Eur. Radiol. 10, 1524±1538 (2000) Ó Springer-Verlag 2000

European Radiology

Review article Imaging of blunt chest trauma S. Wicky, M. Wintermark, P. Schnyder, P. Capasso, A. Denys Department of Radiology, University Hospital, CHUV, CH-1011 Lausanne, Switzerland Received: 29 November 1999; Accepted: 28 January 2000

Abstract. In western European countries most blunt chest traumas are associated with motor vehicle and sport-related accidents. In Switzerland, 39 of 10,000 inhabitants were involved and severely injured in road accidents in 1998. Fifty two percent of them suffered from blunt chest trauma. According to the Swiss Federal Office of Statistics, traumas represented in men the fourth major cause of death (4 %) after cardiovascular disease (38 %), cancer (28 %), and respiratory disease (7 %) in 1998. The outcome of chest trauma patients is determined mainly by the severity of the lesions, the prompt appropriate treatment delivered on the scene of the accident, the time needed to transport the patient to a trauma center, and the immediate recognition of the lesions by a trained emergency team. Other determining factors include age as well as coexisting cardiac, pulmonary, and renal diseases. Our purpose was to review the wide spectrum of pathologies related to blunt chest trauma involving the chest wall, pleura, lungs, trachea and bronchi, aorta, aortic arch vessels, and diaphragm. A particular focus on the diagnostic impact of CT is demonstrated. Key words: Blunt chest trauma ± Thorax ± Injuries ± Multiple trauma ± Spiral CT ± Radiography

Introduction In western European countries blunt trauma of the chest largely exceeds penetrating traumas. According to the Swiss Office for Prevention of Accidents [1], the majority (90 %) of blunt chest injuries are due to motor vehicle accidents (MVA), falls or work-related accidents (7 %), and recreation activities (3 %). In most European countries the ratio between the number of severely injured patients and death related to MVA is in the range of 40:1. In Correspondence to: S. Wicky

Switzerland in 1998, alcohol consumption is responsible for 10.7 % of severely injured patients involved in traffic accident and 16 % of mortalities. Between 1980 and 1998 the number of severe injuries and deaths related to MVA dramatically dropped down (±15 %) from 33,572 to 28,387 casualties. This striking decrease relates to the introduction of active and passive means of protection, such as airbags, lateral reinforcement of the vehicles, and usage of front and rear seat belts. Swiss drivers were found to be wearing seat belts in 67 % of casualties in 1990 and in 74 % in 1998. The use of rear-seat belts became mandatory in 1994; however, only 34 % of rearseat injured patients were wearing the seat belts in 1997. The number of deaths per 1,000,000 vehicles in 1998 varies greatly in industrialized countries. The highest rate is found in France (286), followed by Spain (276), U. S. (206), Italy (179), Germany (174), Japan (149), Switzerland (138), Great Britain (134), and Sweden (122). In Switzerland a striking increase in the amount of the fines for speeding help to account for an additional 10 % reduction of MVA in 1996. The number of MVA-associated severely injured patients and mortality rate reach has reached that obtained in the early 1950 s, although the number of Swiss-registered vehicles increased from 251,952 in 1950, to 4,349,173 in 1998. In 1998, 28,387 Swiss were injured in traffic accidents. Injuries to the extremities, chest, abdomen, and head were encountered in 66, 52, 32, and 30 %, respectively. Severe thoracic trauma is associated with multiple injuries in 70±90 % of cases [2, 3]. Associated extrathoracic injuries involving the head, abdomen, and extremities are observed in 36, 33, and 35 %, respectively, in a reported series of 144 severe blunt chest traumas [3]. Fifty-nine percent of patients had one associated extrathoracic injury and 41 % had two associated extrathoracic injuries. Only 30 % of patients had blunt trauma limited to the chest. This data may be markedly modified with the large usage of front and lateral airbags, with the development of progressive seat-belt systems, and with the extensive improvements of the plastic distortion properties of the cars.

S. Wicky et al.: Imaging of blunt chest trauma

Plastic distortion of the front and rear portions of the chassis of cars is of major importance in the passive prevention. Indeed, the force level applied to passengers can reach 60±80 G during a frontal collision, according to the equation G = (0.036 ” V2)/D, for which V (kilometers per hour) is the speed of the vehicle at the time of the impact and D (centimeters) the plastic distortion of the car [4]. The time needed to transport patients to a trauma center greatly influences the outcome of the patient. Helicopters are routinely used in our country in cases of severe injuries due to motor vehicle, sport, and mountain accidents. The impact of helicopters was demonstrated by the decrease in the mortality rate during the Korean and Vietnam wars compared with World War II. Air transportation reduced the mortality rate per 100 casualties from 4.5 to 2.5 during the Korean conflict and to less than one during the Vietnam war [5]. Blunt chest trauma morbidity is associated with atelectasis, aspiration pneumonitis, lung contusions or lacerations, ARDS, recurrent pneumothorax, tracheobronchial tears, and cardiac or vascular lesions. In a series of 515 patients [6] morbidity from chest trauma was of 36 % with a mortality rate of 15.5 %. Associated shock and head injury, with Glasgow scores of 3±4, increased the mortality rate to 77 % [3]. Lung contusion appears to be one of the most important factors contributing to the increase of the morbidity and mortality rates of blunt chest trauma [7]. Computed tomography is being used more and more frequently in the evaluation of blunt trauma patients, and modern first aid trauma centers are now equipped with spiral CT (SCT) equipment placed in the immediate proximity of emergency rooms and intensive care units. Chest CT examination is usually part of a survey and is associated most of the time with head and abdominal examinations. Therefore, if requested, at our institution we begin with an examination of the head before any contrast media injection has been performed. Chest SCT survey, which requires intravenous contrast injection every time a vascular lesion of the mediastinum has to be ruled out, is obtained during the early vascular enhancement phase, in spiral mode acquisition. It is then followed by the abdominal study, 60±70 s after the beginning of the injection, during the portal enhancement phase. The entire CT data acquisition usually lasts 3±4 min in spiral-mode acquisition. For 4 months we have been using a multiple detector arrays SCT and acquisition time has dropped to less than 1 min. Blunt trauma of the chest wall Soft tissue lesions in blunt chest trauma patients are usually directly related to the severity of the trauma. Subcutaneous emphysema can be attributed to three conditions. The first condition involves a laceration of the parietal pleura due to rib fractures allows air to pass into the pleural space and then to leak into the adjacent soft tissues (Fig. 1). A tension pneumothorax facilitates this mechanism but is not necessary. The second condition

