ajr%2e09%2e2544

AJR:194, May 2010 1197

Corresponding to technical improvements, the role of CT has evolved in the assessment of acute thoracic injury. Previously, CT had a screening role, in which the goal was the identification of a mediastinal hematoma on unenhanced CT [5]. More recently, CT has acquired a diagnostic role, in which the tho-racic aorta is evaluated for direct signs of in-jury on contrast-enhanced CT [68]. How-ever, to our knowledge there are no recent studies addressing whether there has been a change in the radiologic evaluation of blunt thoracic aortic injury in the pediatric popu-lation. The purpose of this study was to re-view our institutional experience with blunt thoracic aortic injury in pediatric patients and to assess whether evaluation of this in-jury has transitioned from thoracic aortogra-

Radiologic Evaluation of Blunt Thoracic Aortic Injury in Pediatric Patients

Waleska M. Pabon-Ramos1David M. WilliamsPeter J. Strouse

Pabon-Ramos WM, Williams DM, Strouse PJ

1All authors: Department of Radiology, University of Michigan Health System, 1500 E Medical Center Dr., Mott Hospital, Rm. F3503, Ann Arbor, MI 48109-0252. Address correspondence to P. J. Strouse ([email protected]).

CMEThis article is available for CME credit.See www.arrs.org for more information.

Pediatr ic Imaging Or ig ina l Research

AJR 2010; 194:11971203

0361803X/10/19451197

American Roentgen Ray Society

Blunt thoracic aortic injury is un-common in the pediatric popula-tion. Traumatic aortic injury ac-counts for only 2.1% of pediatric

deaths resulting from trauma. However, trau-matic aortic injury is highly lethal especially among children, for whom initial survival is reported to be 7% compared with 14% in adults [1, 2].

Prompt diagnosis is essential for the rap-id initiation of therapy because the mortal-ity rate due to untreated acute thoracic aor-tic injury is high. The reported percentage of patients with this injury who survive af-ter reaching the hospital is 33% for pediatric patients compared with 76% for adults [1, 3]. Moreover, a definitive diagnosis is complete-ly dependent on imaging findings [4].

Keywords: blunt thoracic aortic injury, CT, emergency radiology, pediatric thoracic trauma, thoracic aortography, trauma

DOI:10.2214/AJR.09.2544

Received February 6, 2009; accepted after revision September 15, 2009.

OBJECTIVE. The objective of our study was to assess the mechanism of injury, associ-ated injuries, and radiographic findings of pediatric patients presenting with blunt thoracic aortic injury.

MATERIALS AND METHODS. The medical records and imaging studies of all pedi-atric patients presenting with blunt thoracic aortic injury from January 1986 through Decem-ber 2007 (n = 17) were reviewed. The mechanism of injury, associated injuries, imaging find-ings, and surgical findings were recorded. The Fishers exact test was used to assess changes in utilization of chest CT and thoracic aortography.

RESULTS. The most frequent mechanism of injury was motor vehicle crash in which the patient was an unrestrained driver or unrestrained passenger (9/17 = 53%). The most common concurrent injury was solid abdominal organ injury (9/17 = 53%). The most frequent find-ing was a prominent or indistinct aortic knob (16/17 = 94%) on chest radiography, a periaor-tic hematoma and aortic contour abnormality on chest CT (9/10 = 90%), and aortic contour abnormality on thoracic aortography (11/11 = 100%). There was a statistically significant in-crease (p = 0.03) in chest CT examinations performed between January 1986 and December 1997 (4/9 = 44%) compared with between January 1986 and December 2007 (8/8 = 100%). There was a statistically significant decrease (p = 0.05) in thoracic aortography examinations performed between January 1986 and December 1997 (8/9 = 89%) compared with between January 1986 and December 2007 (3/8 = 38%).

CONCLUSION. Blunt thoracic aortic injury is a rare injury in the pediatric population. Radiologic evaluation of pediatric patients presenting with this injury has changed. More chest CT examinations and fewer thoracic aortography examinations are being performed. Furthermore, surgeons are choosing to perform surgery on the basis of chest CT findings con-sistent with aortic injury.

Pabon-Ramos et al. Thoracic Aortic Injury in Pediatric Patients

Pediatric ImagingOriginal Research

FOCU

S O

N:

Dow

nloa

ded

from

ww

w.a

jronli

ne.or

g by 3

6.83.3

7.83 o

n 10/0

3/14 f

rom IP

addre

ss 36

.83.37

.83. C

opyri

ght A

RRS.

