| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Kurt Amplatz Center, Innsbruck University Hospital,
Innsbruck 6020, Austria.
2 Department of Radiology II, Innsbruck University Hospital, Anichstrasse 35,
Innsbruck 6020, Austria.
3 Department of Vascular Surgery, Innsbruck University Hospital, Innsbruck 6020,
Austria.
4 Department of Trauma Surgery, Innsbruck University Hospital, Innsbruck 6020,
Austria.
Received May 11, 2004;
accepted after revision January 31, 2005.
Address correspondence to M. Rieger.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Over 36 months, 87 patients (16 females and 71 males; age range, 16-87 years) with clinically suspected arterial injury after extremity trauma underwent 4-MDCT angiography and 67 ultimately underwent surgery. Eighty patients had blunt injuries, and seven had penetrating injuries. The presence of arterial involvement was investigated prospectively by the radiologist in charge and retrospectively by two independent radiologists. Each detected arterial lesion was then characterized as a spasm, stenosis, occlusion, or rupture. The standard of reference was surgery in 67 patients, angiography in two patients, and clinical and radiologic follow-up findings in 18 patients. MDCT angiography was assessed by means of receiver operating characteristic (ROC) curve analysis for lesion detection and Spearman's rank correlation test for lesion characterization. Image quality, lesion depiction, and artifacts were subjectively assessed.
RESULTS. Sixty-two traumatic arterial lesions were confirmed at
surgery in 55 patients. MDCT angiography yielded high accuracy in detection
(area under the ROC curve [Az] = 0.96; p <
0.001) and characterization (r = 0.94; p < 0.001) of
traumatic arterial injuries and in recognizing an underlying dissection
(Az = 0.82; p < 0.001). Prospective
sensitivity and specificity were 95% and 87%, respectively, and retrospective
sensitivity and specificity were 99% and 87%, respectively. MDCT angiography
was considered to be sufficient for a reliable diagnosis in 83 patients
(p < 0.001). Image quality and lesion depiction on MDCT angiograms
were considered to be good and artifacts were considered mild with substantial
interobserver agreement (
, 0.62-0.69).
CONCLUSION. MDCT angiography provides significant and reproducible technique for the detection and characterization of arterial injuries to the extremities with high image quality and vascular delineation.
Keywords: angiography • arteriography • CT • extremities • MDCT • trauma
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The use of single-detector CT angiography for the evaluation of suspected vascular abnormalities coincided with noninvasive diagnosis and brought further advantages over digital subtraction angiography including 3D visualization of the examined vasculature; shorter acquisition time; and, most importantly, simultaneous visualization of vascular, muscular, and bone structures. Indeed, the advantages of single-detector CT angiography improved the detection and characterization of traumatic vascular lesions in general [11-16] and yielded promising results regarding the evaluation of arterial injuries of the proximal extremities with a high sensitivity (95%) and specificity (99%) [17, 18].
Owing to its advantages [19-21] over single-detector CT angiography, MDCT angiography has been implemented in evaluating large vascular territories because of its fast image acquisition [22, 23] and in assessing small arteries thanks to its ability of thin-section acquisition [24, 25]. Furthermore, MDCT has been recommended for the evaluation of polytrauma patients because it provides a comprehensive diagnostic workup within a relatively short stay in the CT suite [26, 27].
To our knowledge, the role of MDCT angiography in evaluating traumatic arterial injuries of the extremities has not yet been reported. The purpose of this study was to retrospectively assess the accuracy of MDCT angiography as the initial diagnostic technique to depict arterial injury in patients with extremity trauma.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
After patients were admitted to the emergency department, a full clinical examination, particularly of the involved extremity, was performed and further management was decided according to the category of the traumatic vascular injury. That is, for definitive vascular injury (pulsatile bleeding, rapidly expanding hematoma, bruit or thrill over wound, absent pulse with obvious trajectory from penetrating lesion), surgery was performed; for suspected vascular injury (excessive nonpulsatile bleeding, large nonexpanding hematoma, major neurologic deficit, diminished but appreciable pulse), MDCT angiography was performed; and for low likelihood of vascular injury (proximity injury, penetrating wound in proximity to vascular structures without clinical findings to suggest vascular compromise), sonographic follow-up was performed. Provided no immediate surgical intervention was required to maintain hemodynamic stability, patients underwent MDCT angiography based on an interdisciplinary decision by the examining anesthesiologist, traumatologist, vascular surgeon, and CT radiologist.