1525

occurs when a pneumomediastinum due to a bronchial rupture or tear moves upward into the deep compartments of the neck through the thoracic inlet (Fig. 2). The third condition is associated with a laryngeal or extrathoracic tracheal tear, with air extending downwards into the mediastinum and thoracic soft tissues (Fig. 3). Subcutaneous emphysema may be very subtle and must be carefully excluded. Indeed, it may even appear on plain films before a pneumothorax becomes visible. Conversely, extensive subcutaneous emphysema may cast the pectoral muscles and erase a severe underlying pneumothorax. Not uncommonly, subcutaneous emphysema may extend anteriorly and/or posteriorly to the contralateral side of the chest, mimicking bilateral pneumothoraces (Fig. 4). Occasionally, tension subcutaneous emphysema resulting from tension pneumothorax may require drainage in order to prevent hypodermic vascular compromise. Computed tomography is a highly sensitive method in detecting subcutaneous air collections, and frequently enables identification of the precise site of air leakage. Soft tissue hematomas are due to direct compression trauma or to laceration of veins and arteries of the chest wall from rib fractures, the latter of which requires percutaneous embolization of the involved vessels. Hematomas represent a life-threatening complication in anticoagulated trauma patients. Breasts are exposed to seat-belt compression and therefore large breast hematomas are not uncommon. Rib fractures are the most common lesions in blunt chest trauma patients. Although they occur in at least 50 % of cases at admission [8], they do not represent, in and of themselves, a life-threatening condition. Therefore, they require no treatment other than analgesia. Plain films of the chest, as well as additional rib views, are of limited value. It is admitted that initial admission X-ray films demonstrate only 40±50 % of rib fractures [8, 9]. They are mostly insensitive in the demonstration of costochondral fractures but demonstrate associated lesions such as pneumothorax, hemothorax, hemomediastinum, pulmonary contusions and flail chest [9, 10]. Careful scrutiny of CT scans may assist in identifying rib fractures. Flail chest is encountered when three or more consecutive ribs are involved (Fig. 5). This life-threatening condition causes paradoxical movements of the chest wall which may lead to severe respiratory failure, urging prompt adequate intensive respiratory therapy. Flail chests have been described in up to 37 % of cases with violent blunt thoracic traumas and have a very high incidence of associated lung, vascular, and cardiac lesions [2, 6]. Fractures of the first two ribs and clavicles are of special importance and are frequently unaccompanied by fractures of other ribs. They are almost always consecutive to severe trauma and are thus frequently accompanied with spinal fractures, tracheo-bronchial tears, or vascular injuries. These conditions must be ruled out with complementary SCT and/or angiographic procedures. Furthermore, fractures of the lower ribs may be frequently associated with spleen or liver trauma (Fig. 6).

1526

S. Wicky et al.: Imaging of blunt chest trauma

1a

1b

2

3

Fig. 1 a, b. A 44-year-old man involved in a motor vehicle accident (MVA). a Front scout view. b Spiral CT (SCT) axial section. Chest SCT depicts a right pneumothorax and a massive subcutaneous emphysema related to multiple right rib fractures. The 30 % right anterior pneumothorax is not detectable on the scout view or on the admission chest X-ray

4 Fig. 3. A 61-year-old patient involved in a MVA. Spiral CT section (3 mm) displays a complete detachment of the tracheal membrane (arrow), leading to an air leak surrounding the vascular structures and the esophagus (curved arrow) of the supra-aortic mediastinal compartment. Additionally, a right pneumothorax surrounds a right upper lobe contusion Fig. 4. A 43-year-old man involved in a recreational aircraft accident. Admission chest X-ray displays a right tension pneumothorax with complete lung collapse associated with subcutaneous emphysema. Rib fractures are not visible on this admission film. The subcutaneous emphysema extensively extends to the contralateral chest, which led to a false diagnosis of left pneumothorax, not confirmed on follow-up films

Fig. 2. A 44-year-old man involved in a MVA admitted with a severe pneumomediastinum and cervical emphysema related to a distal tracheal fracture. Neck CT, as part of a whole-body CT survey, displays extensive free air collections casting the hypopharynx, carotid artery (white arrow), and jugular vein (curved arrow), submaxillary glands (S), and cervical muscles, bilaterally

Fractures of the sternum are encountered in up to 10 % of severe trauma [10, 11], especially with the use of seat belts or by direct trauma from the steering wheel. Sternal fractures mostly occur 2 cm away from the manubrio-sternal joint. As for fractures of the first two ribs and clavicles, sternal fractures are frequently associated with severe intrathoracic lesions, as well as the internal mammary arteries or veins (Fig. 7). These fractures are almost always overlooked on supine plain films. They can be seen on coned-down lateral views, but such incidences are presently rarely obtained. Sternal fractures are easily demonstrated on SCT studies obtained for all blunt chest trauma patients at our institution. Sternal fractures are not uncommonly associated with large anterior mediastinal hematomas and cardio-

S. Wicky et al.: Imaging of blunt chest trauma

5

7a

1527

6

7b

Fig. 5. A 67-year-old male involved in a lateral car crash, with subsequent side-door syndrome. At admission, the patient was in respiratory distress and a paradoxical chest wall motion was identified. The chest X-ray discloses a striking flail chest, with serial bifocal posterior and lateral rib fractures. The mobile fractured ribs were responsible for the development of subcutaneous emphysema, as well as for extensive lung contusions Fig. 6. A 26-year-old motorcycle passenger admitted with a severe right flail chest and massive hypovolemic shock. Spiral CT section obtained after resuscitation and explorative laparotomy for bleeding displays a right posterior rib fracture (arrow) responsible for a large liver fracture involving segments 6±8 Fig. 7. A 51-year-old male patient involved as a car driver hit by a high-speed train. Admission whole-body SCT displays a left sterno-clavicular fracture responsible for left internal mammary artery and vein lacerations, leading to b active bleeding (arrow) and an anterior hemomediastinum

vascular lesions, such as myocardial contusions, hemopericardium, coronar tears, aortic lesions, and tracheobronchial tears. Scapulo-thoracic dislocation combined with sternoclavicular detachment or clavicle fractures, observed mainly in motorcycle accidents, is a condition which occurs during a posterior displacement and stretching of an upper limb (Fig. 8 a). Scapulo-thoracic dislocations, also called ª closed forequarter amputation of the upper limb,º must be recognized on plain films and CT, since they are associated, in 100 % of cases, to partial or complete brachial plexus injury [12], and tears or complete avulsions of thoracic inlet arteries (Fig. 8 b,c). These result in permanent neurologic compromise of the upper limb, in mediastinal and/or extrathoracic hemorrhage, and in vascular compromise of the upper limb. Chest films, in our experience, have a high specificity of up to 80 % and demonstrate the lateral displacement of the medial border of the scapula. This displacement of more than 2 cm from the spinous process line when compared with the opposite side must be present on at

1528

S. Wicky et al.: Imaging of blunt chest trauma Fig. 8 a±c. A 23-year-old male patient involved in a motorcycle accident. Vascular and neurological examination in the emergency room revealed a complete left brachial plexus palsy. a Admission chest X-ray demonstrates a left clavicle fracture. The medial border of the scapula is not recognized, and a left apical cap (arrow) is identified, which typically relates to an extrapleural blood collection. b Selective subclavian arteriography shows an occlusion of the vertebral (v), thyro-cervical (t), and subclavian (s) arteries, whereas the internal mammary (m) and dorsal scapular (d) arteries are preserved. c Myelogram obtained 10 days later reveals a complete avulsion of the left brachial plexus, with extensive sub-arachnoidal contrast material extravasation from C6 to T2. An intradural hematoma shifts the spinal cord to the right

a

c

selves, which lead to hypovolemia, shock, and to displacements and compressions of vessels and airways (Fig. 9). Blunt trauma of the pleura

b least three consecutive anteroposterior films. The sensitivity of this lateral displacement of the scapula is mediocre, in the range of 60 % [12, 13]. The CT sections obtained in patients with scapulo-thoracic dislocations demonstrate a large subscapular hematoma which is responsible for the lateral displacement of the scapula. Large intrathoracic, extrapleural hematomas can be depicted by plain films and CT, and can be readily differentiated from hemothoraces, due to their crescentshape configuration and apical localization. Associated vascular lesions in scapulo-thoracic dislocations include tears and complete thrombosis or dissection of the thoracic inlet vessels, mainly involving the subclavian and axillary arteries. Not uncommonly, smaller arteries, such as the thyrocervical trunk, left vertebral and circumflex arteries, can be damaged and lead to life-threatening intrathoracic, extrapleural hematomas, which extend centrally into the posterior and mid mediastinum and cranially along the jugular veins and carotid arteries and along the superficial mid and deep cervical fasciae, up to the base of the skull. These large hematomas are life-threatening conditions by them-