For p

erson

al use

only;

all ri

ghts

reserv

ed

1198 AJR:194, May 2010

Pabon-Ramos et al.

phy to contrast-enhanced chest CT, as it has in the adult population [9, 10].

Materials and MethodsData Collection

After having received institutional review board (IRB) approval and in compliance with HIPAA, we accessed pertinent medical records and imag-ing studies of the patients in this study. Subject in-formed consent was waived by the IRB because of the retrospective nature of this investigation.

Patients presenting with blunt thoracic aortic injury to our institution between January 1986 and December 2007 were identified by performing a discharge database search limiting the search terms to the thoracic surgery or trauma surgery team and to patients 17 years old or younger. The resulting list was reviewed to identify patients with blunt thoracic aortic injury. Because of the rarity of blunt thoracic aortic injury in the pediat-ric population, patients who were initially treated at or admitted to a different hospital and were sub-sequently transferred to our institution were in-cluded in the study. Patients 18 years old or older were excluded from the study.

A retrospective review of the hard-copy and electronic medical records of all the identified pediatric patients with blunt thoracic aortic inju-ry was performed. The following characteristics were noted from the medical records: patient age and sex, date of trauma, initial admitting hospital and date of transfer to our institution if applicable, mechanism of injury, concomitant injuries, date of surgical intervention, intraoperative findings, and surgical procedure performed.

A retrospective review of the initial hard-copy or digital chest radiographs and chest CT images or of both hard-copy and digital images was per-formed by one of the authors who recorded the

findings. The hard-copy, digital, or both hard-copy and digital thoracic aortograms were reviewed by another author who also recorded the findings.

The following findings were recorded as pres-ent or absent on the chest radiographs: widened mediastinum; prominent or indistinct aortic knob; obliteration of the aortic outline; obliteration of the aortopulmonary window; rightward tracheal deviation; rightward nasogastric tube deviation; left pleural cap; pleural effusion or hemothorax; widened paratracheal or paravertebral stripe; de-pression of the left mainstem bronchus; fracture of the first rib, second rib, or both first and second ribs; pneumothorax; presence of a chest tube; and scapular fracture [1115].

If chest CT was performed before surgical in-tervention, the following findings were recorded as present or absent: pseudoaneurysm, intimal flap, periaortic hematoma, contour abnormality, and posttraumatic coarctation [15, 16].

If thoracic aortography was performed before surgical intervention, the following findings were recorded as present or absent: contour abnormal-ity, pseudoaneurysm, intimal flap, and posttrau-matic coarctation [16].

Data AnalysisFrequency calculations were performed to de-

termine the rate of the different mechanisms of injuries, concomitant injuries, intraoperative find-ings, and types of surgical interventions. Frequen-cy calculations were also performed to determine the rates of the different findings present on chest radiography, chest CT, and thoracic aortography.

The study period was divided into two subpe-riods: January 1986December 1997 and Janu-ary 1998December 2007. Because of the small number of subjects, the Fishers exact test was per-formed to determine whether there was a statis-tically significant change in the performance of chest CT and thoracic aortography between the subperiods. A p value equal to or less than 0.05 was considered statistically significant.

ResultsDuring the 22-year period from January

1986 through December 2007, 17 pediat-ric patients who had suffered blunt thorac-ic aortic injury were admitted to our institu-tion. This yields a frequency of 0.8 cases per year. The patients ranged in age from 3 to 17 years. The mean age was 14.8 years and the median age, 16 years. There were 13 boys (76%) and four girls (24%).

Three patients (18%) were admitted di-rectly to our institution, whereas 14 patients (82%) were transferred to our institution from another hospital. Nine patients were

admitted on the day of the trauma, two pa-tients on the day after the trauma, three pa-tients 2 days later, one patient 3 days later, one patient 5 days later, and one patient 17 days later.

The reported mechanisms of injury are shown in Table 1. The most frequent mech-anism of injury was motor vehicle crash in which the patient was an unrestrained driver or unrestrained passenger (9/17 = 53%).

The injuries most frequently associated with blunt thoracic aortic injury are shown in Table 2. The most common concurrent injury was solid abdominal organ injury (9/17 = 53%).

