrophic osteoarthropathy. ²⁷ In advanced disease cor pulmonale and cyanosis may develop. ²⁸

The prognosis of patients with biopsy-proven UIP is poor, with mean survival after presentation ranging from 2.8 years to 9 years.^{12.26.62. and 63.} At 5 years after diagnosis, only 12%¹⁵ to 45% of patients with UIP survive.^{12.14. and 64.} Lower mortality rates were reported in earlier series,^{65.66. and 67.} but it is likely that these included other IIP subtypes such as NSIP. The survival of patients with UIP is clearly worse in recent series which have excluded patients with NSIP and DIP.^{14.15.16. and 68.}

Respiratory failure is the cause of death in 50–90% of patients with IPF.^{69.70. and 71.} The majority of the respiratory deaths are related to acute exacerbation of UIP.^{69.70. and 72.} The second most common cause of death is cardiovascular disease, while lung cancer accounts for about 10% of deaths.^{71. and 72.} Given the high mortality, there is substantial interest in developing a treatment for IPF. There is no evidence that immunosuppressive therapy or corticosteroids are effective in UIP, although these agents are still commonly used. ⁷³N-Acetylcysteine may help prevent decline in pulmonary function. ⁷⁴ Bosentan, an endothelin antagonist, showed a trend toward delayed time to death or disease progression: ⁷⁵ this effect was more pronounced in those who had undergone biopsy, and particularly in patients who did not have honeycombing on CT at presentation. ⁷⁶ Other potentially helpful agents currently undergoing evaluation include sildenafil⁷⁷ and perfenidone. ⁷⁸

Almost all patients with UIP pattern due to IPF have abnormal chest radiographs when they present with symptoms. Indeed, when prior chest radiographs are available, basal reticular opacities are usually visible in retrospect for several years prior to the development of symptoms (Fig. 10.1). These presymptomatic opacities are commonly reported by the radiologist but, in the absence of symptoms or other abnormalities, disregarded by the clinician. Since early detection of IPF may be of value in treatment, the radiologist should emphasize the importance of these abnormalities, and ensure that they are investigated by careful review of the patient’s symptoms and signs, and, if appropriate, physiologic and HRCT evaluation. The classical chest radiographic appearance in patients with UIP is of basal reticulonodular pattern which becomes a coarser reticular, or honeycomb, pattern as the disease progresses (Fig. 10.2). The characteristic subpleural distribution of the abnormality may often be recognized in the mid- and upper lungs, along the lateral chest wall. The lung volumes usually decrease with time, particularly in the lower lobes, but in patients with associated emphysema, the radiographic lung volumes may be normal or even increased (Fig. 10.3).

Pneumomediastinum (Fig. 10.2) is quite common in patients with UIP, and is often asymptomatic. Pneumothorax may be seen; while this may be small and asymptomatic, it may also be associated with a significant air leak if there is rupture of a honeycomb cyst. In patients with advanced UIP who develop pneumothorax, the stiffness of the lung often prevents marked lung collapse.

In about 50–70% of cases, UIP has a very characteristic appearance on CT. ⁷⁹ The predominant CT pattern is usually either reticular, honeycombing, or a combination of both (Figs 10.3 and 10.4). ²⁵ Traction bronchiectasis and/or bronchiolectasis is often seen. Although ground-glass abnormality is commonly present,^{25.80. and 81.} it is usually associated with reticular abnormality and traction bronchiectasis, suggesting that it represents lung fibrosis rather than inflammation. Isolated ground-glass attenuation, if present, is usually sparse. Consolidation is uncommon in studies from North America and Europe, but appears to be more common in Asian series.^{79.80. and 81.} The abnormalities are basal predominant in most, but may be diffuse. Peripheral, subpleural predominance is present in over 90%. The distribution of fibrosis may be asymmetric in up to 25% of cases, and asymmetry may be useful in distinguishing UIP from other fibrotic interstitial pneumonias. ⁸⁰ In contrast to the homogeneous appearance of NSIP, the abnormalities of UIP often have a patchy distribution. As the disease progresses, it often appears to ‘creep’ up the periphery of the lung, causing subpleural reticular abnormality in the upper lungs (Fig. 10.4). In a study by Hunninghake et al., ⁸² this finding of subpleural lines in the upper lungs was useful in discriminating between UIP and other conditions.

As with many other ILDs, mediastinal adenopathy may be seen on CT in UIP (Fig. 10.5). ⁸³ The nodal enlargement typically involves only one or two nodal stations and the nodes usually measure less than 15 mm. The adenopathy is thought to be reactive. ⁸⁴ It is never visible on the chest radiograph. The presence of lymph node enlargement does not appear to correlate with the extent of lung fibrosis, ⁸³ but the number of enlarged nodes may relate to disease extent. ⁸⁵ Pulmonary ossification, characterized by areas of punctate or linear calcification, may occur in areas of advanced fibrosis (Fig. 10.6), and will be better seen if the CT images are viewed at mediastinal windows. ⁸⁶ Several other examples of the typical CT findings of UIP may be found in the section on collagen vascular diseases (see Figs 10.29, 10.31, 10.33 and 10.42 below).

Numerous prospective and retrospective studies have shown that a confident or highly confident diagnosis of UIP, based on the CT features outlined above, has a specificity of over 90% for the pathologic diagnosis of UIP.^{25.87.88.89.90. and 91.} Honeycombing in the lower lobes, and linear abnormality in the upper lobes, are the most reliable features for differentiating between UIP and its clinical mimics (Fig. 10.4). ⁸² In a study by Flaherty et al., ¹⁶ the observation of honeycombing on HRCT indicated the presence of UIP with a sensitivity of 90% and specificity of 86%.

Because of the high degree of accuracy of HRCT in many cases of UIP, the diagnosis of UIP is commonly made based on clinical and imaging features, without the need for surgical biopsy. The ATS has published criteria for diagnosis of UIP in the absence of a surgical biopsy (Box 10.6). ¹⁸ These criteria indicate that ‘bibasilar reticular abnormalities with minimal ground glass opacities on HRCT scans’ are a major criterion for diagnosis of IPF. However, since publication of these criteria, the importance of honeycombing in the diagnosis of IPF has become more apparent,16. ^{and 82.} and a confident CT diagnosis of UIP is not usually made unless honeycombing is present. However, there is a substantial minority (30–50%) of cases of histologic UIP in whom a confident diagnosis of UIP cannot be made based on the CT appearances (Fig. 10.7).^{89.91.92. and 93.} In these patients, the diagnosis of UIP can be made only by lung biopsy. A study by Flaherty et al. ⁶⁸ suggested that the patients with histologic UIP who had definite or probable UIP by HRCT criteria had a shorter survival than those who had indeterminate HRCT findings. This is most likely because the typical HRCT criteria for diagnosis of UIP include the presence of honeycombing, and may therefore select patients with more advanced or more severe disease. This study reemphasizes the importance of seeking lung biopsy in patients in whom CT is not diagnostic of UIP. An equivocal, atypical, or nonspecific CT appearance should always prompt biopsy. Most importantly, the CT features must be interpreted in conjunction with a complete clinical evaluation.

Initial reports of the CT features of IPF or fibrosing alveolitis suggested that the finding of ground-glass attenuation was associated with ‘alveolitis’ on biopsy, ⁹⁴ which were potentially reversible on steroids and were associated with a favorable prognosis.^{22. and 95.} However, the patients in those studies who had a large amount of ground-glass abnormality probably had NSIP or DIP rather than UIP. HRCT studies of ‘IPF’, ‘fibrosing alveolitis’, or ‘UIP’ published before 1995 are very difficult to interpret, because it is unclear how many cases of NSIP or DIP were included in these papers. As discussed above, under the current definition of UIP, alveolitis is not regarded as an important histologic component of the disease, and therefore ground-glass attenuation in UIP cannot be regarded as reflecting alveolitis. In contrast to DIP⁹⁶ and NSIP, ⁹⁷ true UIP shows inexorable progression and rarely if ever reverses in response to steroid treatment (Fig. 10.8). ⁹⁶ When extensive ground-glass attenuation is present in patients with known UIP, the possibility of an acute exacerbation or accelerated deterioration should be considered.

Several studies have evaluated sequential changes in IPF.^{24.81.98.99.100. and 101.} Comparison of serial scans in patients with IPF depends on identifying comparable scan levels, by matching anatomic features such as vessels and bronchi. In untreated patients scanned at intervals of greater than 6 months, the extent of IPF increases progressively (Fig. 10.8). In patients who are treated, some areas of ground-glass attenuation progress to reticular abnormality and honeycombing, but some areas regress. Reticular abnormality usually progresses to honeycombing. ⁹⁹ Areas of honeycombing increase inexorably in extent, ⁸¹ and the size of the honeycomb cysts also increases. ⁹⁸ The presence or absence of honeycombing on baseline CT may impact disease progression. In a study by Jeong et al., ¹⁰¹ extent of disease decreased in five of 29 patients with UIP who did not have honeycombing on baseline CT, but in none of 22 with honeycombing. However, progression occurred in six of those without honeycombing, and in only two of those without honeycombing. The appropriate interval for detecting change in patients with IPF would appear to be about a year, though some patients may show clear evidence of change in 3–4 months. Of course, sequential scanning is indicated only if it will change management.

