| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Emergency Radiology |
1 From the Department of Diagnostic and Interventional Radiology (M.W., N.T., P.S.) and Department of Traumatology and Orthopaedics (E.M., P.M., P.F.L.), Centre Hospitalier Universitaire Vaudois, BH07, 1011 Lausanne, Switzerland; and Biostatistics Unit, Institute of Social and Preventive Medicine, Lausanne University, Switzerland (G.v.M.). Received May 17, 2002; revision requested July 16; final revision received November 15; accepted November 27. Address correspondence to M.W. (e-mail: Max_Wintermark@hotmail.com).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: One hundred consecutive severe trauma patients who underwent conventional radiography of the thoracolumbar spine as well as thoracoabdominal multi–detector row CT were prospectively identified. Conventional radiographs were reviewed independently by three radiologists and two orthopedic surgeons; CT images were reviewed by three radiologists. Reviewers were blinded both to one another’s reviews and to the results of initial evaluation. Presence, location, and stability of fractures, as well as quality of reviewed images, were assessed. Statistical analysis was performed to determine sensitivity and interobserver agreement for each procedure, with results of clinical and radiologic follow-up as the standard of reference. The time to perform each examination and the radiation dose involved were evaluated. A resource cost analysis was performed.
RESULTS: Sixty-seven fractured vertebrae were diagnosed in 26 patients. Twelve patients had unstable spine fractures. Mean sensitivity and interobserver agreement, respectively, for detection of unstable fractures were 97.2% and 0.951 for multi–detector row CT and 33.3% and 0.368 for conventional radiography. The median times to perform a conventional radiographic and a multi–detector row CT examination, respectively, were 33 and 40 minutes. Effective radiation doses at conventional radiography of the spine and thoracoabdominal multi–detector row CT, respectively, were 6.36 mSv and 19.42 mSv. Multi–detector row CT enabled identification of 146 associated traumatic lesions. The costs of conventional radiography and multi–detector row CT, respectively, were $145 and $880 per patient.
CONCLUSION: Multi–detector row CT is a better examination for depicting spine fractures than conventional radiography. It can replace conventional radiography and be performed alone in patients who have sustained severe trauma.
© RSNA, 2003
Index terms: Cost-effectiveness • Radiations, exposure to patients and personnel • Spine, CT, 328.12115, 338.12115 • Spine, fractures, 328.41, 338.41 • Spine, radiography, 328.11, 338.11 • Trauma, 328.41, 338.41
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Given these potentially devastating consequences, especially for an individual in whom a spine injury has been missed at examination, and the potential medicolegal consequences for the involved physicians, all patients who have sustained trauma must be assessed for possible spine fractures. Ideally, the spine should be cleared within minutes after the admission of a trauma patient to the emergency room. In many cases, however, patients who have sustained severe trauma are uncooperative, obtunded, or unstable, and are thus difficult to examine clinically. Thus, a rapid yet accurate and thorough imaging assessment of the spine must be performed. The purpose of this imaging survey includes detection and characterization of the type and degree of bone fractures and ligamentous instability, as well as characterization of possible spinal cord injury (3,4).
In choosing the best imaging algorithm to use in screening for spine fractures, the radiologist is frequently encumbered by time limitations, lack of patient cooperation, and constraints related to severe hemodynamic instability or coexistent traumatic lesions. With these conditions, conventional radiography has historically been reported as the screening examination of choice for the injured spine (1). However, other imaging techniques have been proposed. Notably, the role of computed tomographic (CT) screening for cervical spine fractures has been extensively debated (5–8). On the other hand, to our knowledge, the role of CT as a screening examination in the assessment of thoracolumbar fractures has never been evaluated in large series, but it has been the subject of a few abstracts at radiology meetings (9,10). However, CT is recognized as a worthy tool for the characterization of thoracolumbar spine fractures (11–16). A few authors do recommend elective supplementary CT for assessment of the extent and stability of spinal fractures diagnosed at conventional radiography, with the results of conventional radiography enabling the CT examination to be focused on the appropriate region of interest (11–16). However, none of these articles concerns CT as a screening examination for thoracolumbar spine fractures, and none specifically addresses multi–detector row CT.
