| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Geneva University Hospital, 24 rue Micheli-du-Crest,
Geneva 1211, Switzerland.
2 Department of Radiology, University of Maryland School of Medicine, Baltimore,
MD.
Received January 21, 2008;
accepted after revision July 22, 2008.
Address correspondence to P. A. Poletti
(pierre-alexandre.poletti{at}hcuge.ch).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Ninety patients with polytrauma underwent whole-body 16-MDCT with a conventional (n = 30) or one of two single-pass (n = 60) protocols. The conventional protocol included unenhanced scans of the head and cervical spine and contrast-enhanced helical scans (140 mL, 4 mL/s, 300 mg I/mL) of the thorax and abdomen. The single-pass protocols consisted of unenhanced scans of the head followed by one-sweep acquisition from the circle of Willis through the pubic symphysis with a biphasic (150 mL, 6 and 4 mL/s, 300 mg I/mL) or monophasic (110 mL, 4 mL/s, 400 mg I/mL) injection. Acquisition times and interval delays between head, chest, and abdominal scans were recorded. Contrast enhancement was measured in the aortic arch, liver, spleen, and kidney. Diagnostic image quality in the same areas was assessed on a 4-point scale.
RESULTS. Median acquisition times for the single-pass protocols were significantly shorter (-42.5%) than the acquisition time for the conventional protocol. No significant differences were found in mean enhancement values in the aorta, liver, spleen, and kidney for the three protocols. The image quality with both single-pass protocols was better than that with the conventional protocol in assessment of the mediastinum and cervical spine (p < 0.05). There was no significant difference between the single-pass protocols.
CONCLUSION. Use of single-pass continuous whole-body MDCT protocols can significantly decrease examination time for patients with polytrauma and improve image quality compared with a conventional serial scan protocol. Monophasic injection with highly concentrated contrast medium can reduce injection flow rate and should therefore be preferred to a biphasic injection technique.
Keywords: emergency radiology • MDCT • polytrauma
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The aim of our prospective study was to compare the acquisition times and image quality of two single-pass continuous whole-body MDCT protocols with those of our conventional segmental whole-body MDCT protocol. Two single-pass protocols with various injection flow rates and iodine concentrations were evaluated. The first, developed and adopted at the R. Adams Cowley Shock Trauma Center (Mirvis SE, Shanmuganathan K, presented at the 2005 annual meeting of the American Society of Emergency Radiology), consisted of a biphasic injection (6 and 4 mL/s) of contrast medium at a standard concentration (300 mg I/mL). The protocol is based on the hypothesis that two injection rates produce enhancement that is closer to uniform than that achieved with a monophasic injection. The second injection protocol consisted of a monophasic injection of highly concentrated (400 mg I/mL) contrast medium at a lower flow rate (4 mL/s) to reduce the risk of contrast extravasation and catheter rupture. This protocol was adapted from the biphasic protocol at our institution and also was evaluated.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Imaging Techniques
After initial cardiovascular stabilization, whole-body imaging was
performed with a 16-MDCT scanner (Mx8000, Philips Healthcare) at 120 kVp with
a variable tube charge current dose-modulation system (DoseRight, Philips
Health care). Injection of contrast medium was performed with 16- or 18-gauge
peripheral IV catheters only (Optiva, Medex Medical). Two experienced CT
technologists were scheduled together so that various tasks, such as patient
positioning, data acquisition, and multiplanar volume reconstruction, could be
performed simultaneously. A radiology resident using a separate workstation
(Cedara I-SoftView 6.1, Cedara Software) was present for each examination to
provide immediate image interpretation.