Positron emission tomography (PET) scans performed in individuals with lung fibrosis usually show increased metabolic activity in areas of lung fibrosis (Fig. 10.9); increased uptake of [¹⁸ F]2-fluoro-2-deoxy-D-glucose (FDG) was found in the fibrotic lungs in six of seven male patients with IPF. ¹⁰² One patient with end-stage lung fibrosis did not have increased uptake. In one of these patients the degree of uptake decreased after steroid treatment (when dyspnea decreased), and uptake increased again when steroids were tapered and symptoms increased. In a different study of 21 patients with a variety of infiltrative lung diseases, activity on FDG-PET did not help discriminate between IPF and non-IPF conditions. ¹⁰³ There has also been interest in the use of ¹⁸F-proline as an agent to monitor fibrotic activity in the lung, but a recent study of seven patients with fibrotic lung disease studied with this agent showed low uptake in all patients with IPF. ¹⁰⁴

Important complications of IPF include infection, lung cancer, and accelerated deterioration (Box 10.7). ⁷¹ A variety of opportunistic infectious organisms may occur in treated patients with IPF, including Pneumocystis jirovecii (Fig. 10.10), Mycobacterium avium complex (Fig. 10.11), and mycetoma due to Aspergillus species or other organisms (Fig. 10.12).

Lung cancer (Fig. 10.13) develops five times more frequently in patients with IPF than in the general population. ¹⁰⁵ Lung cancer appears to be primarily associated with UIP, rather than with the other IIPs. It occurs most commonly in male smokers. ¹⁰⁶ The prevalence of carcinoma in various series has ranged between 5% and 13%.^{65.71.107. and 108.} Synchronous multiple lung cancers occur in up to 15% of patients with fibrosing alveolitis. ¹⁰⁹ Adenocarcinoma and squamous carcinoma are the two most common cell types, with approximately equal prevalence in most series, and smaller numbers of patients with bronchioloalveolar cell carcinoma and small cell cancer.^{106. and 108.} Lee et al. ¹⁰⁸ found lung cancer in 32 (13%) of 244 cases of IPF seen over a 7.5-year period. The majority (55–65%) of these lung cancers occur peripherally, usually adjacent to honeycombing (Fig. 10.13), and most of the remainder are found in the upper lobes^{106. and 108.} (Box 10.8).^{108. and 109.} The cancers may appear as areas of poorly defined consolidation, ¹⁰⁸ but are more commonly lobulated nodules. ¹⁰⁶ The preferential occurrence of these lung cancers within areas of fibrosis may make them difficult to detect. Meticulous comparison of serial images is important, and evaluation for additional nodules is also important, since multiple synchronous cancers are relatively common. ¹⁰⁹ The presence of lung fibrosis does not preclude resection for lung cancer, but the mortality and morbidity are higher than in those without lung cancer.^{110. and 111.}

Between 10% and 25% of patients with IPF develop accelerated deterioration, or acute exacerbations, progressing to acute respiratory failure.70.^{112.113.114. and 115.} Acute exacerbation is more acute than accelerated deterioration, but the precise distinction between these entities is unclear, and the terms are sometimes used interchangeably. The lungs of most of these patients have a histologic appearance of diffuse alveolar damage, similar to that of acute interstitial pneumonitis, though organizing pneumonia may also be seen, and has a better prognosis. Acute exacerbation is one of the most common causes of death in patients with IPF. ⁶⁹

Accelerated deterioration or acute exacerbation of IPF presents with a relatively short onset of progressive dyspnea or cough, occasionally associated with systemic symptoms. There is usually a short prodrome of 4–8 weeks’ duration. Mortality in different series has varied from 20% to 100%, but exceeds 80% in most studies. ¹¹⁶ A recently published consensus definition for this entity requires the following: a previous or concurrent diagnosis of IPF, with unexplained worsening or development of dyspnea within 30 days, HRCT showing new bilateral ground-glass abnormality and/or consolidation superimposed on a background of usual interstitial pneumonia, and exclusion of other causes of decompensation including infection, left heart failure, and pulmonary embolism. ¹¹⁶ Precipitants for acute exacerbation may include thoracic biopsy or resection, chemotherapy, infection, and irradiation.^{117.118.119.120. and 121.} Even patients with subclinical lung fibrosis evident on CT may be at risk for acute respiratory distress following thoracic resection surgery. ¹²² Acute exacerbation of IPF may be more common in Asian individuals (Box 10.9). ¹²³

The chest radiograph usually shows ground-glass abnormality or consolidation superimposed on background reticular abnormality (Fig. 10.14). On CT, acute exacerbation is characterized by diffuse, multifocal, or peripheral ground-glass attenuation (Fig. 10.15), ¹¹² which must be distinguished clinically from opportunistic viral or pneumocystis infection. Consolidation may also be seen. In 17 patients with accelerated IPF, Akira et al. ¹¹² classified the CT distribution of ground-glass attenuation as peripheral (adjacent to preexisting fibrosis), diffuse, or multifocal. Peripheral opacities were more likely to improve with corticosteroid treatment. On biopsy, the peripheral pattern of abnormality correlated with the presence of active fibroblastic foci in two patients, while the multifocal and diffuse patterns correlated with acute diffuse alveolar damage superimposed on UIP. On sequential evaluation of patients with acute exacerbation or accelerated deterioration, ground-glass attenuation was seen to evolve into consolidation with architectural distortion, and cystic lesions often developed. ¹¹³ In a subsequent study from the same group, evaluating 58 patients with acute exacerbation, the diffuse pattern of parenchymal abnormality, found in 34 patients, was associated with significantly worse survival than the multifocal and peripheral patterns (median survival 16 days, compared with 240 and 540 days respectively). ¹²⁴

In addition to IPF, acute exacerbation may occur in other fibrotic lung diseases, including nonspecific interstitial pneumonia, lung disease associated with collagen vascular disease, and hypersensitivity pneumonitis.^{125.126. and 127.}

The chest radiograph12.^{129.136. and 137.} is abnormal in more than 90% of patients with NSIP, but changes are not well characterized in the literature. Abnormalities are usually bilateral with a basal predominance. They may be consolidative and patchy, reticulonodular, or mixed. Lower lobe volume loss is common, and traction bronchiectasis is frequently present (Fig. 10.16).

Published descriptions of the CT appearances of NSIP have varied quite widely,13.^{25.97.132.136.141.142.143. and 144.} probably because of differing pathologic diagnostic criteria. However, a multidisciplinary workshop identified a relatively typical CT appearance among 67 patients who received a clinical–imaging–pathologic consensus diagnosis of NSIP (Box 10.12). ¹²⁸

Ground-glass pattern is the most common CT abnormality in all series of patients with NSIP, being found in 90–100% of patients.81. ^{and 145.} It is usually confluent, bilateral, and symmetric, with basal predominance (Fig. 10.17). It is commonly peribronchovascular in predominant distribution, ¹⁴¹ sometimes with sparing of the subpleural lung. ¹²⁸ The peribronchovascular distribution and subpleural sparing may be helpful in distinguishing NSIP from UIP. Ground-glass opacity occurs as an isolated finding in about a third of patients, but is more commonly associated with other findings such as reticular abnormality or consolidation. Consolidation is seen in 0%¹³² to 98%¹⁴⁴ of patients, almost always combined with ground-glass opacity. The wide variation in prevalence of consolidation may be due to the fact that some patients with NSIP have considerable amounts of histologic organizing pneumonia, so that it can be difficult for the pathologist to decide whether to classify an individual case as NSIP or organizing pneumonia. ⁴ Criteria for distinguishing NSIP from organizing pneumonia remain poorly defined. Like ground-glass opacity, consolidation is bilateral, basal and subpleurally predominant or peribronchovascular. Irregular linear or reticular opacities (Fig. 10.17) occur in about 75% of cases (range 29–91%).^{81.136.141.142. and 145.} Traction bronchiectasis and bronchiolectasis is a characteristic feature, present in areas of ground-glass attenuation or reticular abnormality in 70–90% of cases. ¹⁴⁵ (For other examples of the typical CT findings of NSIP see the section on scleroderma [Figs 10.44 and 10.45].)

Honeycomb changes were recorded in 20–30% of patients in some series^{13.25.81.134.142. and 144.} but were absent in others.^{97.136. and 141.} This apparent discrepancy may be accounted for by interobserver variability in diagnosis of NSIP and UIP, and by the fact that histologic changes of UIP may coexist with those of NSIP. In general, honeycombing, if present in NSIP, is sparse. However when patients with NSIP are followed over time, honeycombing may increase in prevalence and extent. ⁸¹ Mildly enlarged mediastinal nodes are present in up to 80% of cases. ⁸³ Pathologic/HRCT correlation¹⁴¹ has shown that ground-glass opacity is caused by cellular or fibrotic interstitial thickening; bronchial dilatation and irregular linear opacity indicates predominant fibrosis, and some areas of consolidation are caused by areas of organizing pneumonia.

Since the prognosis of cellular NSIP is better than that of the fibrotic type, it would be useful if CT were able to reliably discriminate between these entities. However, MacDonald et al. ¹³² suggested that there was substantial overlap between the CT findings of these subtypes, although cellular NSIP was less likely to be subpleural in distribution, and had a higher proportion of ground-glass type abnormality. Johkoh et al. ¹⁴⁴ found that the extent of reticular abnormality and traction bronchiectasis correlated with increasing amounts of fibrosis on histologic grading. Honeycombing was seen mainly in patients with purely fibrotic NSIP. While traction bronchiectasis was present in almost all cases, it tended to involve more proximal airways (lobar and segmental bronchi) in patients with fibrotic NSIP. Patients with cellular NSIP were more likely to have pure ground-glass abnormality without reticular abnormality. However analysis of this issue is difficult because the proportion of patients with purely cellular NSIP is relatively small (six of 55 patients in the paper by Johkoh et al., ¹⁴⁴ and five of 21 patients in the paper by MacDonald et al. ¹³²).