Nowadays, most severe trauma patients undergo a fixed thoracoabdominal CT protocol (multi–detector row CT at our institution) after the conventional radiographic survey. This fixed thoracoabdominal multi– detector row CT protocol is performed to rule out aortic and other visceral traumatic lesions. The purpose of our study was to determine if multi–detector row CT can replace conventional radiography and be performed alone for the depiction of thoracolumbar spine fractures in patients who have sustained severe trauma.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
In no case were conventional radiography of the thoracolumbar spine and thoracoabdominal multi–detector row CT or additional imaging examinations performed only for this study. This study protocol was approved by the review board of our institution, and institutional informed consent guidelines were followed.
Imaging Studies and Timing
Conventional radiographic survey of the spine consisted of anteroposterior and lateral views of the thoracic and lumbar spine with the patient in the supine position. Occasionally, a so-called swimmer’s view (a lateral radiograph of the cervical spine acquired with the patient having one arm up and one arm down) was obtained to clarify the status of the spine at the cervicothoracic junction. The total number of conventional radiographs obtained in each trauma patient, including those of body parts other than the spine, was recorded. Radiographs that had to be reacquired because of insufficient quality were quantified. Finally, the technologists were instructed to record the time needed to perform the conventional radiographic survey in each trauma patient, from the patient’s admission to the radiology unit to the time when images were available for viewing. The technologists were required to record the time dedicated to the spinal survey separately.
The thoracoabdominal multi–detector row CT survey (performed with a four–detector row Lightspeed CT unit; GE Medical Systems, Milwaukee, Wis) is typically performed after non-contrast material–enhanced cranial CT and is most often performed after cervical spine multi–detector row CT.
Our protocol for thoracoabdominal multi–detector row CT included series of thoracic and abdominopelvic images, both acquired in helical mode. The scanning directions for these thoracic and abdominal series were caudocranial and craniocaudal, respectively. The thoracic series consisted of 2.5-mm-thick CT sections (collimation, 4 x 2.5 mm) reconstructed at 2-mm intervals with a pitch of 1.5:1; acquisition parameters were 140 kVp and 200 mAs. The abdominal series, which included the pelvis, consisted of 5-mm-thick CT sections (collimation, 4 x 2.5 mm) reconstructed at 5-mm intervals with a pitch of 1.5:1; acquisition parameters were 120 kVp and 200 mAs. This thoracoabdominal multi–detector row CT acquisition was performed during and after intravenous administration of 100 mL of nonionic iodinated contrast material (iohexol, Accupaque 300; Nycomed, Oslo, Norway). Finally, when possible, the patient’s arms were raised over his or her head and image acquisition was performed during suspended respiration or superficial breathing.
After this thoracoabdominal multi–detector row CT acquisition was performed, the raw data were used to reconstruct transverse 2.5-mm-thick CT sections every 2 mm with a field of view adequate for visualization of the spine, as well as sagittal and coronal reformats of the thoracolumbar spine.
The CT technologists were instructed to record the time it took to perform the multi–detector row CT survey in each trauma patient. They were also required to record separately the time dedicated to patient transportation, multi–detector row CT data acquisition, and multi–detector row CT data postprocessing (reformats and filming) separately.
Review of Images
The review process was designed to reproduce the situation in the emergency radiology unit as closely as possible. Three radiologists (M.W., N.T., P.S.) and two orthopedic surgeons (E.M., P.M.) reviewed the conventional radiographs of the spine independently, whereas the same three radiologists reviewed the multi–detector row CT images independently. Orthopedic surgeons were not required to review multi–detector row CT images because they do not routinely read them in emergency settings.
Special attention was devoted to making the patient data anonymous to ensure blinding, because conventional radiographs and multi–detector row CT images were reviewed by the same three radiologists. Moreover, for the same reason, the reviews of conventional radiographs and multi–detector row CT images were performed successively, with a 1-month delay between review sessions. Finally, the reviewers, who performed the readings independently, were blinded to each other’s identity as well as to the results of initial clinical evaluation of the images.
The review of multi–detector row CT images was not performed with films, but at CT workstations. The raw data were evaluated with a postprocessing software program (Volume Analysis; GE Medical Systems) that enabled visualization of not only transverse images but also two-dimensional multiplanar reformats and three-dimensional surface-shaded display images. The reviewer chose the format of each image according to the particular situation of each patient and the reviewer’s preferences. We could not perform interpretation of soft-copy conventional radiographs because our emergency radiology unit was not equipped with phosphorus plates at the time of this study.
