| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, University of Washington, Box 357115, Seattle, WA
98195-7115.
2 Harborview Injury Prevention and Research Center, Seattle, WA.
Received September 15, 2006;
accepted after revision April 25, 2007.
Address correspondence to M. Sammer.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. This study is a retrospective analysis of patients presenting with blunt trauma over 4.5 years at a level 1 trauma center. Indeterminate CTA findings in patients with blunt injury were identified through a database search of imaging reports. CTA findings and final outcomes, including catheter angiography and clinicopathologic records, were reviewed independently by blinded observers.
RESULTS. One hundred seven patients (age range, 11-88 years) met the inclusion criteria. Seventy-two (age range, 15-88 years) had a reference standard of subsequent catheter angiography, and 35 subjects (age range, 11-87 years) did not undergo catheter angiography and therefore had a reference standard of clinicopathologic review. No subjects with isolated mediastinal hematoma on CTA had aortic or intrathoracic great vessel injury, for a positive predictive value of 0% (95% CI, 0-0.028). Using our proposed classification scheme, we found a direct correlation between the percentage of cases that underwent catheter angiography and hematoma severity.
CONCLUSION. When CTA is indeterminate in blunt thoracic trauma, conventional angiography is unlikely to show an aortic or intrathoracic great vessel injury and may be unnecessary. A grading system for mediastinal hematomas could help triage patients to conventional angiography when further imaging is desired.
Keywords: aorta • blunt trauma • cardiac imaging • catheter aortography • CT angiography • emergency radiology • trauma
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Diagnosis of blunt aortic and intrathoracic great vessel injuries has historically relied on clinical suspicion and the gold standard of catheter aortography [2-4]. However, in studies, chest CT and CT angiography (CTA) have shown high sensitivity and specificity for these injuries, and CTA is now commonly used as the primary imaging approach to rule out BAI [5-7]. However, CTA can be indeterminate for BAI by revealing a mediastinal hematoma (a possible surrogate for injury) but depicting no direct evidence of in jury (pseudoaneurysm, dissection flap, caliber change, or intramural hematoma). In these indeterminate cases, current recommendations suggest subsequent catheter aortography [8-10]. To our knowledge, these recommendations have not been validated with a study dedicated to this group of patients. Rather, these recommendations are based on either studies using traditional contrast-enhanced CT rather than CTA [11-13] or studies focused on other hypotheses that include patients with indeterminate CTA [14, 15].
The primary objective of this study was to determine the frequency of aortic or intrathoracic great vessel injury in subjects with a mediastinal hematoma on CTA but no direct CTA signs of aortic or great vessel injury. Our secondary objective was to determine whether the location of the mediastinal hematoma can be used to predict the likelihood of aortic or intrathoracic great vessel injury; to this end, a classification scheme was devised.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
The database search identified 743 subjects. The initial data set was then reviewed by one of the authors. CTA examinations with direct signs of aortic or intrathoracic great vessel injury and duplicate examinations were excluded, leaving 166 subjects. Most cases excluded were duplicate examinations, each producing three data search hits because of separate coding entries for chest, abdomen, and pelvis CT examinations performed concurrently. Finally, after review of images on the PACS, an additional 59 cases were excluded because they were not blunt trauma (n = 17), definite aortic or intrathoracic great vessel injury was present (n= 15 and n = 5, respectively), there was no mediastinal hematoma (n = 21), or the imaging examination was not CTA (n = 1). In all 166 cases, the original radiologist's report was reviewed, and the author reviewer concurred with the original radiologist about the presence of either aortic or intrathoracic great vessel injury and presence or lack of mediastinal hematoma. This left 107 CTA examinations that were indeterminate for aortic or intrathoracic great vessel injury (Fig. 1). However, in five of the 107 cases, a distinction could not be made between thymus and hematoma.
The CTA technique evolved over the 4.5-year period as technology and protocols progressed; however, all were performed in the arterial phase on an MDCT unit: 71 (66.4%) on a 4-MDCT unit and 36 (33.6%) on a 16-MDCT unit (LightSpeed, GE Healthcare). Our current protocol is as follows: 150 mL of iohexol (Omnipaque 350, Nycomed) administered at a rate of 4.0 mL/s; arterial phase imaging at peak plus 3 seconds; helical, full 0.6 second at 2.5 mm; 140 kV with automatic tube current (mA) settings (120-440); noise index, 10.5; and maximum-intensity-projection reconstructions (10-mm slice thickness at 2.5 mm).
Catheter angiography was performed with standard intraarterial digital subtraction technique (Integris, Philips Medical Systems). Nonionic contrast material was injected, typically at a rate of 20-40 mL/s for a total volume of 40-50 mL depending on patient age and cardiac output, with a 6-French pigtail catheter in the proximal ascending aorta in both obliquities. If the initial run was of limited diagnostic quality, imaging was repeated using a different technique, and if any abnormalities were present or if oblique views provided limited coverage, frontal projections were performed.