Group A: conventional segmental whole-body scan protocol—Group A underwent MDCT according to a standardized protocol for different body regions. Each patient was positioned on the CT table head first with arms at the sides. Sequential cranial CT scans were acquired from the vertex to the occipital condyle. When a facial bone fracture was clinically suspected or depicted on cranial scans, helical face scans were obtained from the vertex to the occipital condyle at thinner collimation (Table 1). A cervical scan was obtained from the skull base to the second thoracic vertebra. Once head and cervical images were reconstructed, thoracic and abdominal helical scans were obtained after positioning of the arms above the head. The arms were left down when upper limb or scapular girdle fracture was suspected at clinical examination, on the CT scout image, or on the admission chest radiograph. The injection protocol included 140 mL of IV contrast medium (iopromide, Ultravist 300, Bayer Schering Pharma) at a rate of 4.0 mL/s with an automatic triggering threshold of 150 HU at the level of the aortic arch. To achieve both an arterial phase for the chest imaging and a portal venous phase for upper abdominal imaging, a time delay was inserted between the last chest image and the first abdominal image. Abdominal scanning was triggered 60 seconds after the beginning of the contrast injection. The area included between the upper part of the liver and the posterior aspect of the 12th ribs (overlap region) therefore was scanned in both the arterial and portal venous phases. Conventional scan parameters are shown in Table 1.
|
Group B: single-pass continuous whole-body scan protocol—The single-pass continuous whole-body scan protocol was divided into two subgroups, B1 and B2, with different injection protocols at an automatic triggering threshold of 100 HU at the level of the aortic arch. In subgroup B1, the injection protocol was biphasic and began with a 90-mL bolus of iopromide (300 mg I/mL, Ultravist 300) at a rate of 6.0 mL/s that was immediately followed by a 60-mL bolus of the same agent at a rate of 4.0 mL/s. In subgroup B2, the injection protocol was a single 110-mL bolus of iomeprol (400 mg I/mL, Iomeron 400, Bracco) at a rate of 4.0 mL/s.
The first step for both subgroups was an initial sequential cranial acquisition (similar to protocol A) without contrast enhancement to evaluate for intracranial hemorrhage. Each patient was positioned on the CT table head first with arms at the sides. As in group A, the arms were repositioned above the head except when upper limb fracture was suspected or the patient was judged to have severe injuries. A contrast-enhanced whole-body single-pass acquisition from vertex to pelvis was performed with the scan parameters shown in Table 1.
|
Data Acquisition
Scan acquisition and reconstruction time— Image acquisition
times were recorded and represent the total time elapsed from acquisition of
the first scout image to acquisition of the last axial image. If delayed
imaging was performed, the additional acquisition time was not taken into
account in the scan acquisition time. Interscan delay, which was part of the
total acquisition time, was analyzed as follows: group A, time elapsed between
the last cervical image and the first chest image as well as the delay between
the last chest image and the first abdominal image; group B, time elapsed
between the last image of the cranial acquisition and the first image of the
whole-body acquisition.
The number of patients who needed delayed images was recorded. Mean and median times, SD, and 95% CI for the means were calculated for each examination. Arm positions (above head or at sides) also were recorded. Reconstruction times were recorded for conventional and single-pass protocols by scanning an adult phantom to avoid bias associated with various body types and sizes.
Image quality evaluation—After imaging was complete, two attending radiologists (9 and 12 years of experience in emergency imaging) independently interpreted the sets of images for each patient in a random manner and using the same workstation and the same visualization software (Cedara I-SoftView 6.1, Cedara Software). Both radiologists were unaware of demographics and the injection techniques used. Assessments were based on the diagnostic quality of images of the entire mediastinum (analysis of both vascular and nonvascular structures), cervical spine, liver, and spleen. Image quality was rated on the following 4-point scale: 0, poor image quality, not adequate to assess or rule out injury with confidence; 1, reduced image quality presenting difficulty in assessing or ruling out injury; 2, good image quality supporting confidence in assessing the diagnosis; 3, excellent image quality, no problems in assessing or ruling out injury. Consensus was determined to be the average of the two radiologists' scores. Objective enhancement was measured in HU for every patient for regions of interest placed in the aortic arch, liver, spleen, and renal cortex. In the solid organs, the region of interest was chosen in an area of parenchyma that had maximal density and was free of injury, vessels, and artifacts. Mean and SD enhancement values were calculated for each scan protocol. Total image values for each examination were recorded to determine average values for images obtained with protocols A and B.
