| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology and Nuclear Medicine, Campus Benjamin Franklin,
Charité Universitätsmedizin Berlin, Freie Universität Berlin,
and Humboldt-Universität zu Berlin, Hindenburgdamm 30, 12200 Berlin,
Germany.
2 Bracco Imaging SpA, Milan, Italy.
3 Department of Surgery, Campus Benjamin Franklin, Charité
Universitätsmedizin Berlin, Freie Universität Berlin, and
Humboldt-Universität zu Berlin, Berlin, Germany.
Received January 27, 2007;
accepted after revision April 19, 2007.
Address correspondence to T. Albrecht
(thomas.albrecht{at}charite.de).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. CTA and DSA were compared in 50 patients. CTA was independently evaluated by two blinded observers. DSA was evaluated by two additional blinded observers in consensus. Consensus DSA served as the reference standard for comparisons with CTA in terms of diagnostic quality, grading of stenoocclusive lesions, visualization of collaterals, impact on patient management, and time required for analysis.
RESULTS. No significant differences in diagnostic quality were observed between CTA and DSA above the ankle; both CTA observers noted significantly better visualization of pedal arteries (70 and 72 segments, respectively) than on DSA (57 segments). Of 958 stenoocclusive lesions on DSA, CTA observers 1 and 2 detected 933 and 929 lesions, respectively. Sensitivity and specificity for the detection of hemodynamically relevant (> 50%) lesions was 93.3% and 96.5% for observer 1 and 90.1% and 95.6% for observer 2. Collaterals were seen at 150 arterial levels on DSA compared with 97 and 92 levels on CTA (p < 0.05, both observers). Patient management decisions based on CTA were equivalent to those based on DSA in 49 of the 50 patients.
CONCLUSION. CTA is an effective noninvasive alternative to DSA for the evaluation of peripheral arterial disease.
Keywords: aortoiliac arteries • digital subtraction angiography • lower extremity • MDCT angiography • peripheral arterial disease
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The technical limitations of single-detector technology for CTA of the peripheral vasculature have largely been overcome with the introduction of multidetector technology and, in particular, with the introduction of 16-MDCT systems [6]. The considerably more rapid scanning times, improved z-axis resolution, and resulting thinner slice thicknesses achievable with 16-MDCT angiography (16-MDCTA) mean that diagnostic imaging of the aortoiliac and peripheral runoff vasculature to the level of the foot is now a practical procedure at many centers. Additional benefits of 16-MDCTA compared with CTA performed using older CT technologies and with catheter angiography include a lower overall effective radiation dose [6] and, given the rapid scanning times, the possibility of using less contrast agent.
Although 16-MDCT systems are in widespread use, little has yet been reported about the comparative value of 16-MDCTA relative to conventional DSA for diagnostic imaging of the peripheral runoff vasculature of the lower extremities [6]. The purpose of this prospective study was to compare 16-MDCTA and DSA in patients with peripheral arterial disease with regard to diagnostic quality, grading of stenoocclusive lesions, visualization of collaterals, impact on patient management, and time required for analysis.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The study was approved by the ethics committee of our institution and all patients gave informed written consent.
16-MDCTA
CTA was performed using a 16-MDCT scanner (Somatom Sensation 16, Siemens
Medical Solutions). All patients were placed in the supine position with their
feet entering the gantry first. Tourniquets were applied to both thighs for
temporary reduction of venous return. Imaging was performed from 5 cm below
the dome of the diaphragm (suprarenal aorta) to the toes. The scanning
parameters used were as follows: tube voltage, 120 kV; effective tube current,
140 mAs with dose modulation (Care Dose, Siemens); rotation time, 0.5 second;
collimation, 16 x 1.5 mm; and table feed, 40 mm/s. Images were
reconstructed at an effective slice thickness of 2 mm, and a reconstruction
interval of 1.2 mm was used. The mean scanning range along the z-axis
was 126 cm (range, 116–140 cm), which resulted in a mean of 1,051 axial
images (range, 966–1,164). The average scanning duration was 31.5
seconds (range, 29–35 seconds).