Pneumothorax is the second most common injury after rib fracture occurring in up to 40 % of patients [14]. Pneumothorax can be asymptomatic and indicates a disruption of the visceral pleura and is most commonly secondary to rib fractures; however, pneumothorax can occur without rib fractures or predisposing factors, particularly in infants and young adults. Gravity, lung recoil, and the anatomy of pleural space influence the X-ray appearance of pneumothoraces. On supine chest films, free air in the pleural space tends to accumulate anteriorly, making small collections very difficult to detect [14, 15]. Basic semiologic signs have been described by Tocino et al. [14]. For that purpose, chest films from 88 patients with 112 pneumothoraces obtained from 1697 trauma patients were reviewed. Results demonstrated that readers missed 30 % of pneumothoraces. Fifty percent of these were tension pneumothoraces. Pneumothoraces may occupy several recesses in the supine chest: the anteromedial recess is the earliest one to be filled, allowing the free air collection to underline both right and left heart borders; the superior vena cava; the right aspect of the ascending aorta; the aortic knob and the innominate and left subclavian arteries. The lateral border of the inferior vena cava and the paracardiac fat pad are also not uncommonly underlined by the free air collection (Fig. 10). Anteromedial pneumothoraces account for 30 % of Tocino et al.'s [14] series. The subpulmonic recess is the second most frequent site occupied by pneumothoraces, as demonstrated on

S. Wicky et al.: Imaging of blunt chest trauma

a

c

1529

b

d

30 % of supine films. The free air collection is trapped between the lower pulmonary lobes and the diaphragm. The subpulmonic recess represents an extent of the anteromedial one; thus, subpulmonic pneumothoraces are most of the time synchronous to anteromedial ones (Fig. 10). The apico-lateral recess is occupied on supine films by free air collections in 22 % of Tocino's series, but in less than 5 % of our cases. Posteromedial pneumothoraces account for 25 % of pleural space free air collections in Tocino et al.'s [14] series and for less than 1 % in our experience. Posteromedial pneumothoraces typically present as a lucent band outlining the mediastinal border leading to an increased delineation of the paraspinal line, descending aorta, and posterior costophrenic sulcus. It has a slightly triangular shape. The posteromedial pneumothorax should not be confused with a free air collection in the pulmonary ligament (Fig. 11), which is an infrequent condition and which is limited cranially by the hilum,

Fig. 9 a±d. A 38-year-old drug-abused male involved in a motorcycle accident. a Admission chest X-ray displays a lateral displacement of the right scapula, a fractured right clavicle, and a fracture of the first and second right posterior rib arches. The large right upper chest opacity mimicks a right upper lung atelectasis, which, according to b the subsequent CT examination, relates to c a large extrapleural hematoma extending to the mediastinum and cranially up to the skull base. d These radiological patterns are consistent and highly suggestive of scapulothoracic dissociation with thyreo-cervical trunk (arrow) damage confirmed by a selective brachio-cephalic arteriogram

having a convex lateral surface, rather than the triangular shape of the posteromedial air-field recess. Both conditions are readily identified and differentiated with CT. Small pneumothoraces presenting by trapped air within the minor fissure are, when carefully scrutinized, not uncommon in chest trauma patients [16]. They appear as sharply defined ovoid air collections whose long axis is parallel to the minor fissure. Another sign of pneumothorax is the deep lateral costophrenic sulcus sign [17]. Air collects in a sub-pulmonary location and supine plain films disclosed deep lateral costophrenic angle. The increasing number of CT examinations obtained in trauma patients enable detection of a high incidence of occult pneumothoraces compared with plain films [18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30]. Karaaslan et al. [20] showed that 20.5 % of chest studies obtained in trauma patients referred for brain injuries had an occult pneumothorax which were undetectable on plain

1530

10

S. Wicky et al.: Imaging of blunt chest trauma

11

12 a

Fig. 10. A 26-year-old male who fell from a 9-m height; persistent tension pneumothorax despite insertion of a large chest tube, which turned out to be occluded by blood clots. Chest X-ray displays free air pleural space collections, which fill the antero-medial (black arrow), subpulmonic (open arrow) and apico-lateral (oblique white arrow) recesses. At this level the air collection strikingly bulges laterally between the second and third ribs. The minor fissure (curved white arrow) also contains a linear air collection. The right lung condensation relates to a massive lung contusion and blood aspiration Fig. 11. A 28-year-old female patient involved in an MVA. Chest X-ray obtained after chest tube insertion and splenectomy discloses air in the left pulmonary ligament (black arrows), a left lower lobe laceration with a pneumatocele (white open arrow) Fig. 12 a, b. A 50-year-old pedestrian involved in an MVA. a Admission chest X-ray displays a right clavicle and comminutive scapula fractures. Lungs and mediastinum are unremarkable. b Spiral CT survey (3 mm) displays occult bilateral anterior pneumothoraces

films. For this reason, as a general rule of our institution, any patient undergoing cranial CT scanning for major trauma receives an SCT survey of the chest in order to identify pneumothoraces undetectable on plain films (Fig. 12). In these clinical circumstances, patients who may be candidates for surgery or mechanical ventilation may greatly benefit from prophylactic chest tube insertion. Tension pneumothorax is a particular form of pneumothorax, which occurs when the intrapleural pressure surpasses the atmospheric one. Tension pneumothorax is characterized by a partially or totally collapsed lung, a large hypolucent air collection, which lowers and flattens the ipsilateral diaphragm, opens the costophrenic angle and induces a contralateral shift of the heart and mediastinum (Fig. 13). This condition may be lifethreatening due to temponnade and impairment of the venous return to the right atrium (Fig. 14). Traumatic pleural fluid collections result from bleeding of the chest wall, mediastinum, or diaphragmatic blood vessels. Bleeding from the lungs is usually selflimited because of the low pulmonary arterial pressure and the large amount of pulmonary thromboplastin. Pleural effusions frequently appear several hours after

12 b trauma and thus one must repeat the chest X-ray when the diagnosis is clinically suspected. Pleural fluid collections, as large as 200 ml, usually are undetectable [31]. Increasing collections induce an increased density of the ipsilateral chest, already present in 91 % of patients when the fluid volume ranges from 200 to 500 ml. When the collection reaches 500±600 ml, the fluid accumulates in the subpulmonic recess and erases the diaphragm and the costophrenic angle. The classical apical cap represents the cephalic extension of the subpulmonic fluid collection. It is present in only 54 % of cases when the effusion is equal or larger than 1200 ml. Chylothorax [32, 33] comes from a laceration of the thoracic duct, usually at the level of its arch. It is associated with trauma of the medial end of the left clavicle in very rare case [34] and is more often due to a penetrating injury with direct section. Ultrasonography of the chest is the most sensitive method in characterizing, detecting and quantifying pleural fluid collections. However, small collections accumulated in the dependent portions of the pleural space are rarely accessible to the sonographic probe in the supine chest trauma patient.