Statistical Analysis
A two-tailed Student's t test was performed to assess the
significance of the observed differences between the conventional and
single-pass protocols (acquisition times with arms above the head or along the
body, contrast enhancement values, and image quality). A value of p
< 0.05 was considered statistically significant. Agreement of both re
viewers on image quality classification was assessed with linear weighted
kappa statistics obtained with MedCalc software (version 9.3 for Microsoft
Windows XP, MedCalc Software). The strength of agreement was categorized as
slight (
= 0.00-0.20), fair ( = 0.21-0.40), moderate ( =
0.41-0.60), substantial ( = 0.61-0.80), or almost perfect to perfect
( = 0.81-1.00) according to an adapted Landis and Koch method
[9].
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The mean ages of the study subjects were 37.5 years in group A and 39.8 years in group B. The median injury severity scores were 26 for group A and 24.5 for group B. Ten of the 90 patients (11%) included in the study had a liver laceration, eight (9%) a splenic laceration, nine (10%) a renal injury, 23 (25%) a spinal injury, and 39 (43%) a head or face injury. There were no significant differences in the frequencies of the various injuries in each subgroup.
Acquisition times for groups A and B are summarized in Figure 1 and Tables 2 and 3. Acquisition time decreased 42.5% when the single-pass protocol was used, and the median imaging time for group B was 7.9 minutes shorter than that for group A. The difference of means between the two groups was significant (p < 0.0001). In group A, eight patients (27%) needed an additional face scan, resulting in a mean acquisition time of 21.2 minutes for those patients. The mean acquisition time for the rest of the group was 17.7 minutes, and the difference of means was significant (p < 0.05). In group A, six patients (20%) underwent a delayed additional scan, adding 8.8 ± 5.8 minutes to the mean acquisition time. In group B, 12 patients (20%) underwent a delayed scan, adding 12.1 ± 5.1 minutes to the mean acquisition time. These differences were not significant (p > 0.05).
|
|
The interscan delays for groups A and B are summarized in Table 2. The mean delay between the last cervical image and the first thoracic image in group A accounted for up to 67% of the mean total acquisition time. During this delay, the mean time elapsed between the thoracoabdominal scout image and first thoracic image was 4.1 ± 2.7 minutes. In subgroups B1 and B2, the mean delay between the last cranial image and the first single-pass image accounted for up to 49.3% and 45.2%, respectively, of mean total acquisition time. The difference of means between groups A and B was significant (p < 0.0001).
Arms were positioned above the head for 19 patients (63%) in group A and 32 patients (53%) in group B with no significant difference (p > 0.05) in the mean acquisition times relative to arm positioning (Table 4). In patients with an injury severity score of 25 or greater, however, mean acquisition times were 54.6 and 72 seconds longer in groups A and B, respectively, when arms were positioned above the head rather than at the sides. Concerning interscan delay between the last cervical and head image and the first thoracic or single-pass image, there were no significant differences between group A and group B relative to arm positioning (p > 0.05).
|
The average number of images produced per study was 1,321.2 ± 533.03 for group A and 1,277.17 ± 513.65 for group B with no significant differences of means between the two groups (p > 0.05). For the conventional scan protocol, the image production rates (including processing and reconstruction) measured on a phantom were cranial scan, 1.12 images/s; facial scan, 1.81 images/s; cervical scan, 4.2 images/s; and thoracoabdominal scan, 2.4 images/s. For the single-pass protocol, the image production rate of the 3-mm-thick images with a 3-mm gap was 1.5 images/s. Therefore, a 3.3-minute time delay was required for image reconstruction of a 90-cm length (vertex to pelvis distance in our phantom model). A direct comparison could not be made with the conventional protocol because acquisitions of the various body segments were not performed in a continuous manner.