TABLE 1: Mechanism of Injury in 17 Patients

MechanismNo. (%) of Patients

Motor vehicle crash 13 (76.5)

Restrained driver 2 (11.8)

Unrestrained driver 3 (17.6)

Restrained front seat passenger 1 (5.9)

Unrestrained front seat passenger 3 (17.6)

Restrained rear seat passenger 1 (5.9)

Unrestrained rear seat passenger 3 (17.6)

Motorcycle 1 (5.9)

Dirt bike 1 (5.9)

Pinned by tractor 1 (5.9)

Car versus pedestrian 1 (5.9)

TABLE 2: Associated Injuries in 17 Patients

InjuryNo. (%) of Patients

Solid abdominal organ 9 (53)

Spleen 6 (35)

Liver 5 (29)

Kidney 3 (18)

Brain 8 (47)

Lung 7 (42)

Pelvic fractures 7 (42)

Appendicular fractures 7 (42)

TABLE 3: Findings on Chest Radiography in 17 Patients

FindingNo. (%) of Patients

Prominent or indistinct aortic knob 16 (94)

Wide paratracheal stripe 12 (71)

Widened mediastinum 11 (65)

Obliteration of the aortopulmonary window

10 (59)

Pleural effusion or hemothorax 10 (59)

Obliteration of descending aortas outline

9 (53)

Rightward deviation of nasogastric tube

8 (47)

Left pleural cap 7 (42)

Rightward deviation of trachea or endotracheal tube

5 (29)

Pneumothoraxa 2 (12)

Fracture of first rib, second rib, or both ribs

2 (12)

Depression of left mainstem bronchus 1 (6)

Scapular fracture 1 (6)aSeven patients already had a chest tube in place in the reviewed chest radiograph, four on the left and three on the right.

Dow

nloa

ded

from

ww

w.a

jronli

ne.or

g by 3

6.83.3

7.83 o

n 10/0

3/14 f

rom IP

addre

ss 36

.83.37

.83. C

opyri

ght A

RRS.

For p

erson

al use

only;

all ri

ghts

reserv

ed

AJR:194, May 2010 1199

Thoracic Aortic Injury in Pediatric Patients

Chest radiographs for all patients were re-viewed; in 10 patients (59%), these radiographs were obtained initially after the trauma. Chest radiographs of the remaining seven patients were obtained at our institution after the pa-tient had been transferred from the initial hospital. Of these examinations, radiographs were obtained on the same day of the trauma in three patients, the day after the trauma in two patients, and 2 days after the trauma in two patients. The chest radiograph findings are listed in Table 3. A prominent or indis-tinct aortic knob was the most frequent ra-diographic finding (16/17 = 94%) (Fig. 1).

Chest CT scans were obtained for 12 pa-tients. Eleven (92%) of these chest CT exami-nations were performed the day of the trauma. One patient (8%) did not undergo chest CT the day of the trauma because he became un-stable in the CT scanner that day and the ex-amination had to be aborted; instead, this pa-tient underwent chest CT the day after the trauma. Of the chest CT scans initially ob-tained, images of two patients were obtained at outside institutions and were not available and images for one patient, also obtained at an outside institution, were excluded from the study because an unenhanced CT examina-tion was performed; however, it did show a periaortic hematoma. One of the patients for whom the initial chest CT examination was not available underwent a second chest CT ex-

amination at our institution the day after the trauma, and this study was reviewed. There-fore, a total of 10 contrast-enhanced chest CT examinationssix performed at outside insti-tutions and four at our institutionwere re-viewed. The findings are listed in Table 4.

Periaortic hematoma and aortic contour ab-normality were the most frequent CT findings (9/10 = 90%) (Fig. 1). Of note, four of the pa-tients underwent CT of the abdomen and pelvis the day of trauma but not CT of the chest. In the most cephalad images, generally referred to as the lung bases, two of these studies showed a periaortic hematoma and one revealed an intra-mural hematoma at the level of the distal de-scending thoracic aorta, strongly suggesting a thoracic aortic injury (Fig. 2). Based on these findings, three of these four patients proceeded

to angiography and subsequently to the operat-ing room, where traumatic aortic injury at the level of the aortic isthmus was confirmed in all three patients. The patient who did not proceed to angiography went directly to surgery with-out undergoing additional imaging, and trau-matic aortic injury at the same location was confirmed intraoperatively.

Thoracic aortograms were obtained for 11 patients. Six (55%) of these examinations were obtained the day of the trauma; two (18%), the day after the trauma; two (18%), 2 days later; and one (9%), 9 days later. Of the 11 thoracic aortography examinations ini-tially performed, two were performed at out-side institutions. One of these examinations, performed 2 days after the trauma, was un-available, but a second thoracic aortography

A

Fig. 116-year-old female unrestrained front-seat passenger who was extricated from car after motor vehicle collision.A, Portable supine chest radiograph shows prominent aortic knob (black arrow) and left apical cap (white arrow).B, Axial contrast-enhanced chest CT image shows pseudoaneurysm, intimal flap, and periaortic hematoma (arrow).C, Sagittal reconstruction of contrast-enhanced chest CT (B) shows pseudoaneurysm and intimal flap (arrow).