The delay between the start of contrast medium administration and the start
of scanning was determined individually for each patient using standard
bolus-tracking software (Care Bolus, Siemens). Scanning began 4 seconds after
a threshold attenuation of 
250 H was reached in the suprarenal aorta.
For each patient, 100 mL of iomeprol (400 mg I/mL [Iomeron 400, Bracco
ALTANA]) followed by 70 mL of normal saline was injected IV via a 20-gauge or
larger IV canula using a single-barrel injector as described previously
[7]. Contrast medium and saline
solution were each injected monophasically at a rate of 4 mL/s.
Image Reconstruction and Review
CTA data sets were reconstructed and analyzed using a workstation
(Leonardo, software version VB30A, Siemens). Maximum-intensity-projection
(MIP) reconstructions were produced by one operator (3 years of experience
with 16-MDCTA) after semiautomatic bone removal. Bone removal was performed
using the 3D function of the workstation: Bones were manually tagged and a
threshold-based region-growing algorithm was applied that automatically
identified the entire bone within the image volume. Axial slices were
carefully reviewed to ensure that no vascular structures (e.g., calcified
plaque or lumen) were included in the bone volume, which was subsequently
subtracted. MIP reconstructions were performed and displayed in eight standard
projections at 22.5° increments rotating around the z-axis. This
was done separately at three levels (aorta and pelvis, upper legs, and lower
legs) to make the best use of the image matrix and thus optimize spatial
resolution. At the level of the feet, a volume-rendering technique without
bone removal was used because bone removal is difficult and time consuming at
that level due to the close contact of arteries and bones. The time required
for 3D reconstruction including bone removal was recorded.
All CTA examinations were reviewed independently by two observers on a workstation. Both observers were blinded to the name, age, and clinical history of the patient and to the results of the DSA examination. Image review was performed initially of MIP and volume-rendering technique reconstructions and axial images together. If this was considered insufficient for determination of the degree of any stenosis by the observer, he or she could perform additional review using curved multiplanar reformations (MPRs), which were calculated along the manually drawn course of the vessel in question in two freely selectable orthogonal planes. Window settings were selected interactively by the observers.
DSA
DSA was performed using an angiography system (Integris 3000, Philips
Medical Systems). The location of the arterial access and the extent of image
coverage depended on the clinical question to be answered. In 21 patients, a
bilateral runoff study after retrograde puncture of the common femoral artery
was performed. In the remaining 29 patients, only one leg was examined
(antegrade common femoral artery puncture before intervention in 23 patients,
retrograde common femoral artery puncture with ipsilateral imaging in two
patients, and retrograde common femoral artery puncture with contralateral
imaging [crossover technique] in four patients). The DSA approach adopted was
influenced by the findings obtained at CTA and was tailored to the clinical
needs of the patient.
For bilateral studies, a 4-French pigtail catheter was placed in the abdominal aorta. Twenty milliliters of iomeprol (300 mg I/mL [Iomeron 300, Bracco ALTANA]) was injected at a rate of 20 mL/s per run. For the lower legs and feet, the volume of contrast medium injected per run could be increased to 40 mL at the discretion of the angiographer. The aorta, upper and lower legs, and proximal feet were imaged sequentially in the posteroanterior projection. Additional oblique or lateral projections of relevant areas were performed if considered necessary. The pelvis was always imaged in two oblique planes. In cases of high-grade stenoses or occlusions, runs using higher volumes of contrast medium were performed at the discretion of the angiographer to improve visualization of collateral and distal vessels.