The traumatic injury involved the upper extremity in 36 patients and the lower extremity in 51. Blunt injuries were found in 80 patients and penetrating injuries in seven. In all cases, MDCT angiography was performed preoperatively to determine the management strategy. Sixty-seven patients underwent surgery after MDCT angiography. Examinations performed for the evaluation of postoperative results were excluded from this study. None of the patients had contraindications to IV injection of nonionic contrast material. Our institutional review board approved the study, and informed consent for MDCT angiography was obtained from all patients.
MDCT Angiography
Helical CT was performed on a 4-MDCT scanner (LightSpeed QX/i, GE
Healthcare) with a 0.8-sec gantry rotation period. MDCT angiography was
performed with the patient in the supine position with arms along the body,
except in five patients who sustained a traumatic injury to only the forearm
or hand. In the latter cases, the patient was placed in the prone position
with the involved arm extended over the head. In all cases, the involved
extremity was immobilized by means of adhesive tape. MDCT angiography images
were acquired, in a proximal-to-distal direction, using 1.25-mm collimation,
1.25-mm reconstructed slice thickness, and the standard reconstruction kernel.
The pitch and the reconstruction interval were adjusted depending on the
anatomic region of interest (Table
1). When more than one anatomic region was examined, a pitch of
1.5 and reconstruction interval of 0.8 mm were used. To improve axial
resolution, we used the smallest possible display field of view and the
display field of view captured only the involved extremity.
|
X-ray tube voltage and current were adjusted for the examined part of the
extremity and ranged from 100 to 140 kV and from 80 to 200 mA. Similarly,
depending on the patient's weight and scan coverage, a total volume of 100-150
mL of nonionic contrast material was IV administered at a rate of 3-5 mL/sec
according to the following formula:
 |
Image Analysis
After acquisition, MDCT data sets were transferred to a PACS and to a 3D
rendering workstation (Ultra 60, Sun Microsystems) running the Advantage
Windows software (Advantage Windows 4.0, GE Healthcare), both of which are
located in the emergency CT suite. Prospectively, MDCT angiograms were
evaluated by the radiologist in charge of the emergency CT suite. While the
radiologist evaluated the axial images on a dedicated PACS workstation to
assess the bone, muscular, and vascular structures of the examined region, the
CT technical assistant used the rendering workstation to create 3D
reconstructions—that is, multiplanar reformations (multiplanar
reconstructions) and volume-rendered views that were interactively interpreted
by the radiologist.
The prospective evaluation was performed on an emergency basis. Therefore, a blinded double-reviewer evaluation was not always available before surgery, particularly at night and on weekends. For this reason, all MDCT examinations were reevaluated in a retrospective fashion within 48 hr by two independent radiologists with 7 and 4 years' experience in CT angiography. Neither reviewer participated in the prospective analysis, and both were blinded to surgical, clinical, and radiologic outcomes. The reviewers evaluated the axial source images and reconstructed each data set to create multiplanar reconstruction and volume-rendered views that were interactively assessed on the rendering workstation. At prospective analysis, the radiologist determined whether arterial involvement was present or not. When an arterial lesion was determined to be present, it was then classified as an arterial spasm, stenosis caused by compression, occlusion, pseudoaneurysm, arteriovenous fistula, or arterial rupture. Arterial spasm was defined as a concentric, smooth, tapered narrowing of the involved artery, and stenosis was defined as an eccentric, irregular narrowing of arterial lumen caused by an extraluminal abnormality (intramural hematoma, soft-tissue hematoma, bone fragment, or foreign body). Atherosclerotic stenoses were not included in this study.
Retrospectively, the diagnostic confidence in the presence of a vascular injury was scored on a 5-point scale, where 1 stood for definitely present; 2, probably present; 3, uncertain; 4, probably absent; or 5, definitely absent. When an arterial lesion was categorized as definitely present or probably present, it was then classified according to the same system used prospectively (i.e., arterial spasm, stenosis caused by compression, occlusion, pseudoaneurysm, arteriovenous fistula, or arterial rupture). Further, the presence of an underlying arterial dissection was sought in those patients with an arterial injury. The diagnostic confidence in the presence of an underlying dissection was scored on a 5-point scale (i.e., from 1, definitely present, to 5, definitely absent). Criteria for diagnosing a dissection included the eccentric stenosis of the involved artery causing a semilunar deformation of the lumen, dissection flap, or segmental thrombotic occlusion in the injured region.