Nishiyama et al. ⁹⁷ evaluated the CT changes on follow-up of patients with NSIP after treatment. Parenchymal abnormalities showed improvement or resolution in 12 of 14 patients. Reticular abnormality, traction bronchiectasis, and ground-glass abnormality were all completely or partially reversible. One patient progressed, with ground-glass abnormality being replaced by honeycombing. In a similar study by Kim et al., ¹⁴¹ the extent of ground-glass abnormality decreased substantially on follow-up, but the extent of reticular abnormality decreased only slightly. In a serial CT evaluation of 24 subjects with NSIP, extent of lung fibrosis increased in nine, decreased in three, and was unchanged in 12. ¹⁰¹ A more long-term serial study of 23 patients with NSIP, with median follow-up of 61 months, found that the extent of ground-glass abnormality decreased over time, but tended to be replaced by increased fibrotic abnormality, with increased prevalence and extent of honeycombing, and development of a pattern suggestive of UIP in five patients. ⁸¹

COP was recognized as a distinct clinicopathologic entity in the 1980s.7.^{8. and 146.} In the USA and Canada, it was called BOOP,^{8.147.148. and 149.} or proliferative bronchiolitis, ¹⁵⁰ because of the tendency of plugs of organizing pneumonia to occlude the smaller bronchioles. For this reason, it was often misleadingly classified as a small airways disease. However, its clinical, imaging, and pathologic features are distinct from those of small airways disease. There is some controversy about its inclusion as an IIP, ¹³¹ because the areas of organizing pneumonia involve the pulmonary airspaces rather than the interstitium, its clinical presentation is more subacute than UIP, DIP or NSIP, and its imaging features are those of airspace disease. However, it is included in the ATS/ERS classification of IIPs, ⁴ because of its idiopathic nature, its association with collagen vascular diseases, and its tendency to overlap with the other interstitial pneumonias, particularly NSIP.

COP is a rare condition, with a prevalence of 12 per 100 000 hospital admissions in a Canadian series, ¹⁵¹ and a reported incidence of 1.1/100 000 in Iceland. ¹⁵² Mean patient age is about 58 years (range 20–80 years),^{153. and 154.} and sex incidence is equal. ¹⁵¹ Twenty-five to fifty per cent of patients give a history of an influenzalike prodrome followed by an illness lasting between a week and several months, characterized by cough (persistent and commonly nonproductive), exertional dyspnea, malaise, fever, and weight loss. Less common complaints include pleuritic pain^{147. and 155.} and hemoptysis.^{155.156. and 157.} Given these symptoms, and radiographic findings of airspace opacity, it is not surprising that 50% of patients are initially diagnosed as having infective pneumonia. ¹⁵¹ On clinical examination of the lungs, fine, dry crepitations occur in 70–90% of patients; clubbing is rare. ¹⁵⁸ The erythrocyte sedimentation rate and C-reactive protein level are usually raised and may be very high.^{156. and 159.}

Pulmonary function tests show a restrictive abnormality, commonly with a diffusion defect and hypoxemia at rest or on exertion.^{159. and 160.} Occasionally there are obstructive features. ¹⁵¹ Bronchoalveolar lavage is nonspecific and shows mixed cellularity usually with predominant lymphocytes accompanied by neutrophils and eosinophils with or without foam cells, mast cells, and plasma cells. ¹⁶⁰ Neutrophil/eosinophil predominance indicates an unfavorable prognosis. ¹⁶¹ The histologic diagnosis may be made by transbronchial biopsy¹⁶² in up to two-thirds of patients^{151. and 163.} especially if a step sectioning technique is used. ¹⁵⁹ CT may be helpful in identifying a suitable site for transbronchial biopsy. Bronchoscopy may also be important to exclude other causes of consolidation such as infection, aspiration, or bronchioloalveolar cell carcinoma. In cases where there is diagnostic doubt, thoracoscopic biopsy will be necessary.

The imaging hallmark of COP is bilateral, patchy, airspace opacity (Fig. 10.18), with a density which ranges from consolidation to ground-glass opacity (Box 10.14). Airspace opacity is usually bilateral (70–90%) and asymmetric. In about 10% of cases, opacities are unilateral or focal.^{59.156.164. and 165.} Consolidation, seen in about 80% of cases, is nonsegmental and commonly 2–6 cm in diameter. It generally shows no craniocaudal predilection though some series have shown a basal predominance.^{8.158. and 166.} On CT, the consolidation often has a peribronchovascular² or subpleural distribution and the subpleural distribution may be apparent on the chest radiograph.^{151.155.162. and 164.} The consolidation commonly contains air bronchograms (Fig. 10.19) with dilated airways. Consolidation tends to migrate (Box 10.15; Fig. 10.18) and to come and go even without treatment.^{7.155.159.166.167.168. and 169.} The consolidation may uncommonly cavitate.^{156.170. and 171.} A perilobular pattern (poorly defined opacity along interlobular septa) (Fig. 10.20) was found in over 50% of patients with COP in one study and may help suggest the diagnosis. ¹⁷² The reverse halo or atoll sign (ringlike opacity with central ground-glass abnormality), found in six of 31 subjects in one series, ¹⁷³ may also be helpful in suggesting the diagnosis, ¹⁷⁴ although this sign may also be encountered in other conditions such as vasculitis, pulmonary infarction, or fungal infection.

About 60% of patients¹⁷⁵ have ground-glass opacity on CT (Fig. 10.21), often with a lobular, mosaic distribution.^{176. and 177.} About 30–50% of patients demonstrate nodules either as an isolated finding or more commonly as part of a mixed pattern with consolidation (Fig. 10.22). Nodules show a great range in size but are most commonly in the 1–10 mm range, smooth, well defined and sometimes with a centrilobular or peribronchovascular distribution. ¹⁶⁴ They are usually bilateral and grossly distributed in a random fashion. In a small number of patients some nodules are larger (10–20 mm) and may contain air bronchograms or have ill-defined margins (‘airspace’ nodules). Other findings described in a minority of patients include: peripheral reticulation and irregular lines, particularly in the lower zones;^{147.151. and 178.} bronchial dilatation and wall thickening (Fig. 10.21); ¹⁶⁴ large nodular or masslike peripheral opacities; ¹⁴⁷ and satellite lesions; ¹⁷⁹ bandlike opacities with central air bronchograms; ¹⁷⁴ mild lymph node enlargement in 14–42% of patients;^{84.147. and 164.} and small pleural effusions, either bilateral or unilateral, in about 20–30% patients.^{151.164.180. and 181.} The finding of reticular abnormality on the chest radiograph or CT in a patient with biopsy-proven organizing pneumonia suggests that the patient will not respond to treatment, and may develop progressive fibrosis, likely due to evolution to NSIP.^{156. and 178.} Other examples of the typical CT findings of organizing pneumonia may be found in the section on collagen vascular diseases (see Figs 10.32 and 10.52 below, and Box 10.16). Lesions of organizing pneumonia may be hypermetabolic on PET imaging. ¹⁸² In a study of 22 patients with primary and secondary organizing pneumonia, consolidation was associated with mildly increased uptake of FDG on PET scanning (mean SUV 3.1), and the degree of activity correlated with cellular measures of disease activity on bronchoalveolar lavage. ¹⁸³

COP occasionally resolves spontaneously^{8.160. and 184.} but usually requires treatment with steroids. With therapy, clinical and radiographic signs clear in weeks in 30–60% of patients, but there is a substantial relapse rate (50% in two recent studies)^{185. and 186.} and up to one-third of patients are left with persistent radiographic abnormality. ¹⁵⁴ A minority of patients with COP progress, despite treatment, to lung fibrosis with honeycombing.^{156. and 170.} Progression is more common in those with longer duration of symptoms and those who have linear radiographic abnormalities on initial chest radiograph or reticular abnormality on CT.^{156. and 178.} It is possible that some of these patients may have NSIP or UIP, with a nonrepresentative biopsy sample showing predominant organizing pneumonia.

There is a 3–13% mortality in various series, many of which combine cases of primary and secondary OP^{43. and 161.} and generally the prognosis is less good in organizing pneumonia associated with connective tissue disorders.^{154.161. and 180.} A subgroup of patients with cryptogenic or secondary organizing pneumonia is described with a fulminant course and high mortality.^{187.188.189. and 190.} Histologic examination of biopsies in these cases has shown organizing pneumonia together with UIP and diffuse alveolar damage. ¹⁶¹ This entity most likely represents a form of acute exacerbation of diffuse lung disease.

Organizing pneumonia is commonly associated with a recognized precipitating cause and in the Mayo Clinic series of 74 patients, 36% were classified as secondary, compared with 50% cryptogenic cases. ¹⁸⁰ These authors found no distinguishing clinical, imaging, or pathologic features¹⁸⁰ between COP and secondary organizing pneumonia apart from a worse overall prognosis in the secondary group, as noted by others. ¹⁸⁷ The poorer prognosis includes a slower and less complete resolution and a higher respiratory mortality.