For unilateral studies, 10 mL of iomeprol (300 mg I/mL) was injected at a rate of 10 mL/s per run via a 5-French sheath (antegrade puncture) or a 4-French dilator or pigtail catheter (retrograde puncture). For the lower legs and feet, the volume of contrast medium injected per run could be increased to 20 mL at the discretion of the angiographer. Images were obtained sequentially in the posteroanterior projection. Again, additional projections of relevant areas or runs with higher volumes of contrast medium were performed in cases of high-grade stenoses or occlusions if considered necessary.
The extent of distal runoff coverage by DSA depended on the individual situation and clinical question to be answered; for example, in cases with proximal occlusion and insufficient collateral flow, the distal leg was not imaged.
DSA images were printed on hard copy and reviewed independently by two blinded radiologists. In the first step, both observers reviewed the images independently and documented their findings on schematic vascular maps. Thereafter, the two observers met to compare their findings using the maps. In cases of discordant findings, the images in question were jointly reviewed and a consensus sought. If a consensus could not be reached, a third observer judged the lesion in question and his decision was used for the consensus interpretation. The consensus findings on DSA were considered the reference standard for subsequent determinations of the diagnostic performance of CTA.
Image and Data Evaluation
The findings of the two blinded observers of CTA images were compared
against the consensus DSA findings in terms of diagnostic quality, grading of
stenoocclusive lesions, visualization of collateral vessels, therapeutic
consequences, and time required for analysis.
Diagnostic quality—Assessment of diagnostic quality was based on arterial levels. Per patient, a maximum of nine levels were evaluated, comprising the aorta, pelvis (left and right), upper leg (left and right), lower leg (left and right), and foot (left and right). However, the number of levels evaluated per patient depended on the coverage achieved on the DSA examination. Thus, fewer than nine levels were evaluated in patients undergoing unilateral studies or in patients with proximal occlusions without sufficient collateral flow.
Each level was scored as either diagnostic (i.e., good differentiation of arteries from background tissue with image quality sufficient for confident evaluation of stenoocclusive lesions) or nondiagnostic (i.e., poor differentiation of arteries from background tissue insufficient for confident evaluation).
Grading of stenoocclusive lesions—Stenoocclusive lesions on CTA images were graded and compared on the basis of arterial segments. Per patient, a maximum of 25 segments could be assessed, again depending on the coverage on DSA. These segments comprised the aorta; the left and right common and external iliac arteries; common, superficial, and deep femoral arteries; popliteal arteries; tibiofibular trunk; fibular arteries; anterior tibial arteries; posterior tibial arteries above the ankle; dorsalis pedis; and posterior tibial arteries below the ankle.
All lesions were drawn onto schematic vascular maps representing the
arteries under evaluation. Lesions were scored primarily by visual inspection
as "diameter stenoses." However, observers were also permitted to
measure the degree of stenosis using a ruler (DSA) or electronic calipers
(CTA) if considered necessary. Segments above the ankle were scored using a
5-point scale in which 0 meant normal vessel lumen with smooth vessel wall; 1,
wall irregularities or mild circumscript stenosis of 
50% of vessel
diameter; 2, moderate stenosis of 51–75% of vessel diameter; 3, severe
stenosis of 76–99% of vessel diameter; and 4, occlusion. Lesions scored
as grade 2 or higher (i.e., > 50% luminal narrowing) were considered
hemodynamically relevant
[6].
Pedal arteries (dorsalis pedis and posterior tibial arteries below the ankle) were assessed only for patency or occlusion; no attempt was made to grade stenoses. Also, the plantar arch was not assessed because it was usually not visualized on CTA.
Comparisons between CTA and DSA were based on the final maps and were performed for each individual lesion. If grade 1 lesions were present within an arterial segment, this was recorded only once for that segment to indicate the presence of arteriosclerotic changes within the segment. If grade 2, 3, or 4 lesions were present within a segment, these were documented individually because each of these lesions might be therapeutically relevant. If a lesion was present on either CTA or DSA but the artery was not visualized on the corresponding technique, the lesion was scored as not being assessable on the corresponding technique.