S. Wicky et al.: Imaging of blunt chest trauma

13

14 a

Fig. 13. A 56-year-old male climber who fell from a cliff. Admission chest X-ray depicts a typical right tension pneumothorax with complete lung collapse, lowered right diaphragm, deep and wide lateral sulcus, as well as mediastinal and cardiac shifts. This patient presents a marked subcutaneous, mediastinal, and cervical emphysema, related to a tear of the right main bronchus, confirmed at bronchoscopy Fig. 14 a, b. An 83-year-old male patient involved as an unbelted driver in a front car crash and admitted with severe cardiac function impairment. a Admission chest X-ray discloses bilateral tension pneumothoraces, featuring a deep costo-phrenic sulcus sign on the right and a double diaphragm sign (arrows) on the left. b The impaired venous return to the heart was immediately relieved after bilateral chest tube insertion, allowing the heart to recover its normal configuration: Its transverse diameter increases by 2.5 cm, although the left pneumothorax is not yet completely relieved

Blunt trauma of the lung The most common lung lesions resulting from blunt trauma are pulmonary contusions and lacerations, atelectasis, and aspiration of gastric content, blood, or foreign bodies. Frequently, neither CT nor conventional chest films can differentiate among these various causes which all induce air-space consolidations. Lung contusion is the most frequent injury to the lung either directly or by contre-coup. Lung contusion describes lesions of the capillaries of the alveolar walls and septa leading to leakage of fluid and red cells into the alveolar spaces [35, 36]. Histologically the general architecture of the lung is preserved. Radiographically pulmonary contusions appear within 4±6 h after trauma of the chest with various patterns ranging from patchy, poorly defined infiltrates without segmental or sub-segmental distribution, to large areas of consolidation. Pulmonary contusions do not respect the fissures and can be widespread. Contusions change rapidly, from one film to the next, and generally resolve within 72 h and clearing in 3±5 days. Large contusions need more time to resolve completely, up to 2 or 3 weeks [37, 38]. Computed tomography patterns are not specific, even if a lucent band has been described as pathognomonic, lying

1531

14 b between the pleura and the periphery of the air-space consolidation [39]. Admission SCT frequently detects in chest blunt trauma patients faint peripheral infiltrates, which retrospectively turn out to belong to pulmonary contusions in their early stage. Here they are still silent and will sometimes become apparent on plain films [22]. Presently, in our daily practice, we consider lung contusions as peripheral air-space consolidations adjacent to soft tissue hematomas and/or rib fractures. Plain films, showing evidence of lung contusions, are suggestive of a violent blunt chest trauma, and all these patients should undergo SCT to detect other associated mediastinal injuries [40] or occult pneumothoraces. With more important blunt trauma to the lung, lung parenchyma disruption or lung laceration occurs [41], leading to the formation of a space-occupying lesion filled with blood designated by air or both. When the lesion is completely filled with air, it is frequently named pneumatocele. Large areas of lung contusion always circumscribe and erase lung lacerations on initial chest films. After 2±4 days, air space consolidations associated with lung contusions regress or disappear and lacerations then become visible, as well-defined masses or pneumotoceles, sometimes with an air±fluid level on erect or semi-erect films. Most of the time, CT sections enable one to obtain early evidence of lung lacerations, which can be readily differentiated from the surrounding areas of lung contusion (Fig. 15). Lung lacerations do not require any treatment. They persist for weeks or even months until they completely resolve, without evidence of scar. However, radiologists should be aware of the high complication rate following blunt pulmonary contusions and lacerations such as adult respiratory distress syndrome and septic complications. Blunt trauma of the trachea and bronchi Tracheo-bronchial ruptures are observed in 0.7 % of blunt chest trauma admission films and in up to 2.8 % of autopsy series [42, 43]. The site of the rupture includes either the main-stem bronchi or the membranous

1532

a

S. Wicky et al.: Imaging of blunt chest trauma

c

b portion of the distal end of the trachea, 2±3 cm from the carina. The mechanisms of rupture are complex and associated with shearing forces on the tracheo-bronchial tree and abrupt increase of the intrathoracic pressure against a closed glottis (Fig. 16). Up to 68 % of tracheobronchial injuries are overlooked during the first days on plain-film examinations [44], even if, retrospectively with close scrutiny, the diagnosis can be suspected on plain films. Small tears of the tracheo-bronchial tree are not infrequently diagnosed 2±6 weeks after the blunt trauma, when they appear as a scar or stricture reducing the airway diameter and sometimes leading to a lung collapse [44] or to recurrent episodes of pneumonia. The chest film and CT patterns of tracheo-bronchial ruptures vary depending on the presence of a pneumomediastinum, which is associated, in 80 % of cases, with a concomitant ipsilateral pneumothorax [45]. Pneumomediastinum alone occurs when the lesion is located medially with respect to the position of the pulmonary ligament. Pneumothorax alone occurs when the lesion involves the main bronchus, distal to the insertion of the pulmonary ligament. Cervical subcutaneous emphysema represents the cranial extension of a pneumomediastinum, which must always, in any blunt trauma patients, evoke the possibility of a tracheo-bronchial tear.

Fig. 15 a±c. A 34-year-old male patient involved in an MVA. a Admission chest X-ray displays multiple areas of pulmonary contusions, with no evidence of pulmonary laceration. b Chest SCT obtained few minutes later shows four areas of pulmonary lacerations in the left upper lobe, three of them filled by blood (arrows). The largest and fourth one, measuring 2.5 cm, features an air±fluid level. Lacerations are surrounded by areas of ground-glass attenuation, representing lung contusions. c Supine film obtained 4 days after admission shows a fading of the multiple areas of lung contusion and appearance of bilateral well-circumscribed lucencies. The left lacerations described in b became pneumatoceles

In case of complete main-stem bronchial rupture, a ªfallen lungº can be observed on plain films and CT [46]. This sign relates to a totally collapsed lung, falling down into the dependent portion of the chest cavity, usually posteriorly and inferiorly, its remaining attachments to the hilum being the pulmonary arteries and veins. Infrequently, SCT directly demonstrates of a tracheal or bronchial rupture, usually associated with a large mediastinal or intrapulmonary hematomas creating a large gap between the main stem bronchus and the lobar and segmental bronchi. Whatever the chestfilm and CT patterns, any blunt trauma patient presenting a persistent pneumothorax despite adequate positioning of a chest tube, or presenting an increasing pneumomediastinum and/or subcutaneous cervical emphysema, should undergo bronchoscopy to rule out such a lesion. Blunt trauma of the esophagus Esophageal rupture is reported to occur in 1 % of blunt trauma patients, mainly in those who previously sustained iatrogenic perforation or penetrating trauma [47, 48]. In our experience esophageal rupture is extremely rare and occurs in less than 1 per 1000 cases of blunt chest trauma. Fifty-two percent of esophageal ruptures involved the cervical or upper thoracic esophagus [48]. Esophageal ruptures are due to various mechanisms including compression of the esophagus between the thoracic spine and the sternum and an increased esophageal pressure due to a violent gastric reflux against a tight cricopharyngeal muscle. The high mortality rate associated with esophageal ruptures is due to their diagnostic delay and the rapid development of severe mediastinitis. Plain films as well as CT are unable to directly visualize esophageal ruptures. They only display indirect signs of rupture, which include mediastinal hematoma, pneumomediastinum, and pneumothorax. Most of the time, early recognition of esophageal ruptures is achieved by esophagoscopy.