Associations and causes of secondary organizing pneumonia are listed in Box 10.16.^{4.151.160. and 161.} In clinical practice, the common causes of an organizing pneumonia pattern are collagen vascular disease (particularly polymyositis/dermatomyositis, and RA), drugs, and lung or bone marrow transplantation. One interesting clinical entity is the occurrence in women of migratory organizing pneumonia, within a year following radiation therapy for breast cancer (Fig. 10.23).^{169.191.192.193. and 194.} These patients may have typical symptoms of COP or may be asymptomatic, and show typical migratory pulmonary opacities, both within and outside the radiation port. They usually respond well to steroids, though with a high prevalence of relapse on withdrawal of steroids. It is thought that the radiation treatment in some way ‘primes’ the immune system of the lung for the development of organizing pneumonia. ¹⁹²

Focal organizing pneumonia, presenting with a single discrete focal nodule, mass, or consolidation on chest imaging, is an uncommon presentation of organizing pneumonia, usually not associated with the clinical syndrome of COP. ¹⁸⁰ Lung resection is commonly performed because of suspicion of malignancy. Maldonado et al. ²¹⁸described 26 cases, identified as solitary nodules (6–68 mm diameter). About 40% of patients had symptoms, most commonly cough. Four patients had symptoms suggestive of an antecedent respiratory infection, and a further three had infection identified at biopsy. Etiologies identified in other reports have included chemotherapy with bleomycin. ²¹⁹ Underlying COPD may be present. ²¹⁸ Kohno et al. ²²⁰ subdivided the CT appearances in 18 subjects into three types: small irregular masses often with pleural tags, all less than 2 cm diameter; larger oval lesions with broad pleural contact, some with satellite nodules, air bronchograms, and converging peripheral lung vessels; and oval lesions lying along bronchovascular bundles, again sometimes accompanied by pleural tags and satellites. Overall 94% of lesions had irregular margins and about 50% of lesions had air bronchograms, satellites, or pleural tags. Follow-up CT examinations showed partial or complete resolution, and the authors stressed the usefulness of such examinations in management. Cavitation may occur.^{170. and 221.} In the report by Maldonado et al., ²¹⁸ all four lesions studied by FDG-PET were hypermetabolic, and all seven studied with dynamic enhancement CT showed enhancement by more than 15 HU. Focal organizing pneumonia may also present with a ground-glass attenuation nodule, and, in a recent series, organizing pneumonia/NSIP accounted for 10 of 49 subjects with such nodules who underwent surgical biopsy. ²²²

As a result of the marked differences in prognosis and in management, the most important diagnostic distinction is between UIP and other entities. Table 10.2 indicates the CT features which help make this distinction. In a study of 91 patients presenting with a clinical syndrome suggestive of IPF, 60% had UIP on biopsy, 13% had respiratory bronchiolitis, 8% had hypersensitivity pneumonitis, 7% had NSIP, 3% had sarcoidosis, while the remaining patients had a variety of diagnoses including emphysema, bronchioloalveolar cell carcinoma, organizing pneumonia, eosinophilic pneumonia, pulmonary hypertension, silicosis, and Langerhans histiocytosis. ⁸² In this study, ⁹¹ as in many others,^{25.87.88.89.90. and 266.} a confident CT diagnosis of UIP was correct in over 90% of cases. Most recently, in a multicenter treatment study, a ‘highly suggestive’ CT diagnosis of UIP, made in 49% of readings, was correct in 92% of cases. ²⁶⁷ Weighted kappa for expert radiologists determining presence or absence of UIP in this study was 0.40.

It is clear that honeycombing is the most important diagnostic feature of UIP, particularly in making the distinction between UIP and NSIP.^{16. and 82.} Sumikawa et al. ⁷⁹ found that the extent of honeycombing was the strongest predictive feature for distinguishing between UIP and fibrotic NSIP. The presence of upper lobe irregular lines has also been found to be another independent discriminator between UIP and non-UIP. ⁸² Chronic hypersensitivity pneumonitis (CHP) is an important aspect of the differential diagnosis of NSIP and UIP (see Chapter 8, p 461). Features that favor CHP over NSIP or UIP include upper- or mid-lung predominance, the presence of lobular decreased attenuation and air-trapping (Fig. 10.27), and the presence of poorly defined centrilobular nodules.^{268. and 269.} However, CHP may present with a histologic pattern identical to NSIP¹²⁹ or less commonly UIP. ²⁷⁰ Therefore, a thorough clinical evaluation for possible sources of hypersensitivity pneumonitis is important in all patients presenting with CT features suggestive of UIP or NSIP.

Pulmonary sarcoidosis, with its reputation as a great mimic, may rarely simulate the basal and peripheral pattern of UIP, with subpleural cysts being found in a small number of cases. ²⁷¹ However, the cysts of sarcoidosis are often larger than those of UIP.

Early studies of the CT differential diagnosis of NSIP noted difficulty in distinguishing it from other ILDs.^{25. and 134.} However, since these studies, our understanding of the CT findings of NSIP has evolved, and the typical pattern of basal predominant ground-glass attenuation and reticular abnormality with traction bronchiectasis has been recognized. The greatest challenge, both for pathologists and radiologists, is distinguishing between fibrotic NSIP and UIP. In a study by MacDonald et al. ¹³² of 21 patients with NSIP and 32 patients with UIP, the sensitivity, specificity and accuracy of a CT diagnosis of NSIP were 70%, 63%, and 66%, respectively. UIP was falsely diagnosed in 21 (33%) of 64 readings of patients with fibrotic NSIP. Conversely, NSIP was diagnosed in 48 (38%) of 128 readings in patients with histologic UIP. In a study by Elliot et al. ²⁷² of 29 patients with NSIP and 24 with UIP, a predominant pattern of ground-glass attenuation and/or reticular opacity, with minimal to no honeycombing, was demonstrated in 57 (98%) of 58 readings in patients with NSIP, and in 21 (43%) of 48 readings of patients with UIP. Conversely, the presence of honeycombing as a predominant feature had a specificity of 98% and positive predictive value of 96% for UIP. The positive predictive value of a confident diagnosis of NSIP was 77%, while the positive predictive value of a confident diagnosis of UIP was 91%. In a more recent study that also included hypersensitivity pneumonitis, the positive predictive values for confident diagnoses of NSIP and UIP were 94% and 100%, respectively. ¹⁴⁰ These studies indicate that it is possible to confidently identify UIP when honeycombing is present, but a substantial minority of patients with UIP have CT findings similar to those of NSIP, and can only be distinguished by biopsy. As indicated above, pathologists show substantial variation in differentiating between UIP and NSIP, so the gold standard for this diagnosis is not established, and the most accurate assessment probably requires clinical–imaging–pathologic correlation. ¹²⁸ For example, if the surgical biopsy is interpreted as NSIP, and substantial honeycombing is present on CT, then the imaging findings trump the pathology findings, in terms of the patient’s prognosis if not the final diagnosis.

At the other end of the spectrum, it may be difficult to distinguish histologic NSIP from organizing pneumonia. Focal areas of organizing pneumonia are seen histologically in about 50% of cases of NSIP. ¹²⁹ As with UIP, the amount of organizing pneumonia present on an individual biopsy specimen may vary with the site of biopsy. On CT, the main distinguishing features between organizing pneumonia and NSIP are that the consolidation of organizing pneumonia is more patchy and that reticular abnormality and traction bronchiectasis are less common. No clear borderline exists to delineate the extent of consolidation in organizing pneumonia from NSIP, and indeed there are cases where organizing pneumonia appears to evolve into NSIP. Organizing pneumonia should be favored in patients with patchy consolidation, while NSIP is favored in patients with basal predominant confluent abnormality associated with evidence of fibrosis.

The imaging features of NSIP may overlap with those of DIP. A subpleural distribution, and the presence of small cysts, should favor the diagnosis of DIP. However, the distinction between NSIP and DIP is not critical, since the treatment of these entities is similar. The long-term survival in DIP is excellent, and similar to that of cellular NSIP. ⁴¹

In patients presenting with the syndrome of AIP, the other IIPs are not usually included in the differential diagnosis. However, the differential diagnosis must include infectious pneumonia, acute hypersensitivity pneumonitis, acute eosinophilic pneumonia, pulmonary hemorrhage, drug or transfusion reaction, and cardiac or noncardiac edema. Tomiyama et al. ²³⁵ compared the CT features of AIP with other causes of acute respiratory failure. In their study, the combination of airspace consolidation, centrilobular nodules, and segmental distribution was found in most cases of infectious pneumonia, but was not seen in AIP. In particular, the finding of centrilobular nodules or tree-in-bud pattern was very suggestive of infection. Traction bronchiectasis and honeycombing, rarely seen in the other acute pneumonias, should suggest AIP. The presence of septal thickening or pleural effusion as predominant findings should suggest acute eosinophilic pneumonia (see Chapter 11, p 660). In the study by Tomiyama et al., ²³⁵ a correct first-choice diagnosis of AIP was made in 90% of cases of AIP, based on the findings indicated above.

There are a large number of papers describing the CT changes in rheumatoid lung disease.312.^{313.314.315.316.317.318. and 319.} In the majority of studies of symptomatic and asymptomatic individuals, bronchial wall thickening and bronchiectasis are the most common findings. Variation in the prevalence of CT findings probably reflects differing CT techniques and diagnostic criteria. Abnormalities found on CT are often not associated with symptoms.