Visualization of collateral vessels—Collateral vessels were
scored for both CTA and DSA on the basis of arterial levels (as described
earlier) when stenoocclusive lesions of grade 2 or higher were present in the
main arterial segment. A 3-point scale was used in which 0 meant no
collaterals; 1, any number of collaterals of 1 mm in caliber or up to two
parallel collaterals of > 1 mm in caliber; and 2, three or more parallel
collaterals of > 1 mm in caliber.
Therapeutic consequences—The possible impact of the CTA findings on patient management was evaluated jointly by radiologists and vascular surgeons in our daily interdisciplinary conference before DSA. Based on the findings of the CTA examination and the patient's clinical situation, a provisional decision was made and documented as to whether a patient would most likely undergo conservative treatment, intervention, or surgery. If intervention or surgery was considered likely, the arterial segments to be treated (e.g., site of percutaneous transluminal angioplasty [PTA], site of proximal and distal anastomoses of a bypass graft) were identified. This provisional decision was subsequently compared with the therapeutic approach that was initiated after DSA.
Time required for analysis—The time required for image analysis and documentation of the findings on the vascular maps was recorded by the observers per patient for both CTA and DSA.
Statistical Analysis
The diagnostic quality of the CTA and DSA examinations overall and at
various anatomic levels was compared using the chi-square test. The
sensitivity and specificity of CTA for the detection and diagnosis of stenoses
> 50% (grade 2, 3, or 4) among peripheral arteries above the ankle and of
occlusions in the pedal arteries were calculated for CTA observers 1 and 2
separately using DSA as the reference standard. Sensitivities and
specificities of certain subgroups and grading of collateral vessels on CTA
and DSA were compared using the chi-square test.
Interobserver variability between the two CTA observers was calculated
using Cohen kappa statistics for the grading of all stenoocclusive lesions on
CTA (all grades considered), for the differentiation of hemodynamically
relevant stenoses (grades 2, 3, and 4) from nonrelevant lesions (grades 0 and
1), and for the visualization of collateral vessels. Interobserver agreement
was classified as very good (
> 0.8), good ( =
0.61–0.8), moderate ( = 0.41–0.6), fair ( =
0.21–0.4), or poor ( 0.2). Agreement between CTA and DSA for
grading of stenoocclusive lesions was similarly assessed using Cohen kappa
statistics.
The duration of image analysis for CTA and DSA was compared between all observers using analysis of variance and posttesting by Dunn's multiple comparisons test.
All statistical analyses were performed using SPSS software (version 13.0,
SPSS). For all statistical determinations a p value of 0.05 was
considered significant.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Diagnostic Quality
Table 1 provides a detailed
overview of the diagnostic performances of DSA and CTA at the various anatomic
levels and results of statistical analyses. A total of 254 arterial levels
were covered by DSA. Of these 254 arterial levels, 237 (93.3%) were considered
diagnostic on DSA and 17 (6.7%) as nondiagnostic. The nondiagnostic levels
comprised five at the lower legs and 12 at the feet. CTA observer 1 scored 247
(97.2%) of 254 arterial levels as diagnostic and seven (2.8%) as nondiagnostic
(three at the lower legs and four at the feet). CTA observer 2 judged 243
(95.7%) levels to be diagnostic and 11 (4.3%) as nondiagnostic (five at the
lower legs and six at the feet).
|
Grading of stenoocclusive lesions above the ankle—A total of 958 lesions were detected on DSA. These were considered the reference number of lesions.
CTA observers 1 and 2 considered a total of 933 and 929 individual lesions, respectively, to be assessable on both CTA and DSA. In the case of observer 1, 15 lesions were considered to be assessable only on CTA and 25, only on DSA. The corresponding numbers of lesions for observer 2 were 14 and 29, respectively. Because these lesions were not assessable on both techniques, they were excluded from subsequent analyses.
