| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Chang Gung Memorial Hospital at Kaohsiung, Chang Gung
University, 123 Ta-Pei Rd., Niao-Sung Hsiang, Kaohsiung Hsien 833,
Taiwan.
2 Department of Thoracic and Vascular Surgery, Chang Gung Memorial Hospital at
Kaohsiung, Chang Gung University, Kaohsiung Hsien, Taiwan.
3 Department of Public Health and Biostatistics, Chang Gung Memorial Hospital at
Kaohsiung, Chang Gung University, Kaohsiung Hsien, Taiwan.
4 Department of General Surgery, Chang Gung Memorial Hospital at Kaohsiung,
Chang Gung University, Kaohsiung Hsien, Taiwan.
5 Division of Nephrology, Department of Internal Medicine, Chang Gung Memorial
Hospital at Kaohsiung, Chang Gung University, Kaohsiung Hsien, Taiwan.
Received October 3, 2004;
accepted after revision November 29, 2004.
S.-F. Ko was supported by grant NSC 92-2314-B-182A-075 from the National
Science Council, Taiwan.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. MDCT angiography of the complete vascular trees of 36 failing AVFs or AVF-related complications (20 native and 16 polytetrafluoroethylene graft AVFs) was reviewed. The numbers and degrees of stenoses at the anastomoses, graft loops, and draining and central veins and the presence of aneurysms or thrombosis were recorded. Wilcoxon's signed rank test was used to compare the findings of MDCT angiography with those of digital subtraction angiography (DSA) (n = 10), surgery (n = 22), or both (n = 4) performed within 2–6 days. Kappa statistics were used to correlate the clinical feasibility of MDCT angiography assessed by two reviewers.
RESULTS. Among the 14 AVFs examined with both MDCT angiography and
DSA, no significant difference was seen in the detection and grading
(p = 0.317 to > 0.999) of stenoses at various segments of the
entire vascular tree. Among the 36 AVFs examined, MDCT angiography also showed
no significant difference from DSA or surgery in revealing vascular stenoses,
aneurysms, and thromboses from the supplying artery to central veins
(p = 0.317 to > 0.999). Overall, the sensitivity, specificity,
positive and negative predictive values, and accuracy of MDCT angiography in
lesion detection were 98.7%, 97.5%, 98.8%, 97.2%, and 98.3%, respectively.
High image quality with superb interobserver correlation (
= 0.809 to
> 0.999) validated the clinical feasibility of MDCT angiography for
assessing AVFs.
CONCLUSION. MDCT angiography is clinically feasible for evaluating the complete vascular tree of failing AVFs and in showing uncommon complications, including brachial aneurysms and central vein lesions.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Among the 43 patients who underwent MDCT angiography, 14 also underwent DSA within 3–5 days after MDCT angiography, and 10 of these 14 patients were subsequently treated with endovascular intervention and four underwent subsequent surgery. Twenty-two patients received surgical treatment within 2–6 days solely based on MDCT angiography findings. Three patients declined any further evaluation or treatment of the failing AVF after MDCT angiography and were subsequently lost to follow-up; four patients underwent creation of a new AVF on the contralateral side, and the original failed AVF was abandoned after MDCT angiography. These seven patients were excluded from this study because of a lack of DSA or surgical findings for comparison. Finally, a total of 36 patients (20 women and 16 men; age range, 22–78 years; mean, 63.7 ± 5.4 [SD] years) with successful MDCT angiography of AVFs (20 in the right and 16 in the left upper extremities) and with DSA or detailed surgical findings were included in the correlation analysis.
MDCT Angiography
MDCT angiography was performed with a 4-MDCT scanner (LightSpeed Plus, GE
Healthcare). Venous access was achieved through a 20-gauge IV catheter
inserted into a peripheral vein of the upper extremity contralateral to the
AVF side in 33 patients. In this way, the streak artifact caused by contrast
material injection was usually limited to the contralateral axillary or
subclavian veins, and the superior mediastinum and the AVF vascular tree could
satisfactorily be evaluated. Three patients had no suitable venous access in
the upper extremity and a contralateral lower extremity peripheral vein was
used.
The contrast injection rate (2.5 mL/sec) was based on a preliminary trial in five patients in whom normal saline was injected at different injection rates (2, 2.5, 3, 3.5, and 4 mL/sec). Only one of these five patients could endure an injection rate of 4 mL/sec, and three had venous rupture at 3 mL/sec and one at 4 mL/sec. However, all five patients could withstand an injection rate of 2.5 mL/sec. Before scanning, a test bolus of normal saline was injected via a mechanical power injector at a rate of 2.5 mL/sec to ensure no leakage from or rupture of the vessel, and then 100 mL of contrast material (Omnipaque [iohexol], 350 mg I/mL, Amersham Health) was administered at the same rate. Bolus tracking (SmartPrep, GE Healthcare) was applied by selecting an engorged vessel proximal to the AVF with a threshold level of 120 H.
The patients were scanned in either the supine or the prone position, depending on comfort and tolerance, in a direction from distal to central and with the arm extended above the head or placed alongside the body. The scanning protocol included a pitch of 3 (2.5/7.5 mm, high quality), 0.5-sec scanner rotation, 120 kV, 160 mAs, and a 35-cm field of view. MDCT angiography scanned the complete vascular tree to the level of the superior vena cava, including the cavoatrial junction (mean coverage, 72 cm; range, 62–76 cm, depending on the AVF type and volume coverage). Patients' reactions to the examination were documented. Mild quivering of the affected upper extremities was common, but neither allergic reaction, extravasation of contrast material, nor any other discomfort was reported during the examinations.