In addition, quantitative and semiquantitative assessments of the quality of the MDCT angiograms were performed by the two reviewers. The quantitative assessment was performed by measuring the arterial attenuation at the proximal and distal sections of the examined anatomic region by each reviewer. The mean attenuation value at each level was then calculated for each examination. The semiquantitative assessment of MDCT angiograms included overall image quality, lesion depiction, and artifacts. The impression of overall image quality was assessed as follows: 1, very good demonstration of vascular anatomy enabling comprehensive and reliable evaluation; 2, good demonstration of vascular anatomy enabling adequate evaluation; 3, reasonable demonstration of vascular anatomy with inadequate evaluation; or 4, unsatisfactory, barely visible vascular anatomy. Depiction of the vascular lesion was scored as follows: 1, sufficient depiction of lesion features enabling definite characterization; 2, barely recognizable lesion features with limited lesion characterization; or 3, ambiguous lesion features with impossible characterization. The presence of artifacts was scored as follows: 1, none; 2, minor; 3, mild, partially affecting diagnostic evaluation; or 4, major, substantially affecting diagnostic evaluation.
To determine their impact as a complement to the axial source images on the perception of the arterial injury, multiplanar reconstruction and volume-rendered views were categorized by reviewers as 1, enabled diagnostic information that was not disclosed on source images; 2, facilitated the evaluation of the arterial vasculature; or 3, provided no advantage over source images. In comparison with surgical and clinical findings, the reviewers also assessed the overall impact of MDCT angiography on the final diagnosis as sufficient for establishing a diagnosis or insufficient, with additional imaging techniques (i.e., digital subtraction angiography) required.
Standard of Reference
In this study, the standard of reference was surgery with inspection of the
involved arterial segment in 67 patients; digital subtraction angiography in
two patients; and, in the remaining 18 patients who did not require surgery or
digital subtraction angiography, clinical and radiologic follow-up findings.
Clinical follow-up included physical examination of the involved extremity,
oscillometric arterial blood pressure measurements, pulse wave propagation
velocity, and skin capillary blood circulation; radiologic follow-up included
Doppler sonography in all patients. The mean follow-up was 6 days (range, 2-14
days). Surgical results and clinical and radiologic follow-up outcomes were
determined by reviewing patients' data files stored in the hospital's digital
database by one of two surgeons.
Statistical Analysis
Data entry procedures and statistical analysis were performed with a
statistical software system (SPSS version 11.0.0, Statistical Package for the
Social Sciences) for Windows (Microsoft). In the first step, prospective data
were analyzed for lesion detectability by calculating the sensitivity,
specificity, and accuracy and for lesion characterization by calculating the
correlation to the standard of reference by means of Spearman's rank
correlation test. In the second step, retrospective data were analyzed.
Receiver operating characteristic (ROC) curves were generated to show the
relative accuracy of MDCT angiography in detecting vascular injuries of the
extremities by comparing the areas under the ROC curves
(Az) calculated by each reviewer. From the observed data
points, sensitivity, specificity, and accuracy were calculated on a per-lesion
(i.e., the ability to correctly identify all vascular lesions) basis whereby
"definitely present" and "probably present" were
positive and all other categories negative. ROC analysis was also applied to
evaluate the detectability of an underlying dissection. Spearman's rank
correlation test was used to investigate the correlation between lesion
characterization at MDCT angiography and that confirmed by the standard of
reference.
Ordinal data arising from the score modes of image quality, lesion depiction, and artifacts were then evaluated to assess the performance of MDCT angiography by calculating the means and 95% confidence intervals (CIs). Ordinal data arising from the comparative analysis of multiplanar reformations, volume-rendered images, and axial source images were tested with the nonparametric binominal test.
Finally, the kappa statistic was used to compare observer performance.
Kappa values were calculated on the basis of each reviewer's confidence level
for the ROC analysis. Interobserver agreement was considered slight if
< 0.2; fair, = 0.21-0.40; moderate, = 0.41-0.60;
substantial, = 0.61-0.80; or almost perfect, = 0.81-1.00
[28].