Complete agreement between CTA and DSA was recorded for 781 (83.7%) of 933 lesions evaluated by observer 1 and for 719 (77.4%) of 929 lesions evaluated by observer 2 (Table 2). The 152 lesions for which agreement was not observed by observer 1 comprised 86 (9.2% overall) that were underestimated on CTA (85 lesions by one grade and one lesion by two grades) compared with DSA and 66 (7.1% overall) that were overestimated on CTA (65 lesions by one grade and one lesion by two grades). In the case of observer 2, 128 (13.8% overall) of 210 discordant lesions were underestimated on CTA compared with DSA (126 lesions by one grade and two lesions by two grades), whereas 82 (8.8% overall) lesions were overestimated (all by just one grade).
|
CTA observer 1 considered 313 (33.5%) of 933 lesions hemodynamically relevant, whereas 620 (66.5%) lesions were hemodynamically nonrelevant. A total of 22 (3.5%) of the 620 hemodynamically nonrelevant lesions on DSA were overestimated as relevant on CTA, whereas 21 (6.7%) of the 313 relevant lesions on DSA were underestimated as hemodynamically nonrelevant on CTA (Table 3). For observer 2, 312 (33.6%) of 929 lesions were considered hemodynamically relevant and 617 (66.4%) lesions were hemodynamically nonrelevant. Of the 312 relevant lesions on DSA, observer 2 underestimated 31 (9.9%) as hemodynamically nonrelevant on CTA. Conversely, 27 (4.4%) of the 617 hemodynamically nonrelevant lesions on DSA were overestimated as relevant on CTA by observer 2 (Table 3).
|
The overall agreement between CTA and DSA was good to very good both for
the evaluation of all lesions (observers 1 and 2, = 0.75 and =
0.64, respectively) and for the differentiation of hemodynamically relevant
(grades 2, 3, and 4) stenoses from nonrelevant (grades 0 and 1) lesions
(observers 1 and 2, = 0.90 and = 0.86, respectively).
The sensitivity and specificity for the detection of hemodynamically
relevant stenoses (
50% stenosis) were 93.3% (292/313) and 96.5%
(598/620), respectively, for observer 1 and 90.1% (281/312) and 95.6%
(590/617), respectively, for observer 2
(Table 4). For diagnosing
stenoses > 75%, sensitivity and specificity were 86.1% (174/202) and 97.9%
(716/731) for observer 1 and 77.2% (156/202) and 98.2% (714/727) for observer
2. For occlusions, sensitivity and specificity were 92.3% (108/117) and 99.5%
(812/816), for observer 1 and 88% (103/117) and 99.5% (808/812) for observer
2, respectively.
|
The overall agreement between observers 1 and 2 was good ( = 0.77)
for the grading of stenoocclusive lesions on CTA and very good ( =
0.91) for the differentiation of hemodynamically relevant lesions from
nonrelevant lesions.
Figures 1A, 1B, 2A, 2B, 3A, 3B, 4A, 4B, 4C, and 4D show examples of the concordance achieved with CTA compared with DSA.
|
|
|
|
|
|
|
|
|
|
After exclusion of nondiagnostic segments from both DSA and CTA, a total of
51 and 49 pedal artery segments (observers 1 and 2, respectively) remained for
determination of vessel patency. Of the 51 segments evaluated by observer 1,
13 were considered to be occluded on DSA. The sensitivity and specificity of
CTA for the detection of these occlusions was 92% (12/13) and 95% (36/38),
respectively. Of the 49 segments evaluated by observer 2, 12 were considered
to be occluded on DSA. The corresponding sensitivity and specificity of CTA
for the detection of these occlusions was 92% (11/12) and 97% (36/37),
respectively. The agreement between observers 1 and 2 for the assessment of
pedal artery patency on CTA was very good ( =0.96).