In an early study of 84 patients, the most common findings were bronchial wall thickening with bronchiectasis (30%), nodules (22%), nonseptal lines (18%), pleural opacity (16%), ground-glass opacity (14%), honeycombing (10%), and consolidation (6%). ³¹² In 36 patients with recent onset RA, 33% had interstitial abnormality on CT, more commonly in men. ³¹⁴ In 126 patients with RA, bronchial dilation was found in 41%, ground-glass attenuation in 27%, parenchymal micronodules in 15%, reticulation in 12%, honeycombing in 9%, and consolidation in 4%. ³²⁰ Interstitial abnormalities were equally prevalent in those with early and longstanding RA, though bronchiolitis pattern was more prevalent in those with longstanding RA. In 75 unselected patients with RA, two-thirds of whom had no pulmonary symptoms, CT was abnormal in 49%, with interstitial abnormalities in 28%, and bronchiectasis in 19%. ³²¹ Factors significantly associated with CT abnormalities were age older than 40 years, IgM rheumatoid factor, hypoxia at rest, and physiologic evidence of distal airway disease. ³²¹ In a study of 34 patients with early RA (duration <1 year), CT showed expiratory air-trapping in 69%, bronchiectasis in 58%, and ground-glass opacity in 35%. ³²² In 52 unselected patients, CT was abnormal in 67%, with reticulonodular abnormality in 63%, ground-glass attenuation in 20%, and bronchiectasis in 17%. ³²³ An interesting recent study³²⁴ evaluated CT findings in 14 asymptomatic first-degree relatives of individuals with known RA. Six of seven individuals who had positive RA-related antibodies had expiratory air-trapping, compared with none of seven antibody-negative subjects. This intriguing finding suggests that the airways or the lung may be involved very early in the evolution of RA.

At postmortem examination, there are pleural changes in some 50% of cases of RA. ³²⁵ About 20% of patients with RA experience an episode of pleurisy. ³²⁶ However, clinical pleural effusion (Box 10.22) is much less common, and in a study of 516 patients with RA, only 17 (3.3%) had pleural effusions, with a higher prevalence in males (7.9%) than in females (1.6%). ³²⁷ Another large series found a 3.3% prevalence and an annual incidence of effusion of about 1% in patients with rheumatoid disease. ³²⁸ Unlike rheumatoid disease in general, but in common with other pulmonary manifestations of the disease, pleural effusion shows a striking preponderance of male patients.^{329. and 330.} Patients are usually middle aged, with a mean age of about 50 years, with high titers of rheumatoid factor, longstanding active articular disease and rheumatoid nodules.^{326. and 329.} The effusions in rheumatoid disease, unlike those seen in SLE, are commonly asymptomatic. ³³¹

Pleural effusions (Fig. 10.30) may be present at any phase of RA, ³²² may predate clinical disease,^{329. and 330.} or may even be present without arthritis. ³³² In about one-third of patients there will be other rheumatoid-related abnormalities on the chest radiograph such as interstitial pneumonia and fibrosis or parenchymal nodules. ³²⁹ Effusions are most commonly small to moderate in size³²⁹ though they can be large.^{333. and 334.} About one-fifth occur bilaterally. ³²⁹ Once formed, the effusions behave in a variable fashion, and although some resolve within weeks, more characteristically they persist for months and indeed sometimes for several years. ³²⁹ They often become loculated. Effusions may recur on the same or opposite sides of the chest. ³³⁵ Effusions may be associated with pneumothorax.^{336. and 337.} Once the effusion has cleared, it commonly leaves residual pleural thickening,^{328. and 330.} fibrothorax, and round atelectasis, ³²⁵ which may warrant decortication.^{327. and 338.} Probably because of the high cellular content, rheumatoid pleural effusions may be intensely hypermetabolic on FDG-PET scan. ³³⁹

The diagnosis may be strongly suspected or established by the findings in the pleural fluid,^{325.330. and 340.} and sometimes by pleural biopsy, which, although frequently nonspecific, may show rheumatoid nodules. The most characteristic features of the fluid are high cellular content, low sugar, low pH, and raised lactic dehydrogenase level. Rheumatoid factor is often present but is unfortunately a nonspecific finding. ³⁴⁰

Lung fibrosis was first associated with RA by Ellman and Ball in 1948. ³⁴¹ In a cohort of 609 patients with RA, the 30-year cumulative incidence of lung fibrosis was 6.8%. ³¹⁰ The two commonest histologic patterns of lung fibrosis in RA are UIP (see p 564) and NSIP³⁴² (see p 572). ³⁴³ Organizing pneumonia may also be seen.^{157.200.344.345. and 346.} A few cases of DIP have been described.^{346. and 347.} Lymphoid aggregates with germinal centers may be seen, ³⁴⁴ and, though nonspecific, are suggestive of rheumatoid lung. Increased numbers of T cells and B cells are present in the pulmonary lesions.^{348. and 349.} When lymphoid aggregates are seen adjacent to bronchioles, follicular or lymphocytic bronchiolitis is diagnosed.

Since most patients with lung fibrosis related to RA do not undergo lung biopsy, it is difficult to determine the precise prevalence of the histologic subtypes of interstitial pneumonia in these patients. Katzenstein et al. ¹²⁹ suggested that UIP was more common than NSIP in patients with rheumatoid disease. A study of 63 patients with rheumatoid lung, 17 of whom had undergone lung biopsy, identified histologic NSIP in 10 cases, UIP in two, organizing pneumonia in two, diffuse alveolar damage in one, DIP in one, and lymphocytic bronchiolitis in one patient. ³⁴³ However, among patients who did not undergo biopsy, a CT pattern of UIP (n = 22) was more common than NSIP (n = 11) or organizing pneumonia (n = 3), probably reflecting a lower threshold for performing lung biopsy in patients with an imaging pattern of NSIP as compared with UIP. A study of 16 patients with RA evaluated by surgical lung biopsy identified a histologic UIP pattern in seven, NSIP in seven, and mixed UIP and NSIP in two. ³⁴²

Rheumatoid lung fibrosis is substantially more common in men than in women,^{314. and 347.} with a mean age of onset of about 50 years. ³²⁹ A higher proportion of male cigarette smokers may account for some of the increased risk among men. Most cases occur after the onset of the arthritis, though some 15–30% occur before or coincident with the initial joint manifestations.^{347. and 350.} In general, whichever develops first, the interval between the onset of arthritis and lung disease is not more than 5 years. ³²⁹ Respiratory symptoms can be absent despite imaging changes³⁵¹ but, if present, consist chiefly of exertional dyspnea and cough.

The prognosis of interstitial pneumonia in rheumatoid disease varies from series to series. Some studies show that the survival is as poor as that seen in idiopathic forms of lung fibrosis,^{347. and 352.} while other authors have suggested that it is less severe.^{331. and 353.} A recent study indicated that the prognosis was similar to IIP, when adjusted for age. ³⁵⁴ The difference between these series is probably related to the clinical and histologic heterogeneity of rheumatoid lung fibrosis. In one series, the mean duration of lung disease to death was 5 years, with a range of 1–17 years. ³⁴⁷ A subset of patients has a fulminant course, dying within months of the initial diagnosis, presumably because of acute exacerbation or AIP.^{114.355. and 356.} Treatment with steroids or immunosuppressants helps improve physiologic and imaging findings of disease in a proportion of the patients, at least in the short term.^{347. and 356.}

The chest radiographic abnormalities in rheumatoid lung fibrosis are indistinguishable from cases of IIP (Fig. 10.30). Pleural effusion or thickening may coexist with interstitial changes in about 5–15% of cases.^{329. and 347.} Pulmonary nodules may also be seen in association with lung fibrosis. ³⁵⁷ CT findings in interstitial pneumonia associated with RA (Figs 10.31 and 10.32) are similar to those of the idiopathic variety.^{313.358. and 359.} However, associated nodules, mosaic attenuation, pulmonary arterial enlargement, and pleural abnormality may provide a clue to the underlying diagnosis (see Fig. 10.29). In a study of 63 patients with rheumatoid lung disease, ³⁴³ 26 had a CT pattern suggestive of UIP, 19 had a pattern of NSIP, 11 had a bronchiolitis pattern, and five had an organizing pneumonia pattern (Fig. 10.32). These CT patterns were in agreement with the histology in 13 of the 17 who underwent biopsy. Two patients who were classified as UIP by CT received a pathologic diagnosis of NSIP, one patient who was classified as NSIP received a pathologic diagnosis of DIP, and one patient with a CT diagnosis of LIP was classified pathologically as NSIP. Organizing pneumonia occurring in patients with RA is characterized on chest radiograph and on chest CT by patchy parenchymal opacities that may be migratory.^{157. and 345.} Drug toxicity and infection should always be considered in the differential diagnosis of this appearance.

There are several case reports of patients with rheumatoid disease who develop upper lobe fibrosis, sometimes associated with cavitation giving rise to a imaging appearance that resembles tuberculosis.^{360. and 361.} However, this complication appears quite rare.

