CT imaging of blunt chest trauma

Trauma-related pneumothorax occurs in 30–40% of cases, and it is most commonly associated with rib fractures that lacerate the lung. Less commonly, pneumothorax may be caused by a disruption of closed airway spaces, such as the alveoli, due to a sudden increase in intrathoracic pressure or to a direct impact or deceleration force to the chest wall. Tracheobronchial injuries are also always associated with pneumothorax [5, 7, 16, 17]. CT is more sensitive in detecting pneumothoraces (Fig. 1), as 78% of them are nowadays believed to be missed on chest radiograph (occult pneumothoraces) [18, 19]. Pneumothorax in supine polytrauma patients tends to accumulate at the anterior and medial aspect of the lung, rendering it difficult to recognise on a supine chest radiograph, although it might be visible on an upright chest radiograph. Radiographic signs that may be present in the case of an occult pneumothorax include:

It is crucial to detect even a small pneumothorax in the trauma patient, as this can significantly enlarge under positive mechanical ventilation in the ICU or during general anaesthesia and endotracheal tube placement. Consequently, a prophylactic chest tube placement is considered [18] in small asymptomatic pneumothoraces (<20%), although controversies exist about this practice in the literature [21]. However, there is growing evidence that occult pneumothorax can be safely treated without thoracostomy in non-ventilated patients [19].

Tension pneumothorax is an urgent clinical diagnosis where progressive accumulation of air—due to the one-valve mechanism—increases the intrathoracic pressure of the hemithorax involved, causing a contralateral shift of the mediastinum, compression of the superior vena cava and loss of venous return to the heart with resultant haemodynamic impairment. Chest radiography and CT will both show a contralateral shift in the mediastinum, hyperexpansion and hyperlucency of the ipsilateral lung with lung collapse towards the hilum, and inversion of the ipsilateral diaphragm (Fig. 2). Sudden evacuation of a large pneumothorax with tube drainage can be complicated by re-expansion pulmonary oedema, presenting the corresponding radiological signs. The complication is more common in younger patients (20–50 years of age) and occurs more often than was previously believed, and although it may be entirely asymptomatic, it has a reported variable mortality rate reaching 20% [22].

Haemothorax occurs in 50% of chest trauma cases, with blood pooling into the pleural space from variable sources: the lung parenchyma, the chest wall, the great vessels, the heart or even the liver and spleen through diaphragmatic rupture [20]. Arterial bleeding (more commonly from the intercostal arteries, and the subclavian and internal mammary arteries) causes a more significant progressive increase in volume and mass effect compared with a venous origin of haemorrhage [23]. Massive haemothorax occurs when the accumulation of blood in the pleural space exceeds 1 l and is accompanied by haemodynamic impairment (Fig. 3) [2]. CT is very sensitive in detecting even a small haemothorax and can further characterise it by measuring accurately the Hounsfield (HU) units attenuation values of the pleural fluid. A reactive pleural effusion will have values not higher than 15 HU, while liquid blood will measure 30 to 45 HU, and the clotted blood should measure around 50–90 HU units [20, 24]. Occasionally a “haematocrit effect” is caused by the layering of different ages and statuses of coagulation of pleural blood products. In the case of active bleeding the fresh extravasated blood in contrast-enhanced CT may have attenuation values similar to the adjacent enhanced thoracic vessels (±10 HU) [17].