Reviewers were asked to evaluate images for the presence or absence of fractures and their location, including that of both the affected portion(s) of the vertebra and the involved vertebral level. Fractures were diagnosed on the basis of strict criteria: the presence of cortical interruption without sclerotic borders. Reviewers had to decide whether the fractures were stable according to the three-column concept (3,4). Finally, they had to mention which vertebral levels could or could not be evaluated at each examination and to judge the global image quality of the examination on a binary scale (0 = not interpretable, 1 = interpretable).
Statistical Analysis
Once each reviewer had evaluated all the conventional radiographs and all the multi–detector row CT images, patients’ clinical files were reviewed in consensus by one radiologist (M.W.) and one orthopedist (E.M.). The definitive presence or absence of thoracolumbar spine fractures was decided on the basis of clinical evolution, results of any repeated imaging examinations (conventional radiography, CT, magnetic resonance [MR] imaging), results of any orthopedic intervention, and/or final diagnoses as reported in the discharge report (or autopsy report, for seven of the 100 patients). Moreover, all the surviving patients underwent delayed orthopedic follow-up with clinical examination, as well as additional imaging examinations when clinically justified (39 of 89 patients). Additional imaging studies were performed either to evaluate for evolution of known fractures, or, in cases of back pain, to rule out an initially missed fracture. This represented a clinical follow-up that ranged from 54 to 163 days (interquartile range) (median, 112 days) for the surviving patients. Autopsy was performed in seven of 11 patients who died. In the four patients who died but in whom autopsy was not performed, the presence or absence of thoracolumbar spine fractures was diagnosed in consensus by all the reviewers after they reevaluated all images obtained in these four patients. Note that the reviewers were blinded to the diagnoses in the discharge or autopsy reports at the time of review of conventional radiographs and thoracoabdominal multi–detector row CT images.
The thoracoabdominal multi–detector row CT images obtained in each trauma patient were also reviewed for associated nonspinal traumatic lesions.
The sensitivity of conventional radiography and multi–detector row CT for thoracolumbar spine fractures were calculated with a weighting coefficient, taking into account the consensus or divergent opinion of the five or three reviewers. On the conventional radiographs, the same fractured vertebra could be assigned from zero to five positive scores from five reviewers, yielding weighting factors of 0%, 20%, 40%, 60%, 80%, and 100%, respectively. Similarly, on the multi–detector row CT images, a fractured vertebra could be assigned from zero to three positive scores from three reviewers, yielding weighting factors of 0%, 33%, 66%, and 100%, respectively. Interobserver agreement was evaluated with the Cohen 
coefficient (17).
In the Results section, mean sensitivity and interobserver agreement values are presented together with their exact binomial CI at the 95% confidence level. Statistical significance was determined with Fisher exact testing, and a P value of less than or equal to .001 was considered to represent a statistically significant threshold (according to results of a Bonferroni correction that allowed for 50 examinations with an overall significance level not exceeding .05).
Radiation Dose
The radiation dose involved in conventional radiography of the thoracolumbar spine and in thoracoabdominal multi–detector row CT was obtained from a recent national survey (18) in which our institution was deeply involved and which included experimental studies on test phantoms performed in conformity with U.S. Food and Drug Administration rules and the European Guidelines on Quality Criteria for Computed Tomography (19). The effective dose was chosen to express the risk represented by each examination. It allowed for the individual sensitivity of organs (different in men and women) to be taken into account, notably that of the gonads, which are exposed to x rays in both conventional radiography of the lumbar spine and abdominal multi–detector row CT. Exposure of the gonads to radiation was a major concern in the younger population of our series.
Financial Cost
We attempted to evaluate the direct resource costs involved in the performance of conventional radiography of the thoracic and lumbar spine and thoracoabdominal multi–detector row CT to get an idea of the amount of money that could be saved if the need to perform the conventional radiographic survey could be ruled out. The resource costs are relevant because they are used as a reference in refunding systems and have been reported to be the most appropriate values for making cost comparisons among providers (6). We used the same scheme proposed in an article about evidence-based evaluation of helical CT for the screening of cervical spine trauma (6).