Necrobiotic nodules occurring in the lung are pathologically identical to subcutaneous rheumatoid nodules³⁴⁷ and consist of a necrotic center bounded by palisading histiocytes (epithelioid cells) and surrounded by plasma cells and lymphocytes. Peripherally there may be a moderate, non-necrotizing vasculitis. ³⁶² Intrapulmonary nodules are rare findings on the chest radiographs of patients with rheumatoid disease; two large series of 955 patients contained no examples;363. ^{and 364.} and another contained only two cases in 516 patients. ³²⁹ Nodules, like pleural effusions and IPF, are more common in men than women, with about a twofold excess. Mean patient age at presentation is 51 years (range 24–64 years). Nodules usually occur in patients with established disease but may occur with or before the onset of arthritis.^{347.365. and 366.} Some patients with histologically proven nodules remain seronegative and never develop arthritis.^{366. and 367.}

Nodules are usually asymptomatic. ³⁶⁸ Occasionally, symptoms result from superadded infection, ³⁶⁹ bronchopleural fistula formation, or pneumothorax. ³⁷⁰ Subcutaneous nodules are found in 80% of those with pulmonary necrobiotic nodules.^{329.335. and 368.} Serum rheumatoid factor is found in nearly 90% of patients with nodules.^{335. and 368.}

The nodules (Fig. 10.33) are usually discrete on imaging, rounded or lobulated, and often subpleural. ³⁵⁸ They may be single or, in about three-quarters of patients, they are multiple, ³⁴⁷ sometimes coalescing, ³⁵⁸ and show some mid/upper zone predilection. ³⁶⁶ They range in size from a few millimeters³⁴⁷ to 7 cm. Occasionally, when the nodules are small and widely disseminated, a miliary pattern is produced. ³³⁵ In about 50% of cases, the nodules cavitate (Fig. 10.34),^{335. and 347.} producing ring opacities often with relatively smooth thick walls.^{335.371. and 372.} Cavitation is more easily appreciated on CT. ³⁵⁸ Occasionally, nodules calcify.^{360. and 371.} Subpleural nodules may erode through the pleura, and even into a rib³⁷³ and cause a bronchopleural fistula and hydropneumothorax.^{335. and 374.} Nodular lesions may increase in size and number, resolve completely, wax and wane,^{375. and 376.} or remain stable for many years. ³⁴⁷ Unless nodules are indolent or resolve, on imaging they are indistinguishable from pulmonary neoplasms and need either careful follow-up for increasing size, or histologic confirmation. ³⁷⁷ Cytologic evaluation of fine needle aspiration can be diagnostic. ³⁷⁸ Small rheumatoid nodules have also been described in the trachea³⁷⁹ and pleura.

CT is clearly more sensitive for lung nodules than chest radiographs. The prevalence of nodules has been recorded in various CT studies as 0%, ³¹⁴ 6%, ³¹³ 10%, ³¹⁵ 22%, ³¹² 28%, ³⁸⁰ and 37%. ³⁵¹ The larger figures contain a high proportion of micronodules. Remy-Jardin et al. ³¹² characterized nodules into three categories:

On FDG-PET scan, rheumatoid nodules may³⁸¹ or may not³³⁹ be metabolically active.

The increasing use of CT in patients with RA has made it plain that airways disease is one of the most common patterns of abnormality seen in these patients (Box 10.23). ³⁸² Both bronchitis (airway wall thickening) and bronchiectasis are more frequent in patients with rheumatoid disease than in matched controls. ³⁴⁰ Bronchiectasis may precede or follow the onset of rheumatoid disease.^{383. and 384.} In CT-based studies, an overall frequency of bronchiectasis of about 20% is recorded, with a range of 6–41%.^{312.314.315.318.320. and 380.} Bronchiectasis may be unassociated with respiratory symptoms. ³¹⁸ Bronchiectasis is usually cylindric in type, and is commonly, though not always, associated with CT and physiologic evidence of small airways disease (Fig. 10.35). ³⁸⁵

There is a recognized association between rheumatoid disease and obliterative bronchiolitis (constrictive bronchiolitis)199.^{360.386.387. and 388.} in which bronchioles are destroyed and replaced by scar tissue. Although penicillamine or gold has been implicated in the pathogenesis of obliterative bronchiolitis, it also occurs in patients taking neither drug. ³⁴⁰ Clinically there is progressive dyspnea and evidence of irreversible airflow obstruction. The chest radiograph is either normal or shows hyperinflation, with or without areas of oligemia. The characteristic CT finding is mosaic perfusion (Fig. 10.35) with air-trapping,^{389.390.391. and 392.} often associated with evidence of mild bronchiectasis.

Follicular bronchiolitis is a second type of small airway disorder recognized in rheumatoid lung disease.^{319.344.346.393. and 394.} It is characterized by lymphoid aggregates, with or without germinal centers, lying in the walls of bronchioles and possibly compressing their lumens. ³⁹⁵ Follicular bronchiolitis probably produces a reticular or reticulonodular pattern on the chest radiograph. ³⁴⁴ The major CT finding is centrilobular nodules, often associated with peribronchial nodules, and with areas of ground-glass abnormality.^{319. and 395.} HRCT may sometimes be normal. ³⁹³ Should lymphoid hyperplasia in rheumatoid lung disease be more concentrated along septa and subpleural areas than in airways, the term lymphoid hyperplasia is used. ³⁴⁶ Radiographically it gives rise to reticular or reticulonodular opacities³⁴⁴ but its HRCT features are not described; Remy-Jardin et al. ³¹² suggested that it may account for subpleural micronodulation, a common finding in rheumatoid lung disease.

Diffuse panbronchiolitis has been described in patients with RA, ³⁹⁶ but this could well be a chance association. A handful of patients with rheumatoid disease have had biopsy-proved bronchocentric granulomatosis (see Chapter 11, p 674).^{397.398. and 399.} On imaging, there were bilateral nodules and focal consolidations ranging in size from 2 cm to 15 cm, some of which were cavitated.

The original description of Caplan syndrome was of multiple, large (0.5–5.0 cm in diameter) rounded nodules seen on the chest radiographs of coal miners with RA. ⁴⁰⁰ Radiographic evidence of simple coal miner’s pneumoconiosis was often absent. The syndrome may occur with exposure to inorganic agents other than silica or coal. ⁴⁰¹ Caplan nodules are pathologically similar to necrobiotic rheumatoid nodules. ⁴⁰²

Although Caplan syndrome was quite common in the UK, ⁴⁰³ the prevalence in the USA has always been low, ⁴⁰⁴ and only a single English-language case report of this entity appears to have been published since 1965. ⁴⁰⁵

The classic imaging finding in Caplan syndrome is unilateral or bilateral pulmonary nodules (Fig. 10.36), usually 1–2 cm in diameter, but ranging from 0.5 cm to 5 cm. ⁴⁰⁰ Imaging changes of coal worker’s pneumoconiosis may or may not be present. ⁴⁰⁰ Nodules are typically situated at the junction of the outer and middle thirds of the lung and tend to appear in crops of lesions having a similar size and rate of growth. Nodules tend to develop rapidly and grow over a period of months. They may heal with scarring, remain stable or grow slowly for several years. Calcification or cavitation may be seen on chest radiograph⁴⁰³ or CT. ⁴⁰⁵ Smaller Caplan nodules may be indistinguishable from silicotic nodules. ⁴⁰⁵

Systemic vasculitis in rheumatoid disease is usually of the small vessel type and affects chiefly the skin. There is some evidence that the prevalence of vasculitis in RA has increased with increasing use of new treatments targeted at tumor necrosis factor. Pulmonary involvement is relatively uncommon.^{406.407.408. and 409.} Prognosis is poor and most patients have died within a year of diagnosis. ⁴⁰⁹ Rheumatoid pulmonary vasculitis must be distinguished from Wegener granulomatosis, which may also occur in patients with RA. ⁴¹⁰

Pulmonary hypertension commonly occurs in patients with RA, but is usually mild. ⁴¹¹ It increases in prevalence with increasing duration of RA. ⁴¹² It may be out of proportion to the visible extent of lung disease, suggesting the presence of an independent pulmonary vasculopathy. It should be suggested when the pulmonary arteries appear enlarged on chest radiograph or chest CT (see Fig. 10.29). Pulmonary hemorrhage is very uncommon in patients with RA.

Most studies show that RA is associated with an increased incidence of lymphoma, and of lung cancer (see Fig. 10.31). ⁴¹³

Many of the available treatments for RA, including gold, methotrexate and D-penicillamine have been implicated in the development of infiltrative lung disease (Table 10.5). ⁴¹⁴ On chest CT scans, gold-induced lung disease was characterized in 12 of 20 reported cases by alveolar opacities extending along bronchovascular bundles. ⁴¹⁵ Low-dose methotrexate may be associated with subacute hypersensitivity pneumonitis in 2–5% of cases (Fig. 10.37).^{416. and 417.} Preexisting radiographic evidence of ILD probably predisposes to the development of methotrexate pneumonitis in patients with RA. ⁴¹⁸ Treatment with D-penicillamine is associated with the development of constrictive bronchiolitis. Nonsteroidal anti-inflammatory drugs may be associated with hypersensitivity reactions (Fig. 10.38),^{419. and 420.} while unintentional therapeutic overdoses of salicylate may result in pulmonary edema. Leflunomide, a pyrimidine synthesis inhibitor, may cause a wide array of pulmonary complications, including DAD, acute eosinophilic pneumonia, and COP. ⁴²¹

The new generation of biologic agents used to treat RA have resulted in a new array of potential pulmonary side effects. The most important of these is impaired immunity related to use of antitumor necrosis factor (TNF)α antibodies (etanercept, infliximab, and adalimumab), which results in a substantially increased incidence of tuberculosis (sometimes disseminated or extraarticular) and nontuberculous mycobacterial infection. ⁴²² Fungal and pneumocystis infection may also occur. Screening chest radiographs are usually obtained when anti-TNFα antibody treatment is planned. Mycobacterial or fungal infection should be strongly suspected when new parenchymal abnormalities are identified in these patients.