Lung contusion is a focal parenchymal injury caused by disruption of the capillaries of the alveolar walls and septa, and leakage of blood into the alveolar spaces and interstitium [25]. It is the most common type of lung injury in blunt chest trauma with a reported prevalence of 17–70% [26]. The main mechanism is compression and tearing of the lung parenchyma at the site of impact (it may also occur contralaterally “contre-coup”) against osseous structures, rib fractures or pre-existing pleural adhesions [27]. Lung contusion occurs at the time of injury, but it may be undetectable on chest radiography for the first 6 h after trauma. The pooling of haemorrhage and oedema will blossom at 24 h, rendering the contusion radiographically more evident, although CT may readily reveal it from the initial imaging [28]. The appearance of consolidation on chest radiography after the first 24 h should raise suspicion of other pathological conditions such as aspiration, pneumonia and fat embolism [2]. Contusions appear as geographic, non-segmental areas of ground-glass or nodular opacities or consolidation on CT that do not respect the lobar boundaries and may manifest air bronchograms if the bronchioles are not filled with blood (Fig. 4) [17]. Subpleural sparing of 1–2 mm may be seen, especially in children (Fig. 4b) [29]. Clearance of an uncomplicated contusion begins at 24 to 48 h with complete resolution after 3 to 14 days [9]. Lack of resolution within the expected time frame should raise the suspicion of complications such as pneumonia, abscess or ARDS. Pulmonary contusion—despite the advances in prompt diagnosis with imaging and supportive management with critical care medicine—remains a predictor of ARDS and has a high mortality rate (10–25%) [30].

Pulmonary laceration occurs in major chest trauma when disruption and tearing of the lung parenchyma follows shearing forces, caused by direct impact, compression or inertial deceleration [27]. Lung lacerations have been classified into the following four types according to the mechanism of injury [27]:

Herniation of the lung parenchyma is an uncommon manifestation of blunt chest trauma, and it can occur through a congenital or a traumatic chest wall defect such as multiple rib fractures or sternoclavicular or costochondral dislocations. Surgical repair is indicated when the patient is symptomatic or if the patient needs intubation and general anaesthesia as herniation may increase with positive-pressure ventilation [24, 31].

Blast lung is the most common fatal injury among initial survivors of explosions; 17–47% of people who die from explosions have had primary blast lung injury [8, 9]. However, the in-hospital mortality rate for these patients ranges from 3.4 to 25% because of prompt diagnosis and aggressive treatment. The blast wave causes thoracic acceleration and propagates through lung parenchyma with subsequent severe disruption at the capillary-alveolar interface. This results in parenchymal haemorrhage and contusions, pulmonary oedema, pneumothorax, barotrauma and air embolism from arteriovenous fistulas, causing substantial immediate and delayed injury. Chest radiography and CT (Fig. 7) will reveal the “butterfly or batwing” pattern, representing central bilateral perihilar air space consolidation and ground-glass opacities that may contain air bronchograms [32].

Tracheobronchial injuries are rare, occurring in 0.2–8% of all cases of chest trauma. It is anticipated that the prevalence is higher, as 50% of patients die at the trauma scene within the first 2 h from associated injuries and respiratory insufficiency [2, 33]. They have a mortality rate of 30%, and in two thirds of cases the diagnosis is delayed with subsequent serious complications, such as pneumonia, abscess, empyema, mediastinitis, sepsis, airway obstruction or atelectasis. Bronchial injuries occur more commonly than tracheal, usually on the right side and within 2.5 cm from the carina [5, 24], while 85% of tracheal lacerations occur 2 cm above the carina. Bronchial lacerations are usually parallel to the cartilage rings as opposed to tracheal ones that are vertical to the cartilage rings. A direct CT finding of tracheobronchial injuries is the cutoff of the tracheal and bronchial wall with extraluminal air surrounding the airway (Fig. 8). Indirect findings are the “fallen lung” sign, corresponding to the collapsed lung resting away from the hilum towards the dependent portion of the hemithorax [34], persistent pneumothorax after chest tube placement and herniation or overdistention of an endotracheal balloon if this is placed at the same level as the tracheal laceration [33]. Tracheal lacerations are usually associated with cervical subcutaneous emphysema. Tracheobronchial injuries in general are accompanied by pneumothorax and pneumomediastinum.

Pneumomediastinum occurs in 10% of patients with blunt chest trauma, with less than 2% caused by blunt tracheobronchial injuries. Other sources of air originate from lung parenchymal injury, oesophageal injury, chest wall, neck and retroperitoneal injury. In a number of patients pneumomediastinum is attributed to the Macklin effect caused by alveolar ruptures that lead to air dissecting along bronchovascular bundles and spreading of the pulmonary interstitial emphysema into the mediastinum. Streaks of air surrounding and paralleling the bronchovascular bundles associated with pneumomediastinum may be observed on CT [35]. Pneumomediastinum may be mistaken for pneumothorax, but the presence of septae within it—delineated on wide lung window—may help in differentiating the two findings, especially if they coexist (Fig. 9).