Each case was evaluated by one of two independent radiology residents on a PACS workstation (Centricity, GE Healthcare). Reviewers were blinded to the existence and findings of catheter angiography and to clinical follow-up. Each reviewer analyzed the CTA images and documented the presence of mediastinal hematoma, the location of the hematoma, evidence of direct aortic or intrathoracic great vessel injury, and presence of fractures that might be related to hematoma. The final report from the initial radiologist was then reviewed. In all cases, as on the initial report, no direct evidence of aortic or intrathoracic great vessel injury was present. Findings considered direct evidence of vascular injury were pseudoaneurysms, intimal flaps, intraluminal thrombi, contour irregularities, pseudocoarctation, and contrast extravasation.
The two radiologists classified mediastinal hematomas according to morphology and location. The internally developed classification scheme was as follows: type I, hematoma focused around a fracture with a clear fat plane between the mediastinal hematoma and the aorta; type II, mediastinal hematoma centered away from the aorta, but with minimal abutment of the aorta; type III, no clear fat plane between aorta and mediastinal hematoma, although appearance is not that of type IV; and type IV, mediastinal hemorrhage clearly focused at the aorta or great vessels with hemorrhage typically filling the mediastinum, surrounding these vessels and obliterating the fat planes between hematoma and aorta, intrathoracic great vessels, or both (Figs. 2, 3, 4 and 5A, 5B).
|
|
|
|
|
A second reviewer, blinded to CTA findings, evaluated each case for the presence of aortic or intrathoracic great vessel injury using the reference standard of reports of the catheter angiograms and subsequent CT images, if obtained, and any pertinent operative reports. If the subject was deceased, an attempt was made to retrieve autopsy reports. In addition, the length of the subject's clinical follow-up was documented. Finally, the trauma registry was also searched for possible missed cases.
The CTA findings were compared with the reference standard using standard two-by-two tables and the chi-square test. The analysis was performed using STATA software (version 8.0, Stata).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
In the 72 subjects who underwent conventional angiography, no aortic injuries were identified on imaging, including both the catheter angiograms and the subsequent chest CT images obtained in 16. Additional follow-up on the subjects who underwent catheter angiography revealed that four died during hospital admission. Of these, autopsy was performed on three, and the aorta was not injured in any. Clinical follow-up at our institution ranged from 15 to 1,560 days, with a mean of 410 days.
Of the 35 patients who did not undergo catheter aortography, 17 of the 35 (49%) were followed up at our institution after discharge from the hospital; the range of follow-up was 50-1,050 days, with a mean follow-up of 260 days. Two of these patients underwent chest CT, either contrast-enhanced CT or CTA, during hospitalization 4 and 7 days after injury. Neither showed signs of aortic or intrathoracic great vessel injury. Eleven of the 35 (31%) were not followed up at our institution after hospital discharge. Most of these patients had type I or II mediastinal hematomas, with specifics as follows: type I, n =5; type II, n = 4; type III, n = 1; and type IV, n =1. The patient with the most concerning hematoma on initial CTA (type IV) had under-gone pulmonary CTA during hospitalization 26 days after the initial injury; at that time, CTA was negative for aortic or intrathoracic great vessel injury. Finally, seven of the 35 patients (20%) died during hospitalization: Autopsy was performed on six, and none had an aortic or intrathoracic great vessel injury. Although autopsy was not performed on one of the deceased patients, no clinical signs during that patient's 19-day hospital stay pointed to aortic or intrathoracic great vessel injury, and the death of this 82-year-old man was thought to be secondary to multiple preexisting medical conditions and to trauma related to an acute cervical spine fracture. Finally, the trauma registry search for aortic and intrathoracic great vessel injuries during our study period confirmed that no injuries in any of the 107 patients had been missed.
Based on the composite reference standard of additional imaging, autopsy, and clinical follow-up, none of 107 subjects with mediastinal hematoma but no direct evidence of aortic or intrathoracic great vessel injury on CTA had aortic or intrathoracic great vessel injury. The positive predictive value for aortic or intrathoracic great vessel injury when isolated mediastinal hematoma is found on CTA is 0% (95% CI, 0-0.028).
The extent of mediastinal injury was diverse in our study group, with a relatively even distribution of hematoma across the devised morphology classifications. The classification system incorporates suspicion for injury by relating the location of the mediastinal hematoma to the aorta and intrathoracic great vessels. Correlation with the presence of an angiogram was great (correlation coefficient = 0.95), with most of the more suspicious cases going on to conventional angiography and the overwhelming majority of clearly fracture-related hematomas (type I) being followed clinically (Table 2). Because we did not identify any subjects with isolated mediastinal hematoma who had aortic or intrathoracic great vessel injury, no association was identified between hematoma location and aortic or intrathoracic great vessel injury.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Isolated mediastinal hematoma diagnosed using CTA in the absence of direct signs of BAI has not previously been fully classified or characterized, to our knowledge. We developed the mediastinal hematoma classification scheme to aid in the prediction of which subjects with mediastinal hematoma are at higher probability of aortic or intrathoracic great vessel injury. The high correlation between the hematoma classification scheme and subsequent catheter aortography confirms that clinicians at our center request angiography based on the shape and location of the isolated hematoma. However, our main result indicates that this practice is probably unnecessary because the prevalence of aortic or intrathoracic great vessel injury was 0% even in subjects with hematomas centered around the aorta and intrathoracic great vessels (type IV).