Visualization of collateral vessels—Collateral vessels were
observed at 150 arterial levels on DSA
(Fig. 5). These collateral
vessels were scored as grade 1 in 84 cases and as grade 2 in 66 cases. On CTA,
collateral vessels were detected at 97 arterial levels (grade 1, n =
75; grade 2, n = 22) by observer 1 and at 92 levels (grade 1,
n = 62; grade 2, n = 30) by observer 2. The difference
between the findings on DSA and CTA was statistically significant (p
< 0.001) for both observers. Although the agreement between CTA and DSA
with regard to grading of collateral vessels was only fair for both observers
( = 0.33 and = 0.29, observers 1 and 2, respectively), the
agreement between observers was good ( = 0.72) for the visualization of
collateral vessels on CTA.
|
Time Requirements for Image Analysis
The analysis of DSA images, including documentation of findings, required
an average of 11.8 ± 5 minutes for observer 1 and 9.9 ± 4.4
minutes for observer 2. The two observers of CTA images required an average of
18.0 ± 10.9 minutes and 18.7 ± 9.1 minutes, respectively. This
was in addition to the time required for postprocessing (production of MIP
reconstructions with bone removal: mean, 13.6 ± 4.2 minutes; range,
6–22 minutes). The differences between DSA and CTA were statistically
significant for all observers (p < 0.05). CTA observer 1 chose to
perform additional MPR s in 116 of 933 identified stenoocclusive lesions,
whereas observer 2 did so in 260 of 929 lesions.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
50%) stenoses was 93.3%
and 90.1%, respectively, and the specificity was 96.5% and 95.6%,
respectively. Of note is that analysis of stenoocclusive disease was performed
using a lesion-based approach rather than an arterial segment–based
approach, as has been used in other studies
[6,
8–13].
Despite the potentially greater susceptibility of our approach to errors in
lesion matching, the sensitivity and specificity values obtained bear
excellent comparison with reports of 96% and 97%, respectively, for 16-MDCTA
[6] and 91–96% and
92–99%, respectively, for 4-MDCTA
[9,
10,
12,
13]. Moreover, these values
also bear excellent comparison with values of 92–100% and 91–99%,
respectively, for contrast-enhanced MR angiography
[14–19].
Our study might have yielded even better results if the data had been
evaluated on a segmental basis as in previous studies
[6,
9,
10,
12,
13]. However, because the
approach to patient management (e.g., PTA or bypass surgery) may differ
depending on whether one lesion or several lesions are detected in a given
arterial segment, a lesion-based approach to image evaluation was considered
more appropriate for this study rather than a segment-based approach.
Disagreements with DSA were almost invariably underestimation or overestimation by just one grade. Rather than reflecting differences between the two imaging techniques, these disagreements may in large part reflect the subjective semiquantitative nature of the image analysis, which is particularly relevant in the case of lesions at or near a given cutoff level (e.g., 50%). However, it should also be borne in mind that whereas CTA is a 3D technique permitting image reconstruction in multiple planes, DSA is solely a 2D technique in which images are obtained in just two or three planes at best. Recently, a study of the visualization of carotid artery stenosis on contrast-enhanced MR angiography revealed the 3D MR angiography technique to be superior relative to 2D DSA for grading of stenosis when compared with 3D rotational DSA as the standard of reference [20]. Although a similar evaluation was beyond the scope of this study due to the absence of rotational DSA data, further investigation of the reasons for the disagreements of CTA with 2D DSA may be warranted.