Mean enhancement values measured with regions of interest placed in the aortic arch, liver and spleen parenchyma, and renal cortex are displayed in Table 5 along with mean values of image quality score consensus between the two radiologists. To assess the effects of beam hardening resulting from arm position on image quality and mean enhancement values, results are displayed according to arm position. The degree of agreement (kappa) on image quality classification between reviewers was 0.7 ± 0.09 (standard error) for the entire mediastinum, 0.72 ± 0.08 for the liver, 0.66 ± 0.08 for the spleen, 0.75 ± 0.09 for the kidneys, and 0.77 ± 0.1 for the cervical spine. No statistical differences in enhancement were found between groups in any anatomic localization measured (p > 0.05). Nor were statistical differences found in enhancement values between the arms-up and arms-down subgroups for any anatomic location or for any scan protocol (p > 0.05).
|
For image quality evaluation based on the 4-point scale, p values reflecting the statistical differences between the consensus assessment of each scan protocol and between arms-up and arms-down subgroups were calculated. For all scan protocols, image quality at the liver, spleen, and kidneys was scored significantly higher when arms were positioned above the head (p < 0.05). In comparisons of corresponding subgroups (arms up and arms down) in the three scan protocols, no significant differences were found at these anatomic regions (p > 0.05). For the entire mediastinum, image quality also was scored significantly higher within the arms-up subgroups for the A and B1 scan protocols. Image analysis at the cervical spine level showed no significant difference in quality for the three scan protocols or in the arm position subgroups for the B1 and B2 scan protocols (p > 0.05).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Another factor increasing interscan delay in group A was the need for thoracic and abdominal scan programming with a mean time of 4.1 minutes between the scout image and first thoracic image. In group B, CT programming for head and whole-body scans was done only one time, at the beginning of the examination with a whole-body scout image. However, the interscan delay between chest and abdomen in protocol A was negligible (0.7 minutes) and did not contribute to increasing the total scan delay with regard to protocol B. Indeed, the smaller collimation thickness at acquisition in protocol B (0.75 mm) compared with protocol A (1.5 mm) reduced the scanning speed. This reduced speed, along with faster injection of the iodine content, explains why the head and chest are scanned in the arterial phase and the abdomen in the portal venous phase in a continuous helix. Therefore, the total scanning times for chest and abdomen are comparable between the two protocols.
Unlike many other investigators of CT workflow, we did not take into account the total duration of stay in the CT room because this time includes a number of factors independent of scanner and technologist performance. These factors (e.g., stabilization of hemodynamic status, management of complications, parallel clinical evaluation) can influence the total duration of stay in the CT room without providing information relevant to CT technique and efficacy. In a review of the medical literature, we identified only three studies [6, 10, 11] in which single-pass continuous whole-body MDCT protocols were used for polytrauma patients. These studies showed significant decreases in examination times. Ptak et al. [6] found that acquisition times decreased by a factor of 10, but those investigators compared results from 4-MDCT and single-detector helical CT of a small (n = 10) patient group. Using 4-MDCT, Heyer et al. [10] evaluated 80 polytrauma patients and found acquisition time decreased 62%. The lower time reduction factor in our study (42.5%) may be explained by the fact that our single-pass scan protocol included an additional initial head scan (not obtained by Heyer et al.) without contrast administration to evaluate for brain injury. In a study of 16-MDCT, Fanucci et al. [11] found total time spent in the CT examination room was 31% lower for a single-pass protocol than for a conventional protocol.
To our knowledge, the literature indicates no agreement on contrast injection proto cols for single-pass continuous whole-body MDCT in polytrauma patients. In this study, we hypothesized that optimal enhancement of both visceral organs and vessels can be achieved with a biphasic injection protocol with two injection rates (6.0 and 4.0 mL/s) to produce more constant enhancement than can be achieved with a monophasic injection. Awai et al. [12] found that biphasic injection was more suitable than monophasic injection for whole-body MDCT tumor screening. Using a contrast agent with a concentration of 300 mg I/mL, Awai et al. found that the liver and infrarenal inferior vena cava displayed better enhancement with biphasic than with monophasic injection.