S. Wicky et al.: Imaging of blunt chest trauma

1533

Air leak arising from alveolar ruptures extends in a centripetal direction along the peribronchovascular sheaths into the mediastinum (Fig. 17). This mechanism, experimentally demonstrated 60 years ago [51], was only casually reported on plain films and SCT studies [52, 53]. Blunt trauma of the aorta

Fig. 16. A 44-year-old male car driver involved in a high-speed MVA. Admission CT displays a rupture of the left antero-lateral aspect of the distal cervical trachea with a large air leak. The breach (white arrow) is enlarged by the inflated balloon cuff of the endotracheal tube

Fig. 17. A 26-year-old blunt trauma female involved in a horse ridding fall. Spiral CT section (3 mm) demonstrates a Macklin effect featuring linear air collections (arrows) following the course of the pulmonary vessels and entering into the mediastinum

Rarely water-soluble contrast series of the upper gastrointestinal tract can be considered as a diagnostic tool in severely injured trauma patients. Blunt chest trauma and pneumomediastinum Pneumomediastinum is defined as free air collections surrounding mediastinal arteries and veins, trachea, esophagus, and dissecting the mediastinal fat towards the neck. Pneumomediastinum occurs in up to 10 % of blunt chest trauma and, in less than 2 %, is due to tracheo-bronchial ruptures. Detection of pneumomediastinum on a supine chest film or with CT requires immediate bronchoscopy to rule out or confirm the diagnosis of a tracheo-bronchial lesion. In more than 95 % of cases, a pneumomediastinum results from alveolar rupture, positive pressure mechanical ventilation or both [49, 50].

Aortic injury in blunt thoracic trauma has a very dismal prognosis. It occurs most frequently during motor vehicle accidents with high-speed deceleration or chest compression. It is responsible for approximately 16 % of immediate deaths. Eighty-five to 90 % of patients with aortic injury die prior to admission [54, 55]. Survivals have an increased mortality rate of 1±2 % per hour after the onset and only 2 % of untreated patients survive beyond 4 months [54, 55]. Blunt aortic injury is a persistent major problem as the overall mortality was 31 % in a multi-center trial involving 50 trauma canters where 274 aortic injuries were admitted [56]. Major recent improvements in diagnostic modalities, which include SCT and transoesophagal echography (TEE), enable surgeons to screen patients for immediate or delayed surgery [57], or even to apply a conservative treatment. The radiological patterns of aortic lesions vary according to their resulting hemomediastinum. In 87.5 % of cases, however, the hemomediastinum has been shown to arise from small venous tears, and vertebral or first-rib fractures [58]. The mediastinal enlargement may also be due to a fatty mediastinum, frequently observed in obese patients or to the distention of the venous mediastinal compartment after rapid intravenous over-hydration. All these conditions account for the low positive predictive value of chest films obtained on chest trauma patients. Conversely, an unequivocal unenlarged mediastinum has a 98 % negative predictive value for an aortic lesion [55, 59, 60]. Alone on chest plain films, indirect signs of hemomediastinum, which include right tracheal displacement, a depressed left bronchus, a blurring of the aortic knob, or a fracture or the first ribs, have a poor positive predictive value [55, 59, 60]. Until recently, aortography, and mainly digital aortography, with a sensitivity and specificity of 98 % [61], was considered as the gold standard modality. Less invasive procedures, such as CT and SCT aortography, are now considered as efficient in the examination of the thoracic aorta [59, 62, 63]. The SCT criteria for the assessment of aortic trauma must consider the presence of a hemomediastinum, a peri-aortic hematoma, an aortic pseudodiverticulum, an irregular wall or a fuzzy contour of the opacified thoracic aorta, and an intimal flap (Fig. 18). Several studies [59, 60, 62, 64], including ours [40], have shown that a normal mediastinum on conventional or spiral CT has a very high negative predictive value for aortic trauma, in the range of 99.3±99.9 %. Thus, SCT is now recognized not only as a screening tool in many trauma centers, but also as a worthy diagnostic procedure which enables one to decrease the number of unnecessary aortographies and their related costs. Aortog-

1534

Fig. 18. A 47-year-old male climber who fell from a 50-m cliff. Admission SCT displays typical patterns of injured aorta at the isthmic level, featuring an unequivocal peri-aortic hematoma, an outpouching isthmic pseudodiverticulum, and an intimal flap

raphy is reserved for patients with equivocal hemomediastinum or when an obvious hemomediastinum does not clearly relate to an aortic injury on SCT sections. At our institution, data based on 18 proven aortic ruptures provided by the SCT chest examination of 695 blunt trauma patients are in agreement with the data of the literature and gives SCT a sensitivity and specificity of 100 and 99.8 %, respectively [40, 63, 65]. The SCT contrast medium injection and data acquisition have already been described in detail [40]. The requested parameters include a 3-mm X-ray beam collimation, a pitch value of 1.5±2.1, and reconstruction intervals of 1.5 mm. The diagnosis of aortic lesions is assessed mainly by axial sections, whereas the use of 2D and 3D surface rendering maximal intensity projection reconstructions are no longer recommended. Transoesophagal echography is another accurate modality for the demonstration of aortic injury; however, TEE requires the presence of a well-trained cardiologist or anesthesiologist, and in case of severe facial trauma or cervical spine fracture, the procedure cannot be performed. Furthermore, most blunt chest trauma patients require cranial and/or abdominal CT scanning, and the additional time needed by the aortic protocol does not increase the total time of the SCT procedure. Blunt trauma of aortic arch vessels Blunt injuries of the aortic arch vessels usually occur at their origin [66], leading to a large hemomediastinum among survivals. Most of the time, they appear as isolated vascular lesions in motorcycle accidents, associated with complete or partial brachial plexus syndrome and scapulo-thoracic dissociation (Fig. 8). Innominate artery injuries are the cause of a widened mediastinum on supine chest film. This predominates on the right side, displacing the trachea to the left. Blunt

S. Wicky et al.: Imaging of blunt chest trauma

injury of the left subclavian artery in its intrathoracic course leads to supra -aortic and mid-hemomediastinum, which displaces the trachea to the right and blurs the aortic knob. More distally, in their extrapleural course, lesions of subclavian and axillary arteries produce large thoracic-outlet hematomas and apical cap collections (Fig. 9). Isolated injuries of the vertebral arteries are not uncommon in scapulo-thoracic dissociation and lead to similar radiological patterns. Supine plain films can most of the time give a reliable demonstration of intra- and extra-thoracic hematomas resulting from aortic arch vessels lesions. Spiral CT routinely performed in these instances at our institution easily depicts the hematoma, and the involved large vessel, but most of the time axial sections, as well as 2D and 3D reformatted images, are not sufficient to precisely demonstrate the lesion for surgical repair. Semi-selective and superselective angiograms are mandatory in these situations. Internal mammary artery or vein lesions regularly occur in injured blunt chest trauma, when associated with sternal fractures, as a consequence of seat-belt compression or direct shock from the steering wheel (Fig. 7). Resulting large anterior hemorrhages may escape supine chest film, but not CT examinations, except in infants and young adults, for whom the anterior mediastinal hemorrhage sometimes cannot be differentiated from the thymus. These anterior mediastinal hematomas can also barely be differentiated from those resulting from lesions of the root of the ascending aorta, from right atrial and ventricle ruptures, and from coronary artery lesions. Blunt trauma of the heart Cardiac injury is rarely clinically manifest, but cardiac contusions occur in up to 76 % of blunt chest trauma [67, 68, 69, 70]. The right ventricle lays anteriorly in the chest, behind the sternum, and is thus more commonly affected than the left one, resulting in transient wall abnormalities on echocardiography, without cardiac output modification [71]. Arrythmias and conduction defects account for the most common clinical manifestations [68]. Myocardial infarct [72] is observed in cases of coronary dissection, thrombosis, or plaque rupture. It evolves similarly to non-traumatic cardiac infarcts and complications such as aneurysms, mural thrombosis, or papillary muscle ruptures can be observed [73, 74, 75]. Direct chordae tendinae rupture is very rare but can occur due to direct leaflet injury. Myocardial laceration or rupture may cause leakage into the pericardium and induce tamponade [76, 77]. Most of the time, cardiac muscle laceration is lethal. Chest plain film is of no value for the demonstration of cardiac injury. Spiral CT can depict a pericardial collection [78] and eventually ventricular wall injury associated with a mediastinal hemothorax; however, no time should be wasted in performing transthoracic or TEE (Fig. 19) [79]. If pericardic fluid is demonstrated, emergency drainage must be considered without delay. Indeed, the amount of fluid necessary to produce an