We have a limited sample size for each hematoma subtype. Consequently, the statistical power of our analysis is limited by this small sample size for the specific sub-types. For example, although we did not find any aortic injuries in the 26 type IV hematomas, the 95% CI for this proportion is relatively wide (0-10.8%). Further studies are needed to more precisely estimate the exact prevalence of aortic injury in this sub-group, and they may identify rare cases in which catheter aortography is positive when CTA is indeterminate. In this eventuality, a clinical prediction rule based on our classification scheme could help decide which indeterminate CTA cases should be followed up with catheter aortography.
We acknowledge the limitations of our study. As with any study, a prospective multicenter study with catheter aortography performed on all patients would be ideal. However, we have catheter angiograms for the majority of the patients in our study group, especially in those who might be thought to be at greatest risk of aortic or intrathoracic great vessel injury. Further, we have extensive imaging, autopsy, and clinical follow-up data on the subjects who did not undergo angiography. However, we do not have strong clinical follow-up data on 11 of the patients who did not undergo catheter angiography. This limitation is small because the majority of these patients (nine of 11) had type I or II hematomas—those who are at lowest risk for aortic or intrathoracic great vessel injury. In the other two patients, in-house follow-up was good and included CTA for pulmonary embolus late in one of the patient's hospital stay, which also showed no aortic or intrathoracic great vessel injury.
In addition, although catheter angiography is the gold standard for the diagnosis of blunt aortic or intrathoracic great vessel injury, it is not a perfect gold standard—inherently limiting our study's validation. Also, because the study was retrospective, some selection bias may have been introduced. To counter this bias, we looked at consecutive cases and verified that the database search adequately identified such cases at the onset of our study. Another limitation of some retrospective studies is difficulty in blinding; however, we do not believe this is an issue with our work because the CTA images were interpreted before evaluation of catheter aortography, catheter aortography results were recorded by the author who had not evaluated CTA, and we extended every effort to ensure blinding throughout the data collection and analysis.
When compared with the results of other recent studies, our study results lead one to question the necessity of catheter angiography in patients with indeterminate CTA. Although recent guidelines still recommend catheter aortography in cases of mediastinal hematoma detected on CTA without direct aortic injury [8], these recommendations are largely based on a few cases of aortic injury in the presence of mediastinal hematoma on nonhelical, non-CTA chest CT [11-13]. Furthermore, although studies examining this question by Mirvis et al. [14] and Dyer et al. [10] also conclude that a confirming catheter angiogram is mandated in periaortic hematoma, their data actually support our hypothesis that catheter angiography is highly unlikely to detect aortic injury.
In the study by Mirvis et al. [14], there were 90 cases of mediastinal hemorrhage with a normal aorta on contrast-enhanced helical CT; of these, 20 had negative catheter angiograms, and none was subsequently diagnosed with aortic injury. Dyer et al. [10], using both helical and conventional CT, also reported 127 cases of mediastinal hematoma without direct signs of aortic injury, which they define as "caliber change of the aorta, intraluminal irregularity, and abnormal contour of the aorta." All 127 patients had negative catheter angiograms. They do report one case after the conclusion of their study in which aortic injury (pseudoaneurysm) was found on catheter angiography, but the initial CT had been interpreted as showing periaortic hematoma without signs of direct aortic injury. However, in retrospectively reviewing the CT image, they note that a "caliber change" in the aorta, although subtle, was present. This case illustrates the possibility that periaortic hematoma may be the only readily recognized sign of aortic injury in a few cases. Finally, in a more recent study, Scaglione et al. [15] concluded that in cases of isolated mediastinal hematoma, sources of the hemorrhage other than the aorta should be sought before catheter angiography. This recommendation is based on 54 patients with mediastinal hematoma detected on helical CT, all of whom had negative catheter angiograms. Our study results concur with and add additional power with 107 cases (77 with catheter angiography).
In this study, we provide evidence that in cases of blunt thoracic trauma, aortic or intrathoracic great vessel injury is unlikely when CTA is indeterminate and that subsequent catheter angiography is unlikely to reveal aortic or intrathoracic great vessel injury. Based on our data, such patients can be followed up clinically, with subsequent cross-sectional imaging, or both without the need for catheter aortography. Our study adds to the growing body of evidence that even in the presence of mediastinal hematoma, when there is no direct evidence of aortic or intrathoracic great vessel injury on CTA, catheter angiography may be unnecessary.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S. D. Steenburg and J. G. Ravenel Acute Traumatic Thoracic Aortic Injuries: Experience with 64-MDCT Am. J. Roentgenol., November 1, 2008; 191(5): 1564 - 1569. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. D. Steenburg, J. G. Ravenel, J. S. Ikonomidis, C. Schonholz, and S. Reeves Acute Traumatic Aortic Injury: Imaging Evaluation and Management Radiology, September 1, 2008; 248(3): 748 - 762. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
