A CXR is frequently helpful in evaluating patients with dyspnoea. Characteristic roentgenographic findings occur in patients with congestive heart failure and pneumonia, and pulmonary fibrosis. The chest radiograph may also be abnormal in patients with obstructive pulmonary disease, but the chest film (particularly the bedside chest film) have low sensitivity above all for the detection of airflow obstruction or pulmonary embolism (2).

Patients with COPD usually have one or two exacerbations per year, often needing hospitalization, with an overall mortality of 3-4%. Incidence of death is higher in the intensive care unit (24%) (4).

For these reasons CXR is not recommended as a routine exam, but only in cases of suspected pneumonia, or to ruleout other causes of dyspnoea, such as massive pleural effusion, atelectasis, pneumothorax, pulmonary edema.

To these purposes, the radiologic signs and findings to be studied are: the perfusion pattern and the spatial distribution of oedema, the size of the vascular peduncle and the cardiac volume. Moreover, it is highly important the recognition of some specific signs, like lung interstitial oedema, pleural effusion and air bronchogram.

In cardiogenic pulmonary edema, CXR may show cardiomegaly, pulmonary venous hypertension, and pleural effusions. Radiologic signs of cardiogenic APE are related to the severity of the condition, and may be divided into 3 stages (Table 1) (11,12). In stage I, an upright examination demonstrates redistribution of blood flow to the nondependent portions of the lungs and the upper lobes (Figure 2). In stage II, there is evidence of interstitial edema with ill-defined vessels and peribronchial cuffing, as well as interlobular septal thickening (Figure 3). In stage III, perihilar and lower-lobe airspace filling is evident, with features typical of consolidation (e.g., confluent opacities, and the inability to see pulmonary vessels in the area of abnormality) (Figure 4). The airspace edema tends to spare the periphery in the mid and upper lung. The distribution of the alveolar edema can be influenced by:
• Gravity: supine or erect position and right or left decubitus position;
• Obstructive lung disease, i.e. fluid leakage into the less severe diseased areas of the lung.

In non-cardiogenic causes, cardiomegaly and pleural effusions are usually absent. The edema may be interstitial but is more often consolidative. The cephalization of blood flow is missing, though there may be shift of blood flow to less affected areas. The edema is diffuse and does not spare the periphery of the mid or upper lungs (Table 2) (Figure 5).

In cases of large, acute myocardial infarction (MI) and infarction of the mitral valve, support apparatus may produce atypical patterns of pulmonary edema that may mimic noncardiogenic edema or in some cases even a pneumonia.

This pathologic condition is quite frequent and sometimes constitutes a hemodynamic and respiratory emergency, leading to death in 30% of untreated cases (14,15). To date, APT is considered the third leading cause of death in western countries and the most misdiagnosed pathologic condition, being correctly diagnosed only in 20% of cases (16). Physical signs as well as routine diagnostic tests are not enough accurate for a safe diagnosis of the condition. Indeed, history, physical examination and blood d-dimer are useful to hypothesize APT in the emergency setting and determine the pre-test probability according to the criteria published by Wells and co-authors (17).

Quite often, CXR is completely normal in APT. Instead, spiral angio-CT (SCT) scan has a well defined role and it is the first level radiographic test when a clinical suspicion has been hypothesized and classified by the clinician (19,20). SCT has a higher sensitivity (87% vs. 33%) and specificity (95% vs. 59%) over CXR, and indubitable advantages due to its fast execution, broad view and objective interpretation, as well as its ability to allow for other diagnoses when the initial clinical suspicion is excluded (21).