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
= 0.78) and in the detectability of an underlying
dissection ( = 0.60).
|
|
|
Characterization of Arterial Injury
MDCT angiography permitted correct characterization of most of the detected
arterial injuries (Figs. 4A,
4B,
5,
6,
7A,
7B, and
7C). The Spearman's rank
correlation test showed a statistically significant correlation between MDCT
angiography and the reference standard for prospective analysis (both
reviewers: r = 0.93, p < 0.001; reviewer 1, r =
0.95, p < 0.001; reviewer 2, r = 0.94, p <
0.001). Prospectively, nine findings were misclassified: three cases of spasm
that were not identified, two other cases that were believed to be occlusion
but were spasm at surgery, one case believed to be an occlusion but was normal
at surgery, and three cases believed to be a spasm but were normal. Reviewer 1
misclassified six findings (two cases that were believed to be occlusion but
were spasm at surgery, three other cases that were believed to be a spasm but
were normal, and one case believed to be occlusion but that was normal at
surgery), and reviewer 2 misclassified seven findings (one case of a spasm
that was not identified, two cases that were believed to be an occlusion but
were spasm at surgery, three other cases that were believed to be a spasm but
were normal, and one case believed to be occlusion but that was normal at
surgery). For lesion characterization, interobserver agreement was almost
perfect ( = 0.95 at retrospective analysis; = 0.92, reviewer 1
vs prospective analysis; = 0.91, reviewer 2 vs prospective
analysis).
|
|
|
|
|
= 0.62), lesion depiction ( =
0.68), and the presence of artifacts ( = 0.69). Minor to mild artifacts
that did not compromise the diagnosis occurred in a few cases and were due
either to patient movement during CT acquisition, mostly seen while examining
the forearm and hand or the lower leg and foot, or to positioning the upper
extremity along the patient's body.
With regard to the impact of postprocessing algorithms on perception of the arterial injury, multiplanar reformations facilitated the evaluation of 53 MDCT angiograms and provided no advantage over axial images in 34 (p = 0.054, nonparametric binominal test). In comparison, volume rendering facilitated evaluation of the arterial vasculature in 72 patients, but provided no diagnostic advantage in 15 patients (p = 0.001, nonparametric binominal test). Neither multiplanar reconstruction nor volume-rendered views provided, in any case, diagnostic information that had not already been disclosed on axial source images. Multiplanar reconstruction images were superior to volume-rendered views in seven patients because they permitted detection of an arterial dissection, whereas volume-rendered views were preferable to multiplanar reconstruction images in 66 patients because they gave a more comprehensive visualization of the examined region. In 14 patients, multiplanar reconstruction and volume-rendered views were equal in performance. Accordingly, the nonparametric binominal test revealed that axial source images are significantly (p < 0.001) more reliable than multiplanar reconstruction and volume-rendered images for diagnosis of traumatic arterial injuries of the extremities and that volume-rendered views are significantly (p < 0.001) more useful than multiplanar reconstruction images.
In terms of its impact on final diagnosis, MDCT angiography was considered
by reviewers 1 and 2 to be sufficient for reliable diagnosis in 83 and 84
patients (p < 0.001, nonparametric binominal test), respectively,
whereas additional information obtained by digital subtraction angiography was
still required for four and three patients, respectively. Interobserver
agreement on the reliability of MDCT angiography as the initial investigation
of traumatic arterial lesions was almost perfect ( = 0.85).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In patients with traumatic injuries, particularly with suspected vascular involvement, both color Doppler sonography and MR angiography have a substantial limitation—that is, both require a considerable amount of time for diagnosis. Further, color Doppler sonography is operator-dependent, may be limited in the evaluation of arterial flow distal to an arterial injury, and may be inapplicable in patients with open trauma, and the presence of collateral vasculature may produce erroneous results. MR angiography is more susceptible to motion artifacts than is CT angiography and may not be practicable for patients with metallic fixation devices.
In this study, we assessed the clinical usefulness of MDCT angiography as the initial vascular imaging technique for patients with a traumatic injury to the lower or upper extremity. In all cases, MDCT angiography permitted comprehensive identification of the main arteries and their major branches in the examined region. Statistical analysis indicated that MDCT angiography permits accurate and reproducible detection of traumatic arterial injuries. Indeed, our early follow-up showed that no surgically relevant lesion was missed at MDCT angiography. Moreover, lesion characterization was correctly determined in most of the arterial injuries. However, detection of an underlying dissection of the injured segment was slightly less reliable, particularly in segments with an occlusion.
The presence and severity of an underlying dissection are important factors in determining a surgical strategy. Although focal intimal flaps that do not cause a hemodynamic effect on extremity perfusion do not require surgery [3], intimal tears with obliteration of the vascular lumen or with resultant thrombosis necessitate immediate treatment to restore blood flow and perfusion pressure distal to the injury. At our institution, intimal tears with complete arterial obliteration are treated surgically with a venous interposition graft (usually acquired from the greater saphenous vein), and intimal tears with consecutive thrombosis undergo thrombectomy or interventional revascularization.