Visualization of collateral vessels was not as good on CTA as on DSA. Although, to our knowledge, no previous peripheral CTA study has quantitatively evaluated the visualization of collaterals, it has frequently been stated that CTA is superior to DSA in depicting peripheral collateral circulation and in revealing enhancement distal to arterial occlusions because of collateral circulation [6, 12, 13]. The fact that these assertions may not necessarily be true, however, is evident not only from the absence of published data but also from published comparative images that often show more extensive collateral vessels on DSA than on CTA [6, 12]. One explanation for the poorer visualization of collateral circulation on CTA in this study may be that the caliber of these vessels is typically very small, often reaching the limits of spatial resolution on CTA. Also bear in mind concerning the visualization of collateral vessels on CTA that unlike the normal vessels of the leg, which generally run perpendicularly through the xy-plane causing minimal partial volume effects, collateral vessels typically have a more convoluted course, frequently running obliquely or even horizontally to the imaging plane. As a result, these vessels are more prone to partial volume effects and thus reduced contrast resolution. This may have been exacerbated by the effective slice thickness of 2 mm used in this study; thinner slices should reduce partial volume effects and may improve collateral visualization. Regardless of the visualization of collateral circulation, visualization of refilling native arteries distal to occlusions was better on CTA than on DSA in our study (Figs. 3A, 3B, 4A, 4B, 4C, and 4D), which is consistent with previous reports [6, 12, 13].
In terms of patient management decisions, our results revealed equivalence of CTA and DSA in 49 of the 50 patients. This finding further supports the use of CTA as a noninvasive alternative to DSA for diagnostic imaging of the peripheral vasculature.
The mean time for image analysis was longer for the two CTA observers than for the DSA observers. Although the additional time necessary for postprocessing and evaluation of CTA images is offset by savings in table and examination time, the large amount of data generated on CTA is an established problem of the technique.
Until the introduction of MDCT scanners with four detector rows, imaging of the peripheral vascular tree was limited to not more than 40 cm of craniocaudal coverage after a single IV injection of iodine-based contrast medium [3–5, 21–23]. However, the relatively slow speed of acquisition and poor z-axis resolution of single-detector CT scanners precluded full coverage of the peripheral vascular tree from the renal arteries to the toes. The advent of 4-MDCT technology improved the utility of CTA for evaluation of peripheral vascular disease [8–13]. However, typical scanning times ranged from 45 to 65 seconds whereas the effective slice thickness ranged between 3 and 5 mm. Moreover, the volume of contrast medium administered, which is critically dependent on scanning duration, was often as much as 180 mL [8, 11].
The availability of MDCT scanners with 16 detector rows has overcome many of these limitations; in a recent study by Willmann et al. [6], the nominal and reconstructed slice thickness was 0.75 mm with a reconstruction interval of 0.4 mm, and the volume of contrast medium administered was only 100 mL. Although the image acquisition time was not reported, the time from patient entry to the CT suite until scanning had finished was only 12 minutes. Although the effective slice thickness and reconstruction interval were slightly larger in our study than in the study by Willmann et al., the resulting mean number of axial images acquired was only 1,051 (range, 966–1,164) compared with 5,700 (range, 5,160–6,225), resulting in a lower requirement for computing power and archiving space and presumably more rapid image postprocessing and analysis. Moreover, the larger slice thickness permitted imaging down to the level of the feet rather than to the level of the tibial arteries. Finally, the acquisition of 2-mm slices resulted in reduced image noise and less degraded image quality, allowing a lower tube current (140 mAs compared with 210 mAs in the study by Willmann et al.) and thus substantially reduced radiation exposure to the patient.
Important considerations for CTA of the lower limbs are the acquisition and
contrast medium injection protocols. These should ensure sufficient arterial
enhancement throughout the arterial tree including the pedal arteries without
overriding the contrast medium bolus and with minimal venous enhancement. To
obtain a sufficiently diagnostic scan, the selection of contrast injection
duration, scanning delay, and table feed is critical. This is particularly
challenging in the peripheral vasculature because arterial flow velocities may
vary considerably among patients with peripheral arterial disease
[24–26].
The approach we used in our study was a standardized protocol in which the
contrast medium injection duration (25 seconds) was 6.5 seconds shorter than
the mean scan acquisition duration of 31.5 seconds. We used automated bolus
tracking with a high threshold of 250 H in the aorta followed by a
delay of 4 seconds after reaching the threshold to ensure scan acquisition at
high arterial contrast concentration. It has previously been shown that this
protocol provides good arterial enhancement of the entire peripheral arterial
tree with a mean enhancement of > 160 H at the lower legs and feet
[27]. In the present study,
CTA observers 1 and 2 considered the diagnostic quality of images to be good
for 97.2% and 95.7% of arterial levels, respectively.