CB

TABLE 4: Findings on Chest CT in 10 Patients and Thoracic Angiography in 11 Patients

Finding

Contrast-Enhanced Chest CT Thoracic Aortography

No. % No. %

Aortic contour abnormality 9 90 11 100

Periaortic hematoma 9 90 NA NA

Pseudoaneurysm 8 80 7 64

Intimal flap 7 70 4 36

Posttraumatic coarctation 6 60 4 36

NoteNA = not applicable.

Dow

nloa

ded

from

ww

w.a

jronli

ne.or

g by 3

6.83.3

7.83 o

n 10/0

3/14 f

rom IP

addre

ss 36

.83.37

.83. C

opyri

ght A

RRS.

For p

erson

al use

only;

all ri

ghts

reserv

ed

1200 AJR:194, May 2010

Pabon-Ramos et al.

examination was performed for this patient at our institution 3 days after the trauma and was reviewed. Thus, a total of 11 thoracic aortography examinationsone performed at an outside institution and 10 at our institu-tionwere reviewed. The findings are listed in Table 4. Aortic contour abnormality was the most frequent finding (11/11 = 100%) (Fig. 3).

There was a statistically significant in-crease (p = 0.03) in the number of chest CT examinations performed when comparing the subperiod of January 1986December 1997 (4/9 = 44%) with the number per-formed during the subperiod of January 1986December 2007 (8/8 = 100%). Fur-thermore, there was a concomitant statistical-ly significant decrease (p = 0.05) in the num-ber of thoracic aortography examinations performed when comparing the subperiod of January 1986December 1997 (8/9 = 89%) with the subperiod of January 1998 Decem-ber 2007 (3/8 = 38%) (Fig. 4).

Surgical repair was performed on all pa-tients diagnosed with blunt thoracic aortic injury, predominantly within 3 days of the trauma: six patients on the same day as the trauma; five patients, the next day; and two patients, 3 days later. Only a minority of pa-tients underwent repair more than 3 days af-ter trauma: one patient 7 days later, one pa-

tient 9 days later, one patient 10 days later, and one patient 40 days later. Three of these patients had multisystem trauma that pre-cluded immediate aortic repair either because other injuries required emergent surgical in-tervention or because other injuries rendered the patients at high risk to undergo aortic re-pair. Only one patient underwent repair more than 3 days after the trauma because of de-layed diagnosis of the aortic injury.

Intraoperative findings were consistent with aortic injury in all cases. According to the operative notes, the most frequent finding was an aortic tear (9/17 = 53%) compared with aortic transection (6/17 = 35%) and pseudo-aneurysm formation (3/17 = 18%). Most of the surgical procedures (13/17 = 76%) were performed at our institution. The surgical procedures performed for aortic injury repair included graft interposition (13/17 = 76%), primary anastomosis (3/17 = 18%), and patch repair (1/17 = 6%). There were no deaths dur-ing these admissions.

DiscussionBlunt thoracic aortic injury is unusual in

pediatric patients, with incidence reported to range from 0.1% to 7.4% among children who sustain blunt chest trauma [13, 14, 17, 18]. Fac-tors that may explain why this injury is rare

A

Fig. 216-year-old male pedestrian struck by car.A, Axial contrast-enhanced abdominal CT image shows periaortic hematoma at level of distal descending thoracic aorta (arrow) on most cephalad image.B, Portable supine chest radiograph shows only questionable widened mediastinum (arrow). Thoracic aortic injury workup was performed in this patient based on abdominal CT findings.

B

Fig. 314-year-old boy involved in dirt bike crash. Thoracic aortogram shows contour abnormality, intimal flap (white arrow), and posttraumatic coarctation (black arrow).

Dow

nloa

ded

from

ww

w.a

jronli

ne.or

g by 3

6.83.3

7.83 o

n 10/0

3/14 f

rom IP

addre

ss 36

.83.37

.83. C

opyri

ght A

RRS.

For p

erson

al use

only;

all ri

ghts

reserv

ed

AJR:194, May 2010 1201


in the pediatric population include increased compliance of the chest wall and elasticity of the tissues; lack of atherosclerotic disease; decreased body mass, translating into a low-er magnitude of force on impact; and fewer children involved in motor vehicle crashes as drivers or passengers [1, 12, 13, 19]. In our 22-year review, the frequency of blunt thoracic aortic injury was 0.8 cases per year.