Pleuritis is found in 40–60% of patients with SLE and it is the commonest pleuropulmonary manifestation.^{432.433. and 434.} It may be a presenting feature, ⁴³⁵ but occurs more commonly during an exacerbation of established disease. ³³¹

The pleuritis is dry 50% of the time, ⁴³⁶ but in the other 50% it is accompanied by a pleural effusion and sometimes also by a pericardial effusion (Fig. 10.39).^{429. and 437.} The pleural effusions are usually small or moderate in size. ⁴³⁷ Unilateral and bilateral effusions are equally frequent.^{435. and 437.} Sometimes the pleural effusions resolve spontaneously, but many require treatment with steroids or immunosuppressive drugs. Clearing may be complete or incomplete, leaving minor pleural thickening.^{435. and 438.} It is important to exclude other causes of pleural effusion in SLE, including the nephrotic syndrome, cardiac and renal failure, pulmonary embolism, and infective pneumonia.^{325. and 439.} The fact that primary pleural effusions in SLE are almost always painful^{436. and 437.} can be a helpful differentiating feature.

Pleural thickening⁴⁴⁰ is a common finding at autopsy, but in life is underreported as a imaging finding. It was, however, seen in 12% of patients in one series. ⁴²⁸ Pneumothorax is not a feature of pleuropulmonary SLE.

Acute lupus pneumonitis is a poorly defined entity, characterized by a variable degree of respiratory impairment accompanied by focal or diffuse pulmonary consolidation, occurring in patients with lupus. ⁴⁴¹ It is now believed that most cases previously identified as lupus pneumonitis probably represented AIP with or without pulmonary hemorrhage. ⁴⁴² The alveolar damage is most likely mediated by immune complex deposition. ⁴⁴¹ The clinical, histologic, and imaging features overlap substantially with those of pulmonary hemorrhage (Box 10.25; Fig. 10.40).

Pulmonary hemorrhage is common in the lungs at postmortem examination in patients with SLE, ⁴⁴³ but may not be recognized clinically. Lung hemorrhage ranges from a mild, chronic, subclinical process to one that is acute and life-threatening. ⁴⁴¹ In a large series of more than 400 patients with SLE at the National Institutes of Health (NIH), the prevalence of acute hemorrhage was only about 1.5%, ⁴⁴⁴ occurring typically in young (mean age ∼30 years) women. In 10–20% of patients with hemorrhage it is the first manifestation of SLE.^{445. and 446.} Acute pulmonary hemorrhage is an important diagnosis to make since urgent treatment is needed. The pathologic findings are diffuse alveolar hemorrhage with acute necrotizing capillaritis.^{362. and 445.} A minority of cases may be associated with DAD. ⁴⁴⁵ Bland pulmonary hemorrhage appears to be rare.

Clinically, in acute hemorrhage, there is rapid onset of severe dyspnea, fever, and hemoptysis. However, hemoptysis is absent in 33–50% of cases on admission. ⁴⁴⁷ Pulmonary hemorrhage is typically accompanied by signs of active disease elsewhere; fever, arthropathy, and particularly nephritis (14 out of 15 patients in one series).^{445. and 448.} Helpful clinical pointers include hemoptysis and a 3–4 g drop in blood hemoglobin level. ⁴⁴⁷ Confirmation can be obtained by bronchoalveolar lavage.

The radiographic appearances are indistinguishable from other forms of pulmonary hemorrhage and consist of airspace opacity that is usually bilateral and diffuse (Fig. 10.41).^{448. and 449.} The opacities usually clear within days. ⁴⁵⁰ CT shows patchy or diffuse parenchymal consolidation and/or ground-glass abnormality. Acute pulmonary hemorrhage must be differentiated from infection, and from lung edema due to renal or cardiac failure. In one series of 13 patients with 14 episodes of pulmonary hemorrhage, infection was identified in eight, with organisms including Pseudomonas and Aspergillus. ⁴⁵¹

Fibrotic ILD (Fig. 10.42) is less common in SLE than in the other collagen vascular diseases. ³⁷² However, in four autopsy series, chronic interstitial abnormalities were present in just over one-third of cases. ⁴⁴³ In a selected clinical series of 30 patients with pleuropulmonary SLE, there was a 13% frequency of interstitial pulmonary fibrosis judged on imaging (with biopsy confirmation in two), ⁴⁵² and in an unselected outpatient series assessed radiographically there was a 3% prevalence of interstitial pulmonary fibrosis. ⁴⁵³ A number of patients with interstitial pulmonary fibrosis have had preceding acute pneumonitis, ⁴⁵⁴ and the interstitial changes presumably represent the organizing phase of diffuse alveolar damage. The histology of other forms of interstitial pulmonary fibrosis is that of interstitial pneumonia, which is likely to be a mixture of UIP and NSIP, the latter probably being more frequent. ³⁶² In general, interstitial disease in SLE pursues an indolent course, but there are exceptions and it may cause death.

There is limited information on CT changes in SLE. In a study of patients with SLE who did not have previously diagnosed lung disease, Bankier et al. ⁴⁵⁵ found that 17 of 45 patients with a normal chest radiograph had an abnormal HRCT. Interlobular and intralobular lines were the commonest type of abnormality, being seen in 15 patients. Air space nodules, architectural distortion, bronchial wall thickening, bronchial dilatation, pleural irregularity, ground-glass attenuation and airspace consolidation were less common. Middle and lower lung zone involvement was predominant. The extent of abnormality on CT correlated with the length of the clinical history, and with impaired pulmonary function (decreased FEV₁/FVC ratio and DLco). Fenlon and co-workers⁴⁵⁶ studied 34 consecutive patients with SLE, 56% of whom had never smoked and 23% of whom had respiratory symptoms. The commonest signs were thickened interlobar septa (44%), parenchymal bands (44%) (Fig. 10.43B), subpleural bands (21%), and pleural tags and thickening (15%). Only 6% had ground-glass opacity, consolidation, or honeycombing (Fig. 10.42). These interstitial changes were usually mild in degree, less commonly moderate and rarely severe. Overall, 33% of patients were judged to have mild or moderate ILD. In 21% of patients there was bronchiectasis and bronchial wall thickening. In contrast to the study by Bankier et al., ⁴⁵⁵ Fenlon et al. ⁴⁵⁶ did not confirm a relationship between the presence or extent of lung abnormality and pulmonary physiology or duration of disease. Because the extent of disease on CT in patients with SLE is usually relatively minor, the clinical value of CT in asymptomatic patients is questionable.

Diaphragm dysfunction in SLE may be manifest as elevation of one or both hemidiaphragms (Fig. 10.43) and is a cause of dyspnea. Bilateral elevation is a common finding and, in some series, has been the most common imaging pleuropulmonary abnormality in SLE, being seen in as many as 18% of patients. ⁴²⁸ As the diaphragm rises, the lungs lose volume, hence the term ‘shrinking lungs’, ⁴⁵⁷ a finding first noted in 1954. ⁴³² The loss of lung volume was initially ascribed to reduction in lung compliance, but more recent consensus suggests it is due to diaphragmatic weakness⁴⁵⁸ presumably caused by a myopathy,^{459. and 460.} for which there is some pathological evidence. ⁴⁶¹ Pleuritic pain with splinting may be a contributory factor.^{459. and 462.} Most patients stabilize or improve over time. ⁴⁶² CT is helpful in excluding lung fibrosis as a cause of decreased volumes, and usually shows basal atelectasis adjacent to the elevated hemidiaphragms (Fig. 10.43). On fluoroscopy, diaphragmatic motion may be sluggish, or paradoxical upward movement of the diaphragm may be seen with inspiration or sniffing.

A variety of changes are described pathologically in the pulmonary vessels.^{362. and 463.} Chronic lesions consist of intimal thickening, subintimal fibrosis, medial hypertrophy, and periadventitial fibrosis. Plexogenic lesions similar to those of primary pulmonary hypertension, and organizing thrombi may also be seen. ⁴⁶⁴ Marked pulmonary arterial hypertension is unusual. ⁴²⁹ Mild pulmonary artery hypertension is quite common⁴⁶⁵ and in one study with echocardiographic assessment on two occasions, separated by 5 years, there was respectively a 14% and 43% frequency of pulmonary artery hypertension with relatively modest mean pressures of 23 mmHg and 28 mmHg. ⁴⁶⁶ Pulmonary artery hypertension is associated with digital ischemia, livedo, ⁴⁶⁷ and, in 75% of patients, Raynaud phenomenon, ⁴⁶⁸ findings that are consistent with a vasospastic mechanism. ⁴⁶⁹ Some cases of pulmonary arterial hypertension have been associated with the presence of lupus anticoagulant, discussed below. ⁴⁶⁸ In these cases, it is very important to exclude central pulmonary thromboembolic disease.

Up to two-thirds of patients with SLE test positive for antiphospholipid antibodies, ⁴⁴² and in a literature review of more than 1000 patients, 34% had lupus anticoagulant and 44% anticardiolipin. ⁴⁷⁰ The presence of antiphospholipid antibodies is an important risk factor for thrombosis (venous and arterial), neurologic disease, thrombocytopenia, and fetal loss. Subjects with SLE and antiphospholipid antibodies have a sixfold increased risk of venous thromboembolism, compared with those with SLE but no antibodies. ⁴⁴² The risk is greater with the lupus anticoagulant than with the anticardiolipin antibody. A review of the thoracic imaging features of 88 patients with antiphospholipid antibodies⁴⁷¹ found that pulmonary embolism was the primary intrathoracic manifestation, being seen in nine patients, usually associated with thrombosis of lower extremity veins. Thrombosis of the superior vena cava or its tributaries⁴⁷¹ may also occur.