In our study, extravasation of contrast material into the subcutaneous tissues occurred in two patients (6.25%) in subgroup B1. The 18-gauge IV catheter did not support the injection rate of 6.0 mL/s. Biphasic injection is therefore not recommended when the antebrachial catheter diameter does not exceed 18 gauge. In subgroup B2, the injection protocol was monophasic at a lower rate (4.0 mL/s) with more concentrated contrast medium (400 mg I/mL) to increase the iodine administration rate without increasing the injection rate, as suggested by Fleischmann [13]. Enhancement measurements in the aortic arch, the parenchyma of the liver and spleen, and the renal cortex did not differ significantly (p > 0.05) for the three protocols (Figs. 2A, 2B, and 2C).
|
|
|
|
|
|
=
0.66-0.77). However, such results should be evaluated with caution because
image quality assessment is a subjective process. Statistical analysis of our
data resulted in two significant findings. First, image quality at the levels
of the mediastinum, liver, spleen, and kidneys was significantly lower among
patients whose arms were positioned beside their bodies than among those whose
arms were over their heads (p < 0.05), a difference that was true
across all three scan protocols. Second, no globally significant differences
in image quality were identified between the conventional and single-pass
protocols (p > 0.05). At the mediastinal level, image quality was
not related to the degree of contrast enhancement within the aortic lumen.
Indeed, the mean image quality score was higher in groups B1 and B2 than in
group A, whereas the density measured within the aortic lumen was comparable
between groups A and B1 and slightly lower in group B2. This finding is easily
explained by the delayed scan times for the chest in groups B1 and B2 (in
which scanning began at the head), leading to better opacification of the
nonvascular mediastinal structures (e.g., lymph nodes) than in group A. The
difference in mediastinal image quality was significant between groups A and
B2 in the presence of arm artifacts (arms along the body). Even though the total iodine content was close in the two protocols, the lower density within the aortic arch in group B2 was the result of a smaller volume of contrast medium with reduced injection times, resulting in an early first-pass enhancement peak in relation to the scan delay [13]. This phenomenon was not observed in group B1 because biphasic injection leads to closer to uniform enhancement over time [14]. Fanucci et al. [11] reported that image quality was slightly higher for a single-pass protocol compared with a conventional (segmental) protocol, both performed with arms along the body. However, the report of that study gave no information about the injection parameters used in each protocol or about the anatomic regions analyzed for quality.
In our series, image quality results on the cervical spine showed no significant difference in arm position subgroups for each scan protocol (p > 0.05). These results suggest that local artifacts resulting from arm elevation do not hinder optimal analysis of bony structures in the cervical spine in single-pass protocols. This finding can be explained by adaptation of the field of view on the reconstruction images targeting the neck area and by the fact that the shoulders usually are free from lines, catheters, and metallic devices.
Single-pass protocols produce contrast-enhanced images of the head and neck, which is not the case of the conventional protocol. No additional supraaortic CT angiography (CTA) was performed on the conventional protocol group. Therefore, we did not compare image quality for enhancement of the circle of Willis and the supraaortic vessel as we did in other body regions. Image quality with our single-pass protocols is certainly not comparable with that of classic CTA owing to venous enhancement contamination and arm elevation beam hardening. Although our single-pass protocol was not designed to produce CTA quality, our impression was that arterial vessel enhancement was sufficiently high to exclude carotid and vertebral traumatic dissection and for assessment for the presence of intracerebral blood flow (Figs. 3A, 3B, and 3C). This examination is not possible with a conventional protocol unless an additional CTA study is performed with additional contrast material, radiation, and examination time. Further studies are needed to compare the sensitivity and specificity of MDCT performed with a single-pass protocol with those of classic CTA in the detection of arterial dissection.