Direct resource costs were divided into supply cost, labor cost, and equipment cost. Supply cost included the cost of film sheets, film storage jackets, and labeling. Labor cost was limited to the costs for technologist labor and radiologist interpretation. Technologist labor cost was determined according to computerized radiologic information, as recorded by the technologists themselves on dedicated forms. Assessment of opportunity cost, patient transportation staff cost, and archive management cost was beyond the scope of this article. Imaging room purchasing, installation, and maintenance costs were not included.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Of these 100 severe trauma patients, 26 had thoracolumbar spine fractures that were diagnosed according to the reference follow-up criteria detailed above. Among the 1,700 reviewed vertebrae, 67 fractured vertebrae were identified. Each patient with thoracolumbar spine fractures had an average of 2.6 fractured vertebrae.
These 100 patients were involved in severe trauma, as also evidenced by the 146 associated nonspinal traumatic injuries diagnosed in 62 of the patients (craniocerebral trauma in 28, lung injuries in 24, thoracic or abdominal wall trauma in 20, pneumothorax in 17, pelvic injuries in 11, hemothorax in seven, splenic fracture in seven, liver injury in five, pneumomediastinum in four, hemomediastinum in four, kidney lesion in three, free peritoneal fluid in two, aortic rupture in one, diaphragmatic rupture in one, and other injuries in 15 patients). Patients with spine fractures had more associated nonspinal traumatic lesions than did those without spine fractures (19 [73%] of 26 vs 26 [35%] of 74; P < .001). No special injury pattern could be identified as being specifically associated with spinal trauma. Instead, the injury pattern reflected the violence of the trauma.
Conventional Radiography versus Multi–Detector Row CT for Assessment of Thoracolumbar Fractures
Overall sensitivity and interobserver agreement, respectively, for detection of thoracolumbar spine fractures were 32.0% and 0.661 at conventional radiography and 78.1% and 0.787 at multi–detector row CT (Tables 1 and 2). The sensitivity of multi–detector row CT was significantly higher than that of conventional radiography (P < .001) (Fig 1). The spine fractures were revealed to be independent variables. This was reflected by the absence of significant differences among sensitivity values calculated separately for patients with one fracture, patients with two fractures, and patients with more than two fractures.
|
|
|
In our series, 12 patients had unstable fractures of the thoracolumbar spine: One patient had one unstable fracture, three had two unstable fractures, and eight had more than two unstable fractures. Sensitivity and interobserver agreement, respectively, for detection of unstable fractures of the thoracolumbar spine were 33.3% and 0.368 at conventional radiography and 97.2% and 0.951 at multi–detector row CT (Tables 3 and 4). The sensitivity and interobserver agreement at multi–detector row CT were excellent (close to 100%) and much higher than those at conventional radiography (P < .001). An average of eight patients appeared to have no fracture (Fig 2) or no unstable fracture (Fig 3) at conventional radiography but were revealed to have unstable fractures at multi–detector row CT. These patients thus constituted at-risk patients.
|
|
|
|
|
Quality of Conventional Radiographs and Multi–Detector Row CT Images
An average of 11.4 conventional radiographs, including 4.3 views of the thoracic and lumbar spine, were obtained in each of 100 severe trauma patients in our study. This latter number was greater than four because swimmer’s views were obtained in some cases and because some inadequate radiographs had to be reacquired. Thirty-five (9%) of the 395 thoracolumbar spine radiographs, versus 34 (5%) of the 676 nonspinal radiographs, had to be reacquired because of insufficient quality; the difference was statistically nonsignificant (P = .350).
Regarding the image quality of conventional radiographs of the thoracolumbar spine, the reviewers rated 103 (98%) of the 105 anteroposterior views of the thoracic spine, 92 (81%) of the 113 lateral views of the thoracic spine, 102 (99%) of the 103 anteroposterior views of the lumbar spine, and 105 (96%) of the 109 lateral views of the lumbar spine as interpretable. Lateral views of the thoracic spine thus constituted the radiographs that were most often of limited quality.
Images from the multi–detector row CT examinations were qualified as interpretable in 100% of cases. Motion artifacts or misregistration artifacts occurred, but in no case did such artifacts substantially alter the diagnostic value of multi–detector row CT results. Fractures could easily be distinguished from nutrient artery channels and other fracture-mimicking entities.
Vertebral levels that could be evaluated extended from T1 to L5 on anteroposterior conventional radiographs and from a median of T4 (interquartile range, T2–T6) to L5 on lateral conventional radiographs. All vertebral levels were visible at multi–detector row CT.