At both prospective evaluation and retrospective review, false-positive and false-negative interpretations were made. The most miscategorized lesions at MDCT angiography were related to severe spasms that completely or partially abated before surgery. Interestingly, some spasms were so severe that they simulated complete obstruction at MDCT angiography. Indeed, four arterial segments were believed to be occluded at MDCT angiography (two by both reviewers and one by each reviewer), whereas surgery confirmed a patent artery. Similar situations have been reported [6, 17]. In such cases, we believe that continuous pulse monitoring and targeted Doppler sonography distal to the injury site may assist in recognizing spasm abatement before vascular surgery, which may preclude unnecessary surgery. In our study, patent arterial flow in the trauma region was confirmed in 12 patients during surgery. All 12 patients had injuries that necessitated open surgery during which arterial integrity was confirmed.
The results of our study show that MDCT angiography permits not only the proximal portions of the extremities to be imaged, as shown by Soto et al. [17, 18], but also the peripheral arterial supply of an extremity to be comprehensively shown with high image quality, good vascular delineation, and improved lesion depiction. Furthermore, the detection of arterial injuries of the hand and foot was also possible. Despite positioning the upper extremity along the patient's body, only minor to mild streak artifacts were noticed with no diagnostic compromise. Neutral position of the involved extremity was not required for MDCT angiography. The arbitrary position of the involved extremity, which was determined by the treating clinical team, did not complicate evaluation of the CT data set or recognition of anatomic relationships.
Several investigators have recommended the application of postprocessing techniques, such as multiplanar reformations and volume rendering, for more reliable evaluation of CT data sets [25, 31-34]. Our study showed that multiplanar reformations and volume-rendered views of the examined extremity facilitate evaluation of the vascular structures and their relationship to bone structures and have the potential to reduce image interpretation time as compared with interpretation of a large number of axial source scans. Furthermore, although we did not formally study the usefulness of the rendered views for surgeons, our impression was that such images were more readily appreciated by the surgeons than was a review of the axial images in the cine mode on a PACS console. However, all arterial lesions were adequately detected and characterized on the axial source scans, which had the advantage of revealing the dissection flap when present.
A main limitation of our investigation is its retrospective design. However, all examinations were performed on an emergency basis by the radiologic team, including a senior radiologist, and were acquired and reconstructed according to standardized protocols designed prospectively and modified slightly for patient weight and size of the region of interest. Another limitation of our study is that the standard reference, in most patients, was surgery that was performed within 1 hr after MDCT angiography, which produced an increase in the rate of false-positive arterial occlusions that were transient spasms. Although the injection of intraarterial vasodilators may help, the differentiation between severe spasm and occlusion may also be difficult at digital subtraction angiography. Also, surgery confirms MDCT angiographic findings only at the injury level rather than for the entire examined arterial system; therefore, we might have missed some injuries in arterial segments that had no proximity to the trauma region and were therefore not explored. Similar to digital subtraction angiography, the presence of a well-enhanced smooth arterial segment with no depictable intraluminal, mural, and extravascular abnormalities on MDCT angiography rules out a significant anatomic injury that requires surgical or interventional treatment. Furthermore, there might be a verification bias in our study because surgery was performed on the basis of MDCT angiographic findings in some cases. However, surgery was indicated in most patients because of fractures, soft-tissue defects, or foreign bodies. For the patients in whom surgery was not indicated, follow-up was used as the reference and therefore a single gold standard for all patients was not available in our study. Finally, the relatively small patient cohort, particularly those with arterial injury, and the lack of patients with pseudoaneurysm or arteriovenous malformation represent limitations of this study. Although all five patients with a penetrating injury in our study were adequately evaluated with MDCT angiography, their small number may not allow generalizing study results in such cases.
In summary, we conclude that MDCT angiography permits significant and reproducible detectability and characterization of traumatic arterial injury in the extremities with high image quality and vascular delineation. MDCT angiography has substantial potential as the initial diagnostic method in patients with suspected arterial injury after upper or lower extremity trauma.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  E. K. Fishman, K. M. Horton, and P. T. Johnson Multidetector CT and Three-dimensional CT Angiography for Suspected Vascular Trauma of the Extremities RadioGraphics, May 1, 2008; 28(3): 653 - 665. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Covey Invited Commentary RadioGraphics, May 1, 2008; 28(3): 665 - 666. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Karcaaltincaba MDCT angiography in patients with traumatic extremity injuries. Am. J. Roentgenol., July 1, 2006; 187(1): W129 - W129. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