Cardiac injuries are the most lethal in chest trauma patients. They are more common in penetrating trauma, but they can occasionly occur in motor vehicle accidents and from severe blows to the anterior chest wall. Cardiac injury may be difficult to depict acutely in those who survive and should be treated with a high degree of clinical suspicion. It may range from a small focal contusion to a frank rupture of the heart, which is rare (<2%), affecting mainly the right atrium [12]. CT has low sensitivity in detecting cardiac injury and may show haemo-pneumopericardium, cardiac contusion as a hypodense, non-enhancing focal area in the myocardium, contrast medium extravasation in the pericardial sac or the mediastinum, coronary artery laceration or valve injury [36]. Pericardial tears occur extremely rarely (0.11%) and may range from tiny to the entire length of the pericardium. They usually manifest at the diaphragmatic surface of the pericardium on the left side. In cases of a large tear, intrapericardial cardiac herniation may occur, leading to death. Occasionally, there may also be intrapericardial bowel herniation [2]. On CT, the most common findings are haemopericardium (Fig. 10) and pneumopericardium (Fig. 11).

Oesophageal injury is far more common in penetrating or iatrogenic injury and occurs only in 1% of cases of blunt chest trauma. The main mechanisms are a direct blow to the neck mainly affecting the cervical oesophagus, burst-type force or hyperextension injury affecting the distal oesophagus, or rupture by a vertebral body fracture [12]. On CT there will be mainly indirect findings of oesophageal rupture such as pneumomediastinum and peri-oesophageal air or abnormal mediastinal contour secondary to leakage of fluid, haematoma or mediastinitis. Hydropneumothorax is usually seen on the left side. Diagnosis can be confirmed (Fig. 12) by water-soluble contrast oesophagography showing leakage of contrast medium into the mediastinal or pleural space [37].

Aortic injury occurs in 0.5%–2% of all non-lethal MVAs and is responsible for 10–20% of deaths in MVAs. Aortic injuries have a high morbidity and mortality; 90% of the patients die at the trauma scene, while 90% of initial survivors die within 4 months if the injury is undetected and untreated [38]. In the remaining 20% of cases, aortic injury is caused by falls and pedestrian injuries. Mechanisms of injury are variable and may overlap, including rapid deceleration, shearing forces, increased intravascular pressure caused by compression exceeding 2,000 mmHg (the water-hammer effect) or the “osseous pinch”, which represents direct compression of the aorta between the anterior chest wall and the spine [38]. The most common site of injury is the isthmus, corresponding to 90–95% of cases. Uncommon sites include the aortic root-ascending aorta, the aortic arch-branch vessels and the mid-distal descending aorta. The injury may be partial (65%), involving only the intima and media, or transmural (35%), also affecting the adventitia, which is lethal in almost all cases. The injury may be circumferential (45% of cases) or segmental (in 55% of cases), involving either the greater (Fig. 13) or the inferior curve (Fig. 14). CT may show direct and indirect signs. Direct signs include active contrast medium extravasation, dissection (Fig. 15), pseudoaneurysm (Figs. 13 and 14), intimal tear/flap (Fig. 16), thrombus protruding into the lumen (Fig. 17), and abrupt change in calibre (pseudocoarctation). Indirect CT signs are indistinctness of mediastinal flat planes, periaortic haematoma and mediastinal haematoma (Figs. 16‐18). Mediastinal haematoma is less than 20% predictive of aortic injury. In the absence of aortic injury, mediastinal haematoma may originate from venous injuries. In such cases the fat plane with the aorta is preserved, contrary to thoracic aortic injury, where haematoma develops in close contact with the aortic wall (Fig. 16‐18). Minimal aortic injuries affect only the intima. They are diagnosed with increasing frequency—because of improved MDCT technology—and constitute a diagnostic dilemma [39], as most of them remain stable or resolve on follow-up (Fig. 18). MDCT has very high sensitivity and specificity, reaching 98% and 100% accordingly for diagnosing aortic trauma. However, in studies where the presence of mediastinal haematoma (an indirect sign of aortic injury) was considered as a positive criterion for aortic injury, a significant number of false-positive diagnoses occurred, and the specificity for aortic injury dropped significantly by up to 62%. Nevertheless, MDCT has become the gold standard for ruling out aortic injury, and in those patients with unequivocal evidence of aortic injury, no further imaging is required [40]. No further workup is indicated if there is no direct evidence of aortic injury and no mediastinal haematoma on CT [38].