Approximately 3050% of patients with blunt thoracic aortic injury do not show ex-ternal signs of direct chest trauma [2, 3, 20]. Also, they usually present with more evident neurologic, skeletal, or abdominal findings. The lack of visible external thoracic inju-ries, the presence of more evident coexisting physical examination findings, and the low incidence of blunt thoracic aortic injury in the pediatric population contribute to a low clinical index of suspicion for this injury in children. Thus, determining the mechanism of injury is important to raise the clinicians index of suspicion for blunt thoracic aortic injury. If the mechanism of injury involves high-energy trauma to the torso including deceleration, thoracic compression, or tho-racic crush, the possibility of blunt thorac-ic aortic injury should be considered. These mechanisms of injury are present, for exam-ple, in motor vehicle crashes, pedestrian or bicyclist struck by an automobile, and fall from a height [21].

In our study, the most frequent mechanism of injury was motor vehicle crash in which the patient was an unrestrained driver or un-restrained passenger. This finding is in con-trast with that of Eddy et al. [1] who found that the most frequent mechanism of injury was a pedestrian being struck by a car. How-ever, differences in patient inclusion and ex-clusion criteria may explain this difference in findings. First, Eddy et al. included pedi-atric patients 16 years old or younger, and the

study group had a mean age of 12 years with a maximum of 15 years. We included patients 17 years old or younger, and the study group had a higher mean age of 14.8 years, a me-dian of 16 years, and maximum of 17 years. Most of our patients were adolescents (i.e., 1217 years old) except two: a 3 year old and a 10 year old. Moreover, the most significant outlier, the 3-year-old patient, suffered an unusual mechanism of injurypinning by a tractor. Hence, our patients may have been involved in a greater number of motor vehicle crashes because they tended to be older and, presumably, more likely to drive. In fact, five of the 17 patients in our cohort were driving at the time of the crash. The findings in an-other study support this explanation: Ander-son et al. [22] included patients younger than 20 years old, and their study group also had a higher mean age of 14.3 years with a maxi-mum age of 19 years. As in our study, Ander-son et al. found that the most common mech-anism of injury was motor vehicle crash.

Another explanation for the differences in the predominant mechanism of injury is that our study population was composed ex-clusively of patients who survived the ini-tial trauma. The cohort in the Eddy et al. [1] study, on the other hand, was composed of pediatric patients who had and those who had not survived and all of the patients who experienced the most common mechanism of injury in their cohortthat is, pedestrian struck by cardied.

As presented in prior studies, our study showed that the injury most frequently asso-ciated with blunt thoracic aortic injury was intraabdominal solid organ injury. Unlike the study by Eddy et al. [1], however, the in-traabdominal organ most commonly injured in our study was the spleen.

The traditional imaging algorithm for pe-diatric patients with suspected blunt thoracic

aortic injury consists of supine or, if possible, upright chest radiographs followed by thorac-ic aortography if the chest radiographs show signs of a mediastinal hematoma. With a neg-ative predictive value of 96% for the supine chest radiograph and 98% for the erect chest radiograph, a chest radiograph with normal findings nearly excludes blunt thoracic aor-tic injury, unless there is a compelling clini-cal reason to suspect the diagnosis, and usu-ally obviates further studies [23, 24]. Indeed, chest radiographs were obtained for all of our patients. The initial chest radiograph, howev-er, was frequently a supine chest radiograph, which may appear abnormal because of pa-tient positioning and the effects of magnifica-tion. A comparison of whether the initial chest radiograph was supine or upright could not be performed because not all of the initial chest radiographs were available when the images were reviewed. A prominent or indistinct aor-tic knob was the most frequent radiographic finding (Fig. 1). Although all the chest radio-graphs appeared abnormal on review, some showed very subtle findings (Fig. 2).