S. Wicky et al.: Imaging of blunt chest trauma

1535 Fig. 19 a±c. A 64-year-old male truck driver involved in a frontal crash at 100 km/h. a A CT survey (8 mm) displays a large hematic collection, mimicking a hemomediastinum. b A 200-ml hemopericardium. c Computed tomography was immediately followed by TEE, which ruled out an aortic lesion but demonstrated a partially clotted hemopericardium (curved open arrow) leading to a compression of the right atrium (white arrow). Cardiac tamponade justified immediate surgery during which the surgeon identified and sutured a myocardial rupture located at the level of the left atrial appendage

a

b acute cardiac tamponade may be as small as 250 ml, although in cases of chronic pericardial effusion, a volume of 1000 ml is still tolerated. Since pericardial effusion in supine patients preferentially accumulates in the nondependent portion of the peripheral sac, this allows a safe approach to percutaneous drainage under sonographic guidance. Angiography and echocardiography remain the diagnostic devices of choice in ruling out lesions of coronary arteries and left ventricular function. Magnetic resonance imaging is still mostly not considered with severe blunt trauma patients, but cardiac MRI may now be used in the subacute phase among survivals for the demonstration of the extent of myocardial contusions as well as cardiac and valvular functions. Blunt trauma of the diaphragm Diaphragmatic disruption occurs in 0.8±8 % of blunt trauma [80, 81, 82, 83, 84]. They are more frequently observed in abdominal than in chest trauma. Left diaphragmatic injuries account for 77±90 % of cases (Fig. 20), due to the protection accorded by the liver [6, 81, 83, 85, 86]. Tears most frequently involve the muscular posterior and postero-lateral aspects of the diaphragm. When not recognized, the mortality rate of diaphragmatic ruptures is in the range of 30 % [82, 83, 86] due to necrosis and incarceration of abdominal viscera

c which herniate into the chest. Furthermore, diaphragmatic ruptures are almost always associated with other severe life-threatening abdominal or thoracic conditions [80, 81, 82, 87, 88, 89]. More than 50 % of diaphragmatic ruptures are missed on plain films during the first 3 or 4 days following admission [87, 90]. In fact, in 77 % of cases, chest X-rays present only non-specific abnormalities, including pleural effusion, lung atelectasis, lung contusions, and a partially blurred, irregular, and elevated diaphragm [90]. More specific radiological patterns, such as herniated liver, spleen, kidney, bowel loops, and stomach (Fig. 21), including a nasogastric tube, are observed in approximately 20 % of diaphragmatic ruptures. Spiral CT is the quickest diagnostic tool in assessing diaphragmatic rupture. Axial sections are able to demonstrate tears in the diaphragm, which occur most of the time in the viscidity of the diaphragmatic crura. The resulting hemothoraces and/or retroperitoneal hemorrhages are identified in 73±82 % of cases [86, 90]. Axial CT sections also enable visualization of thoracic herniation of abdominal viscera and peritoneal fat in 55 % of cases [86, 90]. The reported sensitivity and specificity of SCT in this field are 61 and 85 %, respectively. Coronal and sagittal 2D reformatted images have been shown to improve diagnostic performances of SCT with a sensitivity and specificity of 92 and 87 %, respectively [83]. Magnetic resonance imaging can be considered only in stable patients [84, 91, 92]. T1-weighted images give

1536

S. Wicky et al.: Imaging of blunt chest trauma

Fig. 20. A 37-year-old male patient involved as a belted driver in a truck accident. Spiral CT section (3 mm) shows an intrathoracic herniation of the stomach (s) and a gap in the left posterior diaphragm suggestive of diaphragmatic rupture (arrows)

a precise demonstration of the diaphragm in the sagittal and coronal plans. It appears as a hypointense homogeneous band underlined by the liver intensity on the right and by peritoneal and mediastinal fat on the left. The access to MR remains, however, difficult in most patients presenting a blunt diaphragmatic injury, due mainly to the associated life-threatening lesions. Diagnostic assessment of penetrating diaphragmatic injuries by video-thoracoscopy has been successfully reported [93] and represents a worthwhile alternative to other imaging modalities. Furthermore, video-thoracoscopy also represents a minimally invasive therapeutic modality. Ultrasonography is, in our experience, a poor diagnostic modality and should no longer be recommended to display diaphragmatic ruptures. Conclusion Supine films of the chest still remains the most efficient diagnostic modality for all chest trauma patients. It gives the precise diagnosis for most life-threatening lesions involving the chest wall, pleura, lung, mediastinum, and diaphragm. Its limited sensitivity, when compared with SCT, relates to lesions which do not immediately threatened the patient's condition, such as occult pneumothorax, lung contusion, or pleural collections which are smaller than 200 ml. However, normal chest X-ray film have a very high negative predictive value, especially for ruling out intrathoracic vascular lesions. Spiral CT examination of the chest obtained in blunt trauma patients, usually performed during the same session as abdominal CT, has unequaled sensitivity and specificity. They give diagnostic information which is of great help for the better comprehension of the admission chest X-ray and additional films which will be obtained later.

Fig. 21. A 30-year-old female patient in a motorcycle accident. Admission chest X-ray obtained prior to death displays a large herniation of the stomach into the thoracic compartment. At autopsy, the stomach as well as the spleen, the left kidney, and the left splenic flexure of the colon were herniated through the ruptured diaphragm. This supine film also displays extensive subcutaneous, mediastinal, and cervical emphysema, a right lung contusion, a large right hemopneumothorax, and multiple rib fractures

Spiral CT angiography is now the method of choice at our institution in demonstrating aortic lesions. Arteriography still remains the modality of choice for the evaluation of aortic arch vessel damage. Percutaneous chest sonography has little place in blunt chest trauma when compared with its role for the assessment of abdominal lesions. Magnetic resonance imaging remains a second procedure mostly reserved for complementary evaluation of blunt lesions in the subacute phase. References 1. Burckhard J, Engler C, Salinas L (1999) La santØ publique en Suisse, prestations, coßts, prix. Pharma Information, Bâle 2. Trupka A, Waydhas C, Hallfeldt KKJ, Nast-Kolb D, Pfeifer KJ, Schweiberer L (1997) Value of thoracic computed tomography in the first assessment of severe injured patients with blunt chest trauma: result of a prospective study. J Trauma 43: 405±412 3. Pinilla JC (1982) Acute respiratory failure in severe blunt chest trauma. J Trauma 22: 221±226 4. Eckert WG (1977) Crash injuries on the road. In: Tedeschi CG et al. (eds) Forsenic medicine, vol 2: a study in trauma and environmental hazards. Saunders, Philadelphia, pp 853±864 5. Baxt WG, Moody P (1983) The impact of a rotocraft aeromedical emergency care service on trauma mortality. J Am Med Assoc 249: 3047±3051