The antiphospholipid syndrome, which may occur without other manifestations of SLE, and may occur in other collagen vascular diseases, is defined by the occurrence of venous or arterial thrombus, and/or fetal loss, associated with the presence of one of the antiphospholipid antibodies. ⁴⁴² The antiphospholipid antibody syndrome may cause thromboembolic pulmonary hypertension.^{472. and 473.} Pulmonary arterial aneurysms, presumably due to vasculitis, have also been described in this syndrome. ⁴⁷⁴ Other manifestations may include nonthrombotic pulmonary hypertension, ⁴⁷⁵ lung fibrosis, ⁴⁷⁶ acute interstitial pneumonia, and cardiac valvular lesions. ⁴⁴² Diffuse alveolar hemorrhage (Fig. 10.44) ⁴⁷⁷ is quite common in our experience. Antiphospholipid syndrome should be suspected in patients with pulmonary embolism or pulmonary hypertension in whom there is no evident predisposing factor, and particularly in those with thrombosis at unusual sites such as central veins or dural sinuses. ⁴⁷⁸ Magnetic resonance angiography may be helpful in visualizing the entire vascular system to document the extent of venous and arterial thrombi. ⁴⁷⁹ The catastrophic antiphospholipid syndrome consists of multiorgan microvascular occlusion, with a mortality rate of about 50%. ⁴⁸⁰

Organizing pneumonia has been recorded in a handful of patients.^{200. and 481.} Acute reversible hypoxemia is thought to be caused by polymorph aggregation in the pulmonary circulation following activation by complement split products. The chest radiograph is normal and the condition resolves with steroid therapy. ⁴⁸²Pulmonary venoocclusive disease has been recorded. ⁴⁸³

This is considered in the section on drug-induced lung diseases (Chapter 9, p 512).

The CREST syndrome is a clinical subtype of scleroderma in which (C)alcinosis, (R)aynaud’s phenomenon, (E)sophageal dysmotility, (S)clerodactyly, and (T)elangiectasia are prominent features.^{559. and 560.} It is commonly associated with the anticentromere antibody (ACA). ⁴⁹⁸ Although patients with this syndrome often have less severe skin and visceral involvement than in scleroderma, about 60% of patients with CREST syndrome develop significant pulmonary hypertension. ⁵⁶¹ Selection bias in various series makes prevalence difficult to assess, but it is probably similar to that of classical systemic sclerosis. ⁵¹⁴ The majority of patients are older women presenting with a long history of Raynaud phenomenon and swollen fingers, but less extensive skin involvement than systemic sclerosis. ⁵¹⁴ Between 50% and 80% of patients have ACA – an unusual finding in systemic sclerosis.^{493. and 514.} It also differs from systemic sclerosis in that life expectancy is greater, and the disease is quite often mild and slowly progressive. Systemic involvement such as of the kidney is less. ⁵⁶² However, the CREST syndrome is by no means benign, and there are several reports of severe pulmonary hypertension and deaths from cor pulmonale.^{562. and 563.} Although earlier reports suggested that the prevalence of pulmonary hypertension in CREST syndrome is about 10%,^{563. and 564.} more recent studies found echocardiographic evidence of pulmonary hypertension in 30–60% of cases.^{542. and 561.} The reported prevalence of lung fibrosis in CREST syndrome has ranged from 0% to 72%, depending at least in part on the criteria used for evaluation^{518.562.565. and 566.} In studies using HRCT, the prevalence of parenchymal abnormality is 25–30%,^{498. and 533.} but the extent of abnormality is less than that seen in diffuse scleroderma. ⁵³³

This is the other major clinical variant of systemic sclerosis, constituting 10–27% of all cases of systemic sclerosis.^{495. and 518.} Overlap occurs with one or more connective tissue disorders: SLE, rheumatoid disease, and dermatomyositis. Pleuropulmonary involvement is more common than in classical systemic sclerosis, with radiographic fibrosis in 25–65%^{495. and 518.} and pleural effusions in 15–36%.^{518.558. and 567.}

Interstitial fibrosis was first described in dermatomyositis in the 1950s. ⁵⁷⁸ It is now a well-recognized association⁵⁷⁹ that occurs in up to 40% of patients. ⁵⁸⁰ It is slightly more common with polymyositis than with dermatomyositis. It may also occur in amyopathic dermatomyositis. The presence of ILD in polymyositis/dermatomyositis correlates strongly with the presence of anti-Jo-1.^{470. and 581.} About 50–70% of patients who are anti-Jo-1 positive have ILD⁵⁷¹ whereas the frequency of ILD falls to about 10% if antibodies are absent. ILD may antedate myositis in patients with anti-Jo-1 antibodies. ⁵⁸²

ILD tends to be accompanied by joint involvement⁵⁸³ and, as with ILD in other collagen vascular diseases, pulmonary changes on the chest radiograph can be the first clinical manifestation, preceding the skin rash or myositis,583.^{584. and 585.} although the usual pattern is for the ILD to follow soon after the onset of muscle weakness. The clinical manifestations vary greatly. At one extreme is an acute, rapidly fatal illness resistant to therapy^{571.586.587. and 588.} in which the muscle disease can be masked by respiratory involvement. At the other is a benign, indolent, and asymptomatic form.^{586. and 588.} It is said that the ILD of polymyositis/dermatomyositis is more steroid responsive than that of systemic sclerosis, ³³¹ perhaps due to a higher prevalence of organizing pneumonia. About half the patients with lung involvement respond with lessening of dyspnea, clearing of the chest radiograph, and improvement in lung function tests.^{578. and 587.} The variation in the clinical course of ILD in polymyositis/dermatomyositis and its generally favorable response to steroid therapy reflects the variety of underlying pathologic changes and their relative frequency. The most common pathologic findings are NSIP^{202.573.589. and 590.} and organizing pneumonia,^{591.592. and 593.} often occurring in combination. A rapidly progressive form of lung disease is associated with DAD.^{594.595. and 596.}

Characteristic imaging changes commonly consist of symmetric, basally predominant reticulonodular opacity, often with superimposed airspace opacity (Fig. 10.50).^{586. and 587.} Lung volumes are often markedly reduced, particularly in the lower lungs. Although pleuritis, fibrosis, and small effusions are common pathologically, ⁵⁸⁷ they are uncommon on imaging. The chest radiograph may show muscle calcifications (Fig. 10.51).

Lung disease associated with polymyositis/dermatomyositis or with the antisynthetase syndrome is often associated with a characteristic CT appearance, characterized by confluent ground-glass opacity and consolidation in the lower lobes, superimposed on a background of reticular abnormality with traction bronchiectasis (Fig. 10.52).^{591.597. and 598.} This pattern reflects the characteristic histologic combination of organizing pneumonia and fibrotic NSIP.^{591. and 598.} Parenchymal band opacities, typically basal and peripheral, and irregular thickening of bronchovascular interfaces may be seen. Honeycombing is relatively uncommon, with recorded prevalence ranging from 0% to 26%, and reflects underlying UIP.^{591.592. and 598.} In patients with acute rapidly progressive lung disease due to DAD, extensive consolidation and ground-glass opacity are seen. ⁵⁹¹ On serial evaluation, the changes of consolidation, ground-glass abnormality, reticular abnormality, and traction bronchiectasis may all be partially reversible with treatment.^{590.591. and 597.} Less commonly, consolidation may progress to reticular abnormality.

When the diaphragm becomes involved in the myositis⁵⁹⁹ functional and radiographic changes are produced similar to those seen with the diaphragmatic myopathy of SLE. Characteristically the myopathy produces bilateral hemidiaphragm elevation, reduced lung volume, and discoid basal atelectasis.^{578. and 599.} Pulmonary arterial hypertension⁵⁷¹ is relatively uncommon in polymyositis/dermatomyositis, except as a complication of ILD (see Fig. 10.51). ⁵⁸⁵

Aspiration pneumonia may occur in patients with pharyngeal or esophageal dysfunction due to polymyositis/dermatomyositis, ⁵⁷⁸ and should be considered in the differential diagnosis of new parenchymal opacities.

There is a two- to sevenfold increase in the frequency of malignant disease in polymyositis/dermatomyositis,^{413.600. and 601.} especially in older patients. The average prevalence in various series is 21% for DM and 15% for PM. ⁶⁰² The carcinomas recorded roughly parallel their frequency in the general population, ⁶⁰² though ovarian cancer may be more common. ⁶⁰³ Polymyositis/dermatomyositis may precede, accompany, or follow the carcinoma, though it is usually found within 1 year of diagnosis. ⁶⁰⁴ The course of the myopathy often parallels the course of the malignancy, improving when the malignancy is treated, and worsening with relapse, suggesting that it is a true paraneoplastic syndrome. ⁶⁰⁵ The prevalence of malignancy in most series is higher in patients with dermatomyositis than in those with polymyositis,^{604. and 606.} and cutaneous necrosis or mucinosis should lead to even higher suspicion for malignancy.^{607. and 608.} Using Callen’s data, ⁶⁰⁹

Stay updated, free articles. Join our Telegram channel

Join