With the evolution of CT technology, CT angiography has become the standard non-invasive imaging technique for the depiction of vascular anatomy and abnormalities. Ad-vantages of a CT evaluation include that it is noninvasive, is less expensive than con-ventional digital subtraction angiography, allows multiplanar visualization of the aor-ta, allows evaluation of adjacent organs, and has shorter acquisition times. Disadvantages include a lack of dynamic information and poor visualization of small collaterals [6]. Some authors have concluded that chest CT is not sufficient to exclude aortic injury [11]. Others have stated that contrast-enhanced chest CT has 100% sensitivity and 100% negative predictive value for the detection of traumatic aortic injury and that it can be used to exclude this type of injury [25]. Fur-thermore, others have stated that CT has be-come the standard of care for the evaluation of blunt thoracic aortic injury in adults [9, 10]. In another medical center, chest CT was the technique used to diagnose all seven pe-diatric thoracic aortic injuries that presented over a 10-year period, and only two confir-matory thoracic aortograms and one con-firmatory transesophageal echocardiogram were obtained [22]. As in the latter study, all contrast-enhanced chest CT studies in our cohort showed direct signs of traumatic aor-tic injury; periaortic hematoma and aortic contour abnormality were the most frequent

1985 1990 1995Thoracic Aortography Only

Thoracic Aortographyand Chest CT

Chest CT Only

2000 2005 2010

Fig. 4Timeline of imaging performed per patient over 22-year study period. One patient was excluded from graph because neither thoracic aortography nor chest CT was performed. = one patient.

Dow

nloa

ded

from

ww

w.a

jronli

ne.or

g by 3

6.83.3

7.83 o

n 10/0

3/14 f

rom IP

addre

ss 36

.83.37

.83. C

opyri

ght A

RRS.

For p

erson

al use

only;

all ri

ghts

reserv

ed

1202 AJR:194, May 2010

Pabon-Ramos et al.

ones (Figs. 1 and 2). Moreover, there was a statistically significant increase in the num-ber of chest CT examinations performed (p = 0.03) during the last decade in the study peri-od compared with earlier. This increase may have been influenced by factors other than the ordering clinicians acceptance of chest CT as a diagnostic tool specifically for the detection of suspected thoracic aortic inju-ry. For example, if a trauma team decides to routinely order chest CT for any level 1 blunt trauma, regardless of chest radiography find-ings or whether thoracic aortic injury is sus-pected, the likelihood of detecting more tho-racic aortic injuries would be greater. During the first subperiod of our study, four patients did not undergo chest CT but did undergo ab-dominal and pelvic CT, which showed a peri-aortic hematoma at the level of the diaphrag-matic hiatus on the most cephalad images. Although thoracic aortic injury was expected at this level, injury at the level of the aortic isthmus was instead confirmed angiographi-cally in three patients and intraoperatively in all four patients. Not only does this finding reflect the higher injury rate at the level of the isthmus (90%) when compared with this rare injury at the level of diaphragmatic hia-tus [26], but it is concordant with the previ-ously reported 94% specificity rate that this sign has for blunt aortic injury [27].

Additionally, during the second subperi-od (January 1986December 2007), five pa-tients did not undergo thoracic aortography and were taken to the operating room sole-ly on the basis of chest CT revealing direct signs of blunt thoracic aortic injury. Most importantly, none of the thoracotomies per-formed on these five patients was nonthera-peutic. These results mean that, like thoracic aortography, chest CT provides crucial guid-ance regarding the need for aortic repair. Surgeons in other institutions seem to have embraced this concept as well. For example, in a study by Anderson et al. [22] of four pe-diatric patients with thoracic aortic injury who were taken to the operating room for re-pair, two patients underwent only chest CT before surgical intervention.

Thoracic aortography has been traditional-ly considered the standard reference technique to evaluate patients with suspected blunt tho-racic aortic injury [25]. Nonetheless, investi-gators have reported that 8092% of patients who undergo thoracic aortography on the ba-sis of the mechanism of injury, clinical sus-picion, and abnormal chest radiography find-ings will not have an aortic injury [3, 4, 11].

As we expected, all thoracic aortograms re-viewed in our study showed signs of traumat-ic aortic injury, of which an aortic contour ab-normality was the most frequent one (Fig. 3). However, there was a statistically significant decrease in the number of diagnostic thorac-ic aortograms obtained (p = 0.05) during the latest decade of our study period compared with previously. Thoracic aortography was omitted in only one patient before 1998. This patient was too unstable for the examination and was taken directly to the operating room on the basis of clinical history and chest radi-ography findings. On the other hand, five pa-tients in the latter subperiod were taken to the operating room without undergoing thoracic aortography because the surgeons were con-fident about the chest CT results. Indeed, the last consecutive four patients did not undergo thoracic aortography.