S. Wicky et al.: Imaging of blunt chest trauma 6. Shorr RM, Crittenden M, Indeck M, Hartunian SL, Rodriguez A (1987) Blunt thoracic trauma. Ann Surg 206: 200±205 7. Johnson JA, Cogbill TH, Winga ER (1986) Determinants of outcome after pulmonary contusion. J Trauma 26: 695±697 8. Deluca SA, Rhea JT, O'Malley TO (1982) Radiographic evaluation of rib fractures. AJR 138: 91±92 9. Thompson BM, Finger W, Tonsfeldt D, Aprahamian C, Troiano P, Hendley G, Mateer J, Stueven H (1986) Rib radiographs for trauma: useful or wasteful? Ann Emerg Med 15: 261±265 10. Roy-Shapira A, Levi I Khoda J (1994) Sternal fractures: a red flag or a redherring? J Trauma 37: 59±61 11. Gazak S, Davidson SJ (1984) Posterior sternoclavicular dislocations: two case reports. J Trauma 24: 80±81 12. Oreck SL, Burgess A, Levine AM (1984) Traumatic lateral displacement of the scapula: a radiographic sign of neurovascular disruption. J Bone Joint Surg 66A:758 13. Rubenstein JD, Ebrahein NA, Kellam JF (1985) Traumatic scapulothoracic dissociation. Radiology 157: 297 14. Tocino IM, Miller MH, Faifax WR (1985) Distribution of pneumothorax in the supine and semirecumbent critically ill adult. AJR 144: 901±905 15. Lacombe P, Cornud F, Grenier P, Nahum H (1982) A new sign of right anterior pneumothorax in the supine adult. Ann Radiol 25: 231±236 16. Spizarny D, Goodman LR (1986) Air in the minor fissure: a sign of right-sided pneumothorax. Radiology 160: 329±331 17. Gordon R (1980) The deep sulcus sign. Radiology 136: 25±27 18. Bridges KG, Welch G, Silver M, Schinco MA, Esposito B (1993) CT detection of occult pneumothorax in multiple trauma patients. J Emerg Med 11: 179±186 19. Wolfman NT, Gilpin JW, Bechtold RE, Meredith JW, Ditesheim JA (1993) Occult pneumothorax in patients with abdominal trauma: CT studies. J Comput Assist Tomogr 17: 56±59 20. Karaaslan T, Meuli R, Androux R, Duvoisin B, Hessler C, Schnyder P (1995) Traumatic chest lesions in patients with severe head trauma: a comparative study with computed tomography and conventional chest roentgenograms. J Trauma 39: 1081±1086 21. McGonigal MD, Schwab CW, Kauder DR, Miller WT, Grumbach K (1990) Supplemental emergent chest computed tomography in the management of blunt torso trauma. J Trauma 30: 1431±1434 22. Poole GV, Morgan DB, Cranston PE, Muakkassa FF, Griswold JA (1993) Computed tomography in the management of blunt thoracic trauma. J Trauma 35: 296±303 23. Enderson BL, Adballa R, Frame SB, Casey MT, Gould H, Maull KI (1993) Tube thoracostomy for occult pneumothorax: a prospective randomized study of its use. J Trauma 35: 726±729 24. Garramone RR, Jacobs LM, Sahder P (1991) An objective method to measure and manage occult pneumothorax. Surg Gynecol Obstet 173: 257±261 25. Engdahl D, Toft T, Boe J (1993) Chest radiograph. A poor method for determining the size of pneumothorax. Chest 103: 26±29 26. Tocino I, Miller MH, Frederick PR, Bahr AL, Thomas F (1984) CT detection of occult pneumothorax in head trauma. AJR 143: 987±990 27. Tocino I, Miller MH (1987) Computed tomography in blunt chest trauma. J Thorac Imaging 2: 45±59 28. Blostein PA, Hodgman CG (1997) Computed tomography of the chest in blunt thoracic trauma: results of a prospective study. J Trauma 43: 13±18 29. Haramati LB, Hochsztein JG, Marciano N, Nathanson N (1996) Evaluation of the role of chest computed tomography in the management of trauma patients. Emerg Radiol 3: 225±230 30. Seow A, Kazerooni EA, Cascade PN, Pernicano PG, Neary M (1995) Comparison of upright inspiratory and expiratory chest radiographs for detecting pneumothoraces. AJR 166: 313±316

1537 31. Woodring JH (1984) Recognition of pleural effusion on supine radiographs: How much fluid is required? AJR 142: 59±64 32. MacFarlan JR, Holman CW (1972) Chylothorax. Am Resp Dis 105: 287 33. Schnyder P, Hauser H, Moss AA,Gamsu G, Brasch R, Bohnet J, Candardjis G (1983) CT of the thoracic duct. Eur J Radiol 3: 18±23 34. Dulchavsky SA, Ledgerwood AM, Lucas CE (1988) Management of chylothorax after blunt chest trauma. J Trauma 28: 1400±1401 35. Fulton RL, Peter ET (1970) The progressive nature of pulmonary contusion. Surgery 67: 499±506 36. Fulton RL, Peter ET (1974) Compositional and histologic effects of fluid therapy following pulmonary contusion. J Trauma 14: 783±790 37. Schnyder P, Essinger A (1991) Radiologic assassment. In: Webb WR, Besson A (eds) Thoracic surgery: surgical management of chest injuries. Mosby, St. Louis, pp 108±122 38. Crawford WO Jr (1973) Pulmonary injury in thoracic and nonthoracic trauma. Radiol Clin North Am 11: 527 39. Donnelly LF, Klosterman LA (1997) Subpleural sparing: a CT finding of lung contusion in children. Radiology 204: 385±387 40. Wicky S, Capasso P, Meuli R, Fischer A, Segesser L von, Schnyder P (1998) Spiral CT aortography: an efficient technique for the diagnosis of traumatic aortic injury. Eur Radiol 8: 828±833 41. Reuter M (1996) Trauma of the chest. Eur Radiol 6: 707±716 42. Kemmerer WT, Eckert WG, Gathright JB, Reemtsma K, Creech O (1961) Patterns of thoracic injuries in fatal traffic accidents. J Trauma 1: 595±599 43. Bertelsen S, Howitz P (1972) Injuries of the trachea and bronchi. Thorax 27: 188±194 44. Taskinen SO, Salo JA, Halttunen PEA, Sovijarvi ARA (1989) Tracheobronchial rupture due to blunt chest trauma: a followup study. Ann Thorac Surg 48: 846±849 45. Cordice J, Cabazon J (1965) Chest trauma with pneumothorax and hemothorax. J Thorac Cardiovasc Surg 50: 316±333 46. Kumpe DA, Oh KS, Wyman SM (1970) A characteristic pulmonary finding in unilateral complete bronchial transection. AJR 110: 704±706 47. Thompson TC, Enderson BL (1993) Distal esophageal rupture following blunt trauma. J Tenn Med Assoc 86: 13±14 48. Maroney MJ, Mirvis SE, Shanmuganathan K (1996) Esophageal occlusion caused by thoracic spine fracture or dislocation: CT diagnosis. AJR 167: 714±715 49. Cirlak D, Milne ENC, Imray TJ (1984) Pneumomediastinum: a diagnostic problem. Crit Rev Diagn Imaging 23: 75±117 50. Groskin SA (1996) Selected topics in chest trauma. Semin Ultrasound CT MR 17: 119±141 51. Macklin CC (1939) Transport of air along sheaths of pulmonic blood vessels from alveoli to mediastinum: clinical implications. Arch Intern Med 64: 913±926 52. Ovenfors CO (1964) Pulmonary interstitial emphysema: experimental roentgen diagnostic study. Acta Radiol 224 [Suppl]: 1±131 53. Wintermark M, Wicky S, Schnyder P, Capasso P (1999) Blunt traumatic pneumomediastinum: CT demonstration of the Macklin effect. AJR 172: 129±130 54. Parmley LF, Mattingly TW, Manion WC, Jahnke EJ (1958) Non-penetrating traumatic injury of the aorta. Circulation 17: 1086±1101 55. Mirvis SE, Bidwell JK, Buddemeyer EU, Diaconis JL, Pais SO, Whitley JE, Goldstein LD (1987) Value of chest radiography in excluding traumatic aortic rupture. Radiology 163: 487±493 56. Fabian TC, Richardson JD, Croce MA, Smith JS, Rodman G, Kearney PA, Flynn W, Ney AL, Cone JB, Luchette FA, Wisner DH, Scholten DJ, Beaver BL, Conn AK, Coscia R, Hoyt DB, Morris JA, Harviel JD, Peitzman AB, Bynoe RP, Diamond DL, Wall M, Gates JD, Asensio JA, Enderson BL et al. (1997) Prospective study of blunt aortic injury: multicenter trial of the American Association for the surgery of trauma. J Trauma 42: 374±380