Concerning the iodine concentration of the contrast medium, investigators have previously chosen contrast medium with iodine concentrations of 300 [8, 10, 11], 320 [4, 9], or 300–370 mg I/mL [13]. In common with Catalano et al. [12], we chose a contrast medium with a high iodine concentration of 400 mg/mL because high iodine concentrations result in higher iodine delivery rates at a given injection rate, resulting in stronger arterial enhancement and thus better visualization of small peripheral arteries [28]. Because injection rates cannot be increased indefinitely, especially in patients with poor peripheral venous access, we consider high-concentration contrast medium advantageous for peripheral CTA. This is particularly the case for CT scanners with 16 or more detector rows because the increased speed of acquisition and coverage demands high contrast density at all levels of the peripheral vasculature. When contrast media that have relatively low concentrations of iodine are used, the speed of scan acquisition may outstrip the ability of the contrast medium to provide sufficient contrast density at all points of the scan in sufficiently rapid time.
Our study has limitations. First, DSA examinations were not standardized but were performed according to the clinical protocol in place at our institution for the relevant question to be answered. Thus, the DSA examinations performed varied in terms of the extent of coverage and comprised both bilateral and selective unilateral studies. A consequence of tailoring the DSA examination to the pertinent clinical question was most clearly seen in the foot where just 57 of 80 arterial segments contained within the field of view were considered of adequate diagnostic quality. Selective DSA of the foot was considered necessary only if below-knee vascular reconstruction was a likely therapeutic approach. On the other hand, suboptimal visualization of the pedal arteries was accepted in patients with proximal stenoses and good lower leg runoff to avoid the risks associated with additional contrast medium load and further angiographic manipulation [29]. Nevertheless, when both imaging techniques were considered, diagnostic, very good values for sensitivity and specificity were obtained for the assessment of pedal artery patency on CTA.
A second limitation that is also a limitation of CTA in general is that evaluation of heavily calcified arteries is sometimes difficult (e.g., in patients with long-term diabetes or terminal renal failure) because of partial volume blooming artifacts that may cause overestimation of stenoses [10, 11, 13, 21]. Because we did not systematically investigate the extent to which calcification impacts diagnostic performance, future work is warranted to clarify this issue. A further limitation of CTA is its difficulty in depicting the plantar arch. This shortcoming may be a particular problem in patients with extensive peripheral vascular disease requiring below-knee arterial reconstruction.
Both of these limitations can be overcome with the recent availability of 64-detector and dual-source and dual-energy CTA systems. In our clinical experience, visualization of the plantar arch is improved by 64-detector row systems because of their better spatial resolution. Dual-energy CT, although still in its infancy, provides automatic bone removal based on the inherent image information without operator-dependent postprocessing, which should render CTA more time efficient. Furthermore, dual-energy CT may allow automatic calcium removal from plaques, thereby potentially permitting better grading of heavily calcified stenosis than currently achievable.
In conclusion, our study shows that 16-MDCTA is a highly sensitive and specific alternative to conventional DSA for the detection of hemodynamically significant stenoses of the aortoiliac and lower extremity arteries. Specifically, no significant difference between CTA and DSA was observed by either observer in terms of diagnostic quality of leg artery segments, but fewer nondiagnostic pedal artery segments were noted on CTA. Although some differences between CTA and DSA were noted for the visualization of collateral vessels, these differences did not affect patient management decisions, which were the same for both techniques in 49 of the 50 patients. The comparable diagnostic performance of CTA to DSA combined with its lower invasiveness, lower risk, and greater patient convenience are factors favoring the use of CTA in routine practice.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  T. A. S. Matalon Invited Commentary RadioGraphics, March 1, 2008; 28(2): 549 - 549. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