We anticipate that, during the era of tho-racic aortic endografts, CT will continue to have a prominent role. Advantages of endo-vascular treatment include a lower operative risk due to circumventing bleeding and pul-monary complications related to open thorac-ic surgery and reduced systemic hepariniza-tion time, lack of significant delay in surgical or medical management of concomitant inju-ries, and shorter hospital stays allowing rapid rehabilitation. Unfortunately, current aortic endografts are not ideal for repair of the pe-diatric aorta because of the required vascular access diameter, which is relatively large for young patients with spastic arteries, smaller aortic diameter and tight arch configuration in pediatric patients, and unknown long-term use in the growing pediatric population with increased life expectancies [2832]. How-ever, technical improvements may result in durable, low-profile, small-diameter, flexible endografts capable of conforming to the pe-diatric aorta. At that time, CT not only will aid in diagnosis but also will be crucial for planning the procedure.

The greatest limitation of this study was the small sample size (n = 17). Nonetheless, to our knowledge this series is the largest se-ries of pediatric patients with blunt thorac-ic aortic injury reported given the rarity of this type of injury in this particular popula-tion. Another limitation was that 14 (82%) of the patients were transferred from an outside hospital. Therefore, their medical records, imaging studies, or both were sometimes in-complete and unavailable for review.

There are also limitations intrinsic to a retrospective study. For example, we were

aware that all the patients in the cohort had surgically proven blunt thoracic aortic injury before reviewing the images. Thus, measure-ment bias was likely introduced on reviewing the chest radiographs given that radiograph-ic interpretation is, to some extent, subjec-tive. On the contrary, that measurement bias was less likely to have been introduced on reviewing the chest CT images and thorac-ic aortograms because the findings on these studies are more definitive.

We did not assess the number of negative chest CT scans or thoracic aortograms ob-tained during the study period. Unfortunate-ly, these data appear incompletely available. Therefore, we do not know the overall use of CT and angiography and cannot calculate or compare specificity and sensitivity of the two techniques for the detection of thoracic aortic injury.

Blunt thoracic aortic injury is rare in children, with 0.8 cases occurring per year in the region served by our level 1 pediat-ric trauma center. Nevertheless, radiologists must be prepared to properly image and ad-equately diagnose this life-threatening inju-ry. Further research is needed to determine proper indications for chest CT for suspect-ed aortic trauma in the pediatric population to avoid overutilization.

Radiographic evaluation of pediatric pa-tients presenting with blunt thoracic aortic injury has changed over the past 20 years. Concurrent to advances in CT technology that now allow the identification of direct signs of aortic injury on contrast-enhanced chest CT, more chest CT examinations are presently being performed for this group of patients. As a result, surgeons are confident in taking patients to the operating room sole-ly on the basis of the chest CT results, obvi-ating thoracic aortography. We believe that thoracic aortography should be performed only when the chest CT examination is in-adequate, when chest CT findings are inde-terminate, or when there is a high index of clinical suspicion for thoracic aortic injury with discordant chest CT findings. Such in-dications, however, are rare.

References 1. Eddy AC, Rusch VW, Fligner CL, Reay DT, Rice

CL. The epidemiology of traumatic rupture of the thoracic aorta in children: a 13-year review. J Trauma 1990; 30:989991

2. Parmley LF, Mattingly TW, Manion WC, Jahnke EJ. Nonpenetrating traumatic injury of the aorta. Circulation 1958; 17:10861101

Dow

nloa

ded

from

ww

w.a

jronli

ne.or

g by 3

6.83.3

7.83 o

n 10/0

3/14 f

rom IP

addre

ss 36

.83.37

.83. C

opyri

ght A

RRS.

For p

erson

al use

only;

all ri

ghts

reserv

ed

AJR:194, May 2010 1203


3. Schmidt CA, Jacobson JG. Thoracic aortic injury: a ten-year experience. Arch Surg 1984; 119:12441246

4. Stark P. Traumatic rupture of the thoracic aorta: a review. Crit Rev Diagn Imaging 1984; 21:229255

5. Harris JH, Horowitz DR, Zelitt DL. Unenhanced dynamic mediastinal computed tomography in the selection of patients requiring thoracic aortog-raphy for the detection of acute traumatic aortic injury. Emerg Radiol 1995; 2:6776

6. Yu T, Xiaomei Z, Lijun T, Dehang W, Nael S. Re-view of CT angiography of aorta. Radiol Clin North Am 2007; 45:461483

7. Gavant ML, Flick P, Menke P, Gold RE. CT aor-tography of thoracic aortic rupture. AJR 1996; 166:955961

8. Mirvis SE, Kethirkamuganathan S, Miller BH, White CS, Turney SZ. Traumatic aortic injury: diagnosis with contrast-enhanced thoracic CTfive-year experience at a major trauma center. Ra-diology 1996; 200:413422