Soft tissue haematomas may occur during direct compression trauma when rib fractures cause laceration of veins or arteries. Soft tissue haematoma may become life-threatening if the patient is under anticoagulant therapy. If it is arterial in origin, embolisation is indicated. Breast haematomas can be serious in direct impact or compression injuries [24].

Rib fractures are the most common injury in blunt chest trauma, occurring in 50% of cases. A single rib fracture is usually not clinically significant, whereas multiple rib fractures indicate severe injury. Fractures of the first three ribs imply high-energy trauma that may be associated with injury of the brachial plexus or subclavian vessels. Fractures of the fourth up to the eighth ribs are the most common, while fractures of the last four ribs are usually associated with intra-abdominal injury. Reconstructed MIP and volume-rendered CT images depict with great detail the number and sites of rib fractures (Fig. 22). Flail chest is a marker of significant intrathoracic injury with increased morbidity, in which three or more contiguous ribs are fractured in two or more sites (Fig. 23). The diagnosis is clinical based on the paradoxical motion during respiration, which may result in ventilatory compromise. More than 50% of cases require surgical treatment and prolonged mechanical ventilation [9].

Sternal fractures have a prevalence of 3–8% in blunt chest trauma. The main mechanism is deceleration injury or a direct blow to the anterior chest wall. It is considered a marker of cardiac contusion (1.5%–6%). Sternal fractures are difficult to detect on lateral chest radiographs and even on axial CT images, as opposed to sagittal and coronal MDCT reformats, which have significant superiority (Fig. 24). It is almost always accompanied by anterior mediastinal haemorrhage, which has a preserved fat plane with the aorta (Fig. 25), as opposed to an anterior mediastinal haemorrhage secondary to aortic injury, which will present with a lost fat plane with the aorta [17].

Sternoclavicular dislocation is rare and occurs in 1–3% of all types of dislocation. Anterior sternoclavicular dislocation is more common and easily detectable, as it is palpable (Fig. 26). It usually has a benign course, but it implies a high-energy trauma and may be associated with haemopneumothorax, rib fractures or pulmonary contusion [43]. Posterior sternoclavicular dislocation is clinically and radiographically silent and carries serious morbidity, as it is associated with injuries of the mediastinal vessels, nerves, trachea and oesophagus (Fig. 27).

Scapular fracture is uncommon, occurring in 3.7% of cases of blunt chest trauma. It is easily detected on initial radiographs and may be masked clinically by other associated serious injuries (Fig. 28). It indicates a high-energy force trauma with a direct blow to the scapula or force transmitted through the humerus. Associated injuries are pneumothorax, haemothorax, clavicular fracture and injuries of the lung parenchyma, subclavian vessels, brachial plexus or spine [44].

Thoracic spine fractures account for up to 30% of all spine fractures. Sixty-two percent of spine fractures will result in neurological deficits. The most vulnerable site is between the ninth and twelfth vertebra. The main mechanism is hyperflexion and axial loading. Plain radiographs may miss fractures of the spine and therefore may be unnecessary in those patients scheduled for CT [45]. Sagittal and coronal MDCT reformats (Fig. 29) readily reveal even small spinal fractures, whereas volume-rendered images are not helpful [12]. MDCT of the spine is highly indicated for spinal survey for possible fractures. However, in the case of suspected compressive myelopathy, MRI is the method of choice (Fig. 30).

It is important to remember that multiple types of injury in a single patient may coexist, and radiologists should not be disorientated by depicting one type of trauma and neglect other coexisting or associated types of injury (Table 2). Therefore, systematic exclusion after thorough investigation of all sites of possible injury in the thorax is warranted.