1538 57. Galli R, Pacini D, Di Bartolomeo R, Fattori R, Turinetto B, Grillone G, Pierangeli A (1998) Surgical indications and timing of repair of traumatic rupture of the thoracic aorta. Ann Thorac Surg 65: 461±464 58. Sandor F (1967) Incidence and significance of traumatic mediastinal haematoma. Thorax 22: 43±62 59. Raptopoulos V, Sheiman RG, Philips DA, Davidoff A, Silva WE (1992) Traumatic aortic tear:screening with chest CT. Radiology 182: 667±673 60. Morgan PW, Goodman LR, Aprahamian C, Foley WD, Lipchik EO (1992) Evaluation of traumatic aortic injury: Does dynamic contrast-enhanced CT play a role? Radiology 182: 661±666 61. Mirvis SE, Pais SO, Shanmuganathan K (1994) Atypical results of thoracic aortography performed to exclude aortic injury. Emerg Radiol 1: 24±31 62. Fisher RG, Chasen MH, Lamki N (1994) Diagnosis of injuries of the aorta and brachiocephalic arteries caused by blunt chest trauma:CT vs aortography. AJR 162: 1047±1052 63. Gavant ML, Flick P, Menke P, Gold RE (1996) CT aortography of thoracic aortic rupture. AJR 66: 955±961 64. Miller FB, Richardson JD, Thomas HA, Cryer HM, Willing SJ (1989) Role of CT in diagnosis of major arterial injury after blunt thoracic trauma. Surgery 106: 596±603 65. Mirvis SE (1997) Major vascular injury in trauma: influence of new technology. Eur Radiol 7:S278 66. Mattok KL (1988) Thoracic vascular trauma. J Vasc Surg 7: 725±729 67. Leinoff HD (1940) Direct non-penetrating injuries to the heart. Ann Intern Med 14: 653 68. Siegler LH (1945) Traumatic injury of the heart. Am Heart J 30: 459±478 69. Bodin L, Rouby JJ, Viars P (1988) Myocardial contusion in patients with blunt chest trauma as evaluated by thallium 201 myocardial scintigraphy. Chest 94: 72±76 70. Helling TS, Duke P, Beggs CW (1989) A prospective evaluation of 68 patients suffering blunt chest trauma for evidence of cardiac injury. J Trauma 29: 961±965 71. Hiatt JR, Yeatman LA, Child JS (1988) The value of echocardiography in blunt chest trauma. J Trauma 28: 914±922 72. Pringle S, Davidson K (1987) Myocardial infarction caused by coronary artery damage from blunt chest injury. Br Heart J 57: 375±376 73. Grieco JG, Montoya A, Sullivan HJ (1989) Ventricular aneurysm due to blunt chest injury. Ann Thorac Surg 47: 322±329 74. Matthews TR, French WJ, Criley JM (1989) Chest trauma and subvalvular left ventricular aneurysms. Chest 95: 474±475 75. Hilton T, Mezei L, Pearson AC (1990) Delayed rupture of tricuspid papillary muscle following blunt chest trauma. Am Heart J 119: 1410±1412 76. Layton TR, Dimarco RF, Pellegrins RV (1979) Rupture of left atrium from blunt trauma. J Trauma 19: 117±118

S. Wicky et al.: Imaging of blunt chest trauma 77. Tobin HM, Hiratzka LF, Vargish T (1986) Ruptured right atrium from non-penetrating trauma of the chest. South Med J 79: 499±501 78. Goldstein L, Mirvis SE, Kostrubiak IS, Turney SZ (1989) CT diagnosis of acute pericardial tamponade after blunt chest trauma. AJR 152: 739±741 79. Shapiro MJ, Yanofsky SD, Trapp J, Durham RM, Labovitz A, Sear JE, Barth CW, Pearson AC (1991) Cardiovascular evaluation in blunt thoracic trauma using transesophageal echocardiography (TEE). J Trauma 31: 835±839 80. Morgan AS, Flancbaum L, Esposito T, Cox EF (1986) Blunt trauma to the diaphragm: an analysis of 44 patients. J Trauma 26: 565±567 81. Gelman R, Mirvis SE, Gens D (1991) Diaphragmatic rupture due to blunt trauma: sensitivity of plain chest radiographs. AJR 156: 51±57 82. Murray JG, Caoili E, Gruden JF, Evans SJJ, Halvorsen RA, Mackersie RC (1996) Acute rupture of the diaphragm due to blunt trauma: diagnostic sensitivity and specificity of CT. AJR 166: 1035±1039 83. Israel RS, Mayberry JC, Primack SL (1996) Diaphragmatic rupture: use of helical CT scanning with multiplanar reformations. AJR 167: 1201±1203 84. Shanmuganathan K, Mirvis SE, Withe CS, Pomerantz SM (1996) MR imaging evaluation of hemidiaphragms in acute blunt trauma: experience with 16 patients. AJR 167: 397±402 85. Estsera AS, Platt MR, Mills LJ (1979) Traumatic injuries of the diaphragm. Chest 75: 306±313 86. Worthy SA, Kang EY, Hartman TE, Kwong JS, Mayo JR, Muller NL (1995) Diaphragmatic rupture: CT findings in 11 patients. Radiology 194: 885±888 87. Rodriguez M, Rodriguez A, Shatney C (1986) Rupture of diaphragm in blunt chest trauma. J Trauma 26: 438±444 88. Nilsson PE, Aspelin P, Ekberg O, Senyk J (1988) Radiologic diagnosis in traumatic rupture of the right diaphragm. Acta Radiol 29: 653±655 89. Heiberg E, Wolverson MK, Hurd RN, Jagannadharao B, Sundaram M (1980) CT recognition of traumatic rupture of the diaphragm. AJR 135: 369±372 90. Kuhlman JF, Pozniak MA, Collins J, Knisely BZ (1998) Radiographic and CT findings of blunt chest trauma: aortic injuries and looking beyond them. Radiographics 18: 1085±1106 91. Shackleton KL, Stewart ET, Andrew JT (1998) Traumatic diaphragmatic injuries: spectrum of radiologic findings. Radiographics 18: 49±59 92. Mirvis SE, Keramati B, Buckman R, Rodriguez A (1988) MR imaging of traumatic diaphragmatic rupture. J Comput Assist Tomogr 12: 147±149 93. Ochsner MG, Razycki GS, Lucente F, Wherry OC, Champion HR (1993) Prospective evaluation of thoracoscopy for diagnosing diaphragmatic injury in thoracoabdominal trauma: a preliminary report. J Trauma 34: 704±709