Limitations of CXR are related to the difficulty to recognize specific signs. Some radiologic findings have been corroborated in many years of experience. They have been argued by the careful observation of CXR studies in patients with confirmed APT, but rarely such signs are found altogether even in case of clear clinic presentations (22). Nevertheless, many authors suggest that a careful observation of CXR images can show some non specific abnormalities in at least 90% of the cases (23-25). The possible findings of standard CXR in APT are the following (16,26):
(I) Pulmonary infiltrates, due to haemorrhagic or oedematous infiltration of secondary lobules, often multiple and presenting as round foci of alveolar consolidations or irregular jeopardized opacities, without a segmental arrangement, more often located to the right base, sometimes associated with signs of atelectasis or pleural effusion.
(II) Atelectasis, often sub-segmental, appearing as curved lines reaching the pleura, secondary to alveolar collapse (line of Fleishner), caused by bronchial obstruction due to mucosa congestion, or alveolar collapse secondary to surfactant reduction, or hypoventilation due to reduced diaphragmatic excursion (Figure 6).
(III) Diaphragm elevation secondary not only to reduction of pulmonary volume due to the reduction in surfactant, but mainly to the dysventilation due to reduced respiration movement during pleural pain (Figure 6).
(IV) Pleural effusion, mainly serous, bilateral and of slight entity, often in association with basal atelectasis.
(V) Westermark sign, uncommon but highly specific, corresponds to a region of impaired vascularisation in the lung region distally to the site of the embolism (Figure 7) (27). Sometimes it is associated with deletion and dilation of the affected pulmonary branch (more often the right pulmonary artery). For a safe interpretation of this sign when present, the film should be compared with an old radiogram where it was absent. Another limitation of this sign is linked to the difficult visualization when CXR is performed in the supine patient.
(VI) Right heart and azygos vein enlargement are signs of severe pulmonary hypertension and right heart failure. They are invariably associated with symmetric enlargement of the ilar regions and other signs previously described. As for the Westermark sign, visualization of these signs should always be compared with previous images and they are unreliable when examination is performed in the recumbent position.
(VII) Hampton’s hump is a triangular opacity with the apex pointing to the hilar region, sometimes with blurred margins and irregular shape. It is a sign of interruption of blood supply from the systemic circulation in the lung region previously excluded by embolic obstruction of the functional circulation. It is more frequent when APT overlaps with some pre-existing conditions, like venous pulmonary hypertension, cardiac illnesses with left heart failure and COPD. Very often this sign is associated to pleural effusion. Often, the differential diagnosis with an alveolar consolidation due to pneumonia is difficult.

When a CXR is not acquired in an orthostatic posterioranterior view, there are some other indirect signs that can be important for diagnosing pneumothorax. These are the emphasized radiolucency of the paracardiac region, the deep sulcus sign (32), the appearance of sharp edges of mediastinum, heart and subcutaneous tissues, or the visibility of the anteriorinferior edge of the lung (33). Anyway, these signs are pathognomonic but not constant.

When possible, in doubtful cases acquisition of a radiogram in the lateral view (Hessen position) or during a forced expiration, can be useful (21,34). In these cases, it is sometimes possible to demonstrate even the smaller layer of pneumothorax.

Free air can also collect in a fissure or behind the triangular ligament, or it can distribute around an atelectasis or a consolidated lobe, sometimes with unusual aspects against the expected gravity distribution. This is due to variations of intrapleural pressure in presence of various chronic pulmonary diseases (Figure 9).

In these cases the differential diagnosis between pneumothorax, pneumo-pericardium and pneumo-mediastinum at CXR can be very difficult.

Diagnosis of tension pneumothorax is generally based mainly on the first clinical evaluation because it gives usually clear physical signs that may evolve rapidly to hemodynamic shock and cardiac arrest. When the clinical conditions are not rapidly evolving, CXR may be helpful in the early diagnosis allowing the emergency physician a greater confidence in deciding aggressive life-saving decompression treatment. The main radiologic signs of tension pneumothorax are the lateral shift of heart and mediastinum, the lowering of the hemi-diaphragm, the flattening of the cardiac profile, the reduced size of the superior vena cava and the protrusion of the parietal pleural layer between the intercostal spaces.

Classical radiologic signs are consistent with a dependent opacity with lateral upward sloping of a meniscus-shaped contour. The diaphragmatic contour is partially or completely obliterated, depending on the amount of collected fluid (silhouette sign) (Figure 10A,B). In case of massive effusion, all the hemi-thorax can be filled and mediastinum can be shifted contra laterally.

If CXR is acquired at bedside in the anterior-posterior view, it is extremely easy to underestimate the real amount of the free effusion (15). Moreover, from 10% to 25% of the milder forms of effusion can be completely misdiagnosed by bedside CXR (4). Some radiologic signs allows diagnosis of pleural effusion at CXR, even if the classical visualization of the basal opacity is lacking. They are the thickening of fissures and of pleural line at the apex, the blurring of the diaphragmatic profile and the haze of costophrenic angle, the complete but slight haze of the hemi-thorax with still visible vascular tree. In a supine patient, one of the more declivous part of the thorax are the apical posterior zones, so in this place can accumulates large amount of pleural effusion for gravity. These signs are useful only when a comparison between the two hemi-thorax can be performed, while in case of massive effusion equally distributed on both sides, they are extremely difficult to be recognized.

When bedside CXR is correctly interpreted, the reader can detect large pleural effusions 92% of the time and can exclude large effusions with high confidence (42).

In selected cases a lateral view with 20° of Trendelemburg inclination (the Hessen view) can obviate to lack of accuracy (37,43) (Figure 11A,B,C). This manoeuvre may visualize even small amount of effusion, normally located in intrapulmonary regions, because fluid move to the pleural space near the costal plane of the superior chest, were concavity is more accentuated. The presence of a short pulmonary ligament allows the accumulation of huge amount of pleural effusion (>500 mL) below the lung, thus mimicking a lifting of the hemi-diaphragm (Figure 12). This approach is now replaced by lung ultrasound.