Time for Performing Conventional Radiography and Multi–Detector Row CT of the Thoracolumbar Spine
Times for performing the conventional radiographic survey in severe trauma patients are summarized in Table 5. The median time needed to perform conventional radiography was 33 minutes, with 70% (23 of 33 minutes) devoted to imaging the thoracolumbar spine. The rather long median time to perform this survey can be explained by the inadequate radiographs that have to be reacquired, but is also related to interruptions caused by restless patients, nursing interventions, and monitoring of and resuscitation maneuvers in unstable patients. The interquartile range was higher for nonspinal radiographs than for spinal radiographs because the number and type of radiographs acquired varied greatly from one patient to another according to the injuries suspected to be present in each patient.
|
Radiation Dose
The effective doses involved in conventional radiography of the thoracolumbar spine are reported in Table 6.
|
Cost of Conventional Radiography of Thoracic and Lumbar Spine
The supply cost for two views of the thoracic spine and two views of the lumbar spine amounted to $25 per patient. Technologist labor cost and radiologist interpretation cost, respectively, amounted to $80 and $40 per patient. The total resource cost that could be spared if conventional radiography is not performed and multi–detector row CT is performed alone is thus $145 per patient.
In comparison, the total resource cost for thoracoabdominal multi–detector row CT in our study was $880, with, however, no incremental cost resulting from the reformatting of thoracolumbar spine images from thoracoabdominal multi–detector row CT data, because neither additional imaging nor soft-copy interpretation had to be performed.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The quality of multi–detector row CT images is superior to that of conventional radiography. Multi–detector row CT allows visualization of the entire thoracolumbar spine in one piece (even if data acquisition is performed successively for the chest and then the abdomen). Use of multi–detector row CT thus eliminates the identification errors regarding the fractured vertebral level that were encountered in 13% of cases at conventional radiography, especially for old osteoporotic fractures. Multi–detector row CT image quality is not impaired by the projection of soft tissues or bone structures that impairs conventional radiographic image quality.
Even though conventional radiographs were judged as being of excellent quality overall in our series, the lateral radiographs of the thoracic spine systematically failed to display the first three thoracic vertebrae adequately. Although the use of phosphorus plates should somewhat improve this situation, this failure constitutes a major drawback, since the cervicothoracic junction is one of the peak areas of spinal injury. Moreover, this region is particularly prone to spinal cord damage, because of the small size of the spinal canal in the upper thoracic spine (1).
Not performing conventional radiography in the screening of thoracolumbar spine fractures in severe trauma patients affords several major advantages relating to imaging examination time, the need for patient manipulation, radiation dose, and cost management.
In our series of severe trauma patients, the time saved in not performing conventional radiography of the thoracolumbar spine amounted to 23 minutes, representing 32% of the 73 minutes it took to perform the entire classic imaging survey (conventional radiography plus multi–detector row CT) in our severe trauma patients. Such a decrease in imaging examination time has considerable importance with respect to the concept of the "golden hour"—the critical period in which definitive treatment of severe trauma patients should begin. Anything that delays treatment, such as prolonged imaging examination time, can adversely affect the patient’s prognosis (20).
Performing multi–detector row CT only rather than conventional radiography followed by multi–detector row CT diminishes the need for patient manipulation and its resulting hazards with respect to unstable fractures. As many as 5%–10% of spinal cord injuries occur in the early postfracture period (1,2). All precautions must therefore be taken to prevent any further injury to the patient during the diagnostic phase of care. From this point of view, multi–detector row CT enables one to perform a complete imaging survey that includes the brain, cervical spine, chest, abdomen, and thoracolumbar spine while the patient remains on the trauma board, with minimal mobilization.
Reformatting of thoracolumbar spine images from thoracoabdominal multi–detector row CT data involves no incremental dose, since no additional imaging is performed. According to the results of this study, not obtaining conventional radiographs of the thoracolumbar spine, which proved to be of lower diagnostic value than multi–detector row CT images, will result in a 24.7% (6.36 mSv divided by 25.78 mSv) decrease in effective dose compared with the effective dose required with the current dual-examination protocol.
Eliminating the practice of using conventional radiography to screen for thoracolumbar spine fractures in severe trauma patients will lead to a substantial saving of $145 per patient in terms of direct resource cost.