9. Novelline RA, Rhea JT, Rao PM, Stuk JL. Helical CT in emergency radiology. Radiology 1999; 213:321339

10. Malloy PC, Richard HM. Thoracic angiography and intervention in trauma. Radiol Clin North Am 2006; 44:239249

11. Eddy AC, Misbach GA, Luna GK. Traumatic rup-ture of the thoracic aorta in the pediatric patient. Pediatr Emerg Care 1989; 5:228230

12. Ali IS, Fitzgerald PG, Gillis DA, Lau HYC. Blunt traumatic disruption of the thoracic aorta: a rare injury in children. J Pediatr Surg 1992; 27:1281

1284 13. Erb RE, Stein SM, Mazer MJ, Nance EP. Blunt

thoracic aortic injury in children. Emerg Radiol 1994; 1:215220

14. Spouge AR, Burrows PE, Armstrong D, Daneman A. Traumatic aortic rupture in the pediatric popu-lation. Pediatr Radiol 1991; 21:324328

15. Moore MA, Wallece EC, Westra SJ. The imaging of paediatric thoracic trauma. Pediatr Radiol 2009; 39:485496

16. Sinclair DS. Traumatic aortic injury: an imaging review. Emerg Radiol 2002; 9:1320

17. Haller JA. Thoracic injuries. In: Welch KJ, ed. Pediatric surgery. Chicago, IL: Year Book Medi-cal Publishers, 1986:143153

18. Nakayama DK, Ramenofsky ML, Rowe MI. Chest injuries in childhood. Ann Surg 1989; 210: 770775

19. Vyas PK, Sivit CJ. Imaging of blunt pediatric tho-racic trauma. Emerg Radiol 1997; 4:1625

20. Greene R. Lung alterations in thoracic trauma. J Thorac Imaging 1987; 2:111

21. Patel NH, Stephens KE, Mirvis SE, Shanmugana-than K, Mann FA. Imaging of acute thoracic aor-tic injury due to blunt trauma: a review. Radiology 1998; 209:335348

22. Anderson SA, Day M, Chen MK, et al. Traumatic aortic injuries in the pediatric population. J Pedi-atr Surg 2008; 43:10771081

23. White CS, Mirvis SE. Pictorial review: imaging of traumatic aortic injury. Clin Radiol 1995; 50:281287

24. Mirvis SE, Bidwell JK, Buddemeyer EU, et al. Value of chest radiography in excluding aortic rupture. Radiology 1987; 163:487493

25. Dyer DS, Moore EE, Mestek MF, et al. Can chest CT be used to exclude aortic injury? Radiology 1999; 213:195202

26. Creasy JD, Chiles C, Routh WD, Dyer RB. Over-view of traumatic injury of the thoracic aorta. Ra-dioGraphics 1997; 17:2745

27. Wong H, Gotway MB, Sasson AD, et al. Periaor-tic hematoma at diaphragmatic crura at helical CT: sign of blunt aortic injury in patients with mediastinal hematoma. Radiology 2004; 231: 185189

28. Fattori R, Napoli G, Lovato L, et al. Indications for, timing of, and results of catheter-based treat-ment of traumatic injury to the aorta. AJR 2002; 179:603609

29. Ferrari E, Tozzi P, von Segesser L. Thoracic aorta emergencies: is the endovascular treatment the new gold standard? Interact Cardiovasc Thorac Surg 2006; 5:730734

30. Marcheix B, Dambrin C, Bolduc JP, et al. Endo-vascular repair of traumatic rupture of the aortic isthmus: midterm results. J Thorac Cardiovasc Surg 2006; 132:10371041

31. Starnes BW, Arthurs ZM. Endovascular manage-ment of vascular trauma. Perspect Vasc Surg En-dovasc Ther 2006; 18:114129

32. Therasse E, Soulzez G, Giroux MF, et al. Stent-graft placement for the treatment of thoracic aor-tic diseases. RadioGraphics 2005; 25:157173

F O R Y O U R I N F O R M A T I O N

This article is available for CME credit. See www.arrs.org for more information.

Dow

nloa

ded

from

ww

w.a

jronli

ne.or

g by 3

6.83.3

7.83 o

n 10/0

3/14 f

rom IP

addre

ss 36

.83.37

.83. C

opyri

ght A

RRS.

For p

erson

al use

only;

all ri

ghts

reserv

ed

ajr%2e09%2e2544

Documents