Some imaging specialists [15, 16] consider arm elevation unacceptable because it is time-consuming and can cause iatrogenic injuries to the shoulder and brachial plexus. Our study results do not substantiate such concerns. In our study, arm positioning did not influence acquisition time (p > 0.05). However, in the group of patients with injury severity scores of 25 or greater, the additional time required for arm elevation was 54.6-72 seconds. Although the difference was not significant (p > 0.05) in this study, the small number of patients with this high injury severity score may not have provided adequate data. The additional time to image these patients can be explained by the fact that in our protocols, both conventional and single-pass, the CT technicians take time for slice positioning and field-of-view adaptation immediately after acquiring cerebral unenhanced images and before starting the contrast-enhanced series. A second technician or the anesthesiologist elevates the patient's arms during these manipulations, which requires only minor additional time in examinations of the most severely injured patients. In our study, elevation of patients' arms over their heads induced no iatrogenic lesions of the shoulder or brachial plexus. Our data also indicate that the risk of subclavian vein compression with arm elevation does not influence the quality of the examination. Mean density measured within the aortic arch and solid intraabdominal organs was not reduced in patients with arms positioned above their heads. Along with authors from other level 1 trauma centers [17, 18], we advocate routine positioning of the arms above the head in CT of trauma patients if upper limb and scapular girdle fractures are not suspected.
Although there was no significant difference in the number of images produced between the conventional and single-pass protocols, the image production rate for the chest and abdomen was almost twofold higher with the conventional scan protocol than with the single-pass protocol. The difference in the image production rate between the two protocols is explained by differences in the initial collimation (1.5 vs 0.75 mm) and pitch (1.35 vs 0.9). Thus in the single-pass protocol, reconstruction of 3-mm-thick images with a 3-mm gap from data obtained with 0.75-mm detectors and a smaller pitch requires additional computerization time compared with the conventional protocol. A 3.3-minute delay for obtaining all 3-mm-thick images with a 3-mm gap in the single-pass protocol can be considered long. However, it is comparable with other reported time delays [5, 7] and is mainly software and hardware dependent. Therefore, image reconstruction times will decrease as technology advances.
Our study design did not take into consideration the radiation dose delivered with the various protocols, which is a major limitation. However, studies [11, 19] have shown marked dose reduction (17%) with use of single-pass protocols instead of conventional segmental protocols. The reduction probably is explained by a decrease in redundant imaging.
We conclude that the single-pass whole-body 16-MDCT protocols evaluated resulted in significantly shorter scanning time than did the segmental conventional protocol. These reductions were achieved by avoiding image reconstruction delays between segmental scans, reducing the need for multiple attempts to reposition the patient, and reducing the number of scout images and time needed for scan programming. Use of the single-pass protocols improved the diagnostic workflow of evaluation of patients with polytrauma, maintained high diagnostic quality, and allowed greater flexibility in reformatting options. Monophasic injection with more concentrated contrast medium (400 mg I/mL) is preferred to biphasic injection because it allows a slower injection rate with maintenance of enhancement quality. On the basis of our observations, we recommend that single-pass protocols with monophasic injection of highly concentrated contrast medium replace segmental conventional CT protocols with 16-MDCT or greater scanners for initial assessment of patients with polytrauma.
Acknowledgments
We thank Nancy Knight, Department of Radiology, University of Maryland
School of Medicine, Baltimore, for invaluable help in editing the manuscript
and Stéphane Montandon, Philips Healthcare, for technical advice and
for performing the phantom tests.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  F. M. Hall Single-Pass Continuous Whole-Body CT for Polytrauma Am. J. Roentgenol., August 1, 2009; 193(2): 594 - 594. [Full Text] [PDF]
|
|
|
|
|
|
P.-A. Poletti and S. E. Mirvis Reply Am. J. Roentgenol., August 1, 2009; 193(2): 594 - 594. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