We acknowledge that our study had several limitations. The main limitation relates to the absence of a real standard of reference for the diagnosis of spine fractures except autopsy results, which were obtained in only seven patients in our series. In the surviving patients, we arbitrarily chose clinical evolution, as recorded during the patient’s hospitalization, and results of orthopedic follow-up, including the results of imaging examinations obtained during this follow-up, as a standard of reference. The initial multi–detector row CT results were included in the standard of reference; this may represent a bias. However, all patients with multi–detector row CT results that were positive for spine fractures underwent repeated imaging examinations (conventional radiography, multi–detector row CT, and MR imaging) during their orthopedic follow-up. Results of all these examinations were taken into consideration in the final diagnosis of spine fractures. For evident ethical reasons, it was not possible to perform systematic imaging studies in all surviving patients just for the purposes of this study because of concerns about radiation dose. In all cases and at all imaging examinations performed during follow-up, strict criteria were used in the diagnosis of spine fractures, and experienced reviewers were involved. Even if some fractures may have been missed, the reference standard protocol we used exactly reflects everyday clinical practice in emergency departments. The fractures that were potentially missed would have been those that involved no pain or secondary complications and necessitated no treatment (ie, those without clinical evidence).
The reformatted images of the thoracolumbar spine that were extracted from thoracoabdominal multi–detector row CT data were sections with 2.5-mm thickness and 2-mm reconstruction intervals. The use of thinner sections might have yielded increased accuracy. However, the use of thinner sections was not justified by the primary purpose of the thoracoabdominal multi–detector row CT survey, which was to enable the ruling out of aortic and other visceral traumatic lesions.
The number of reviewers was not the same for conventional radiographs and for multi–detector row CT images. This may have had an effect on the calculation of interobserver agreement.
Another limitation is that the times required to perform conventional radiography and thoracoabdominal multi–detector row CT, as well as the costs of these examinations, were determined on the basis of data from forms completed by technologists. The possibility of error in each of such recorded values must therefore be kept in mind.
Additionally, we limited our analysis to direct resource costs. We did not take into account costs that cannot be measured accurately and must be estimated (ie, overhead costs and capacity costs).
The radiologic evaluation of spinal trauma in severe trauma patients is a diagnostic challenge. Multi–detector row CT is better than conventional radiography for the identification of thoracolumbar spine fractures; it can replace conventional radiography and be performed alone in severe trauma patients. Radiography may still be indicated in selected cases—for instance, when severe trauma patients are too hemodynamically unstable to undergo a multi–detector row CT survey—since it can be performed during the acute phase of injury when patients are on the resuscitation table. Finally, MR imaging remains mandatory for direct imaging of damage to the spinal cord.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. W. Stuhlfaut, B. C. Lucey, J. C. Varghese, and J. A. Soto Blunt Abdominal Trauma: Utility of 5-minute Delayed CT with a Reduced Radiation Dose Radiology, February 1, 2006; 238(2): 473 - 479. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. C. Lucey, J. W. Stuhlfaut, A. R. Hochberg, J. C. Varghese, and J. A. Soto Evaluation of Blunt Abdominal Trauma Using PACS-Based 2D and 3D MDCT Reformations of the Lumbar Spine and Pelvis Am. J. Roentgenol., December 1, 2005; 185(6): 1435 - 1440. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. W. Stuhlfaut, G. Barest, O. Sakai, B. Lucey, and J. A. Soto Impact of MDCT Angiography on the Use of Catheter Angiography for the Assessment of Cervical Arterial Injury After Blunt or Penetrating Trauma Am. J. Roentgenol., October 1, 2005; 185(4): 1063 - 1068. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R Watura, M Cobby, and J Taylor Multislice CT in imaging of trauma of the spine, pelvis and complex foot injuries Br. J. Radiol., December 1, 2004; 77(suppl_1): S46 - S63. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. W. Stuhlfaut, J. A. Soto, B. C. Lucey, A. Ulrich, N. K. Rathlev, P. A. Burke, and E. F. Hirsch Blunt Abdominal Trauma: Performance of CT without Oral Contrast Material Radiology, December 1, 2004; 233(3): 689 - 694. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. E. Roos, P. Hilfiker, A. Platz, L. Desbiolles, T. Boehm, B. Marincek, and D. Weishaupt MDCT in Emergency Radiology: Is a Standardized Chest or Abdominal Protocol Sufficient for Evaluation of Thoracic and Lumbar Spine Trauma? Am. J. Roentgenol., October 1, 2004; 183(4): 959 - 968. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