After the examination, the raw data were immediately reconstructed on the working console at 2.5-mm thickness and a 1.25-mm interval (50% overlap). The reconstructed image data were then transferred to a commercially available workstation (Advantage Workstation, AW 4.0, GE Healthcare) in a 1024 x 1024 pixel format. In addition to axial images, postprocessing of MDCT angiography was performed with interactive real-time multiplanar or curved reconstructions, and 2D and 3D projection angiograms were created with maximum-intensity-projection (MIP) and shaded-surface display (SSD) volume-rendering techniques to best depict particular anatomic features of interest. MIP images were used to calculate the percentage of stenosis. Because data segmentation methods were used, unnecessary overlapping structures and vascular clips, if present, were removed, so that the entire vascular tree, including central veins, was clearly shown. Two radiologists who were blinded to the clinical status of the patients were asked to independently reformat and analyze these images. Disagreements were resolved by consensus. The time devoted to each case was approximately 30 min.
Digital Subtraction Angiography
DSA examinations were performed by experienced radiologists on a digital
subtraction system (Integris V5000, Philips Medical Systems). Nine patients
underwent DSA according to the method of Staple
[10] by direct puncture of the
draining vein with a 20-gauge needle while the arterial part and graft loops
were retrogradely opacified after the application of a proximal cuff to
interrupt flow. Because of a high location of the PTFE graft, suspected
infection of the AVF, or marked swelling of the upper extremity, five patients
underwent axillary arterial DSA using a 4-French end-hole catheter via a right
transfemoral approach. The total amount of contrast material (Omnipaque) used
varied from 50 to 80 mL, and the number of contrast injections varied from
five to nine, depending on the complexity of the vascular structures and the
number of angled views.
Image Analysis
The number and degree of AVF stenoses revealed on MDCT angiography and DSA
were compared for patients who underwent these two examinations. Measurements
of the stenoses at arteriovenous anastomoses (including artery-to-graft and
graft-to-vein anastomoses), graft loops, and draining veins were made
independently by the CT radiologists and interventional radiologists using
electronic calipers on the magnified MDCT angiography and DSA images. The
percentage of stenosis was calculated as
 |
Using DSA or surgery as the reference standard, the entire vascular tree of the AVF was divided into four parts (or five parts for PTFE grafts) for analysis, including the supplying artery, anastomosis sites, graft loop (if present), draining veins, and central veins. The number of abnormalities, including number of stenoses, aneurysms, thrombosed vessels or grafts, and central vein lesions revealed on MDCT angiography in each patient, was compared with the findings on DSA or surgery using the nonparametric paired Wilcoxon's signed rank test. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of MDCT angiography were calculated for each of these findings.
The assessment of the overall quality of MDCT angiography (including the presence or absence of motion artifacts, the continuity of the vessels, and the sharpness and demarcation of the vascular outline) and clinical value were subjectively graded using a 4-point scale as follows: 1 = poor quality and inadequate information for answering clinical queries and making decisions about further treatment; 2 = fair quality but inadequate information for answering clinical queries and making decisions about further treatment; 3 = good quality and adequate information for answering clinical queries and making decisions about further treatment; and 4 = excellent quality and adequate information for answering clinical queries and making decisions about further treatment. The MDCT angiography reformatted images were preoperatively rated by a vascular surgeon and then independently rated by a radiologist who had not performed the image processing. The interobserver correlation of the quality scores was evaluated with correlation coefficient kappa statistics.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Detection and Grading of Stenoses
The number and grading of stenoses of the AVF shown on MDCT angiography and
DSA in 14 patients are summarized in Table
1. The total numbers of stenoses detected on MDCT angiography
versus DSA at anastomoses (13 vs 13), graft loops (7 vs 7), and draining veins
(18 vs 17) were not significantly different (p = 0.317 to >
0.999). The stenosis scores acquired from MDCT angiography versus DSA at
anastomoses (1–3/1.79 vs 1–3/1.71 [range/mean score]), graft loops
(1–3/1.00 vs 1–3/0.93), and draining veins (1–3/2.64 vs
1–3/2.64) were also not significantly different (p =
0.317–0.655) (Figs. 1A,
1B, and
1C). With DSA used as the
standard, MDCT angiography showed only one grade 1 false-positive stenosis at
the draining vein among these 14 patients.
|
|
|
|
Lesions in the Entire Vascular Tree
Table 2 summarizes the
number of lesions (including stenoses, aneurysms, and thromboses) along the
AVF vascular tree revealed on MDCT angiography versus DSA or surgery. Among
the 36 AVFs examined, MDCT angiography showed no significant difference from
DSA or surgery in revealing aneurysms, thromboses, and vascular stenoses
(Figs. 1A,
1B,
1C,
2A,
2B,
3, and
4) from the supplying artery to
the central veins (p = 0.317 to > 0.999). MDCT angiography
exhibited 100% sensitivity, specificity, PPV, NPV, and accuracy in disclosing
lesions in most portions. However, the numbers of stenoses detected on MDCT
angiography versus DSA at the native AVF draining veins and graft loops were
25 versus 26 and 12 versus 13, respectively, with sensitivity, specificity,
PPV, NPV, and accuracy ranging from 80% to 100%. Overall, the sensitivity,
specificity, PPV, NPV, and accuracy of MDCT angiography in lesion detection
were 98.7%, 97.5%, 98.8%, 97.2%, and 98.3%, respectively. In addition, MDCT
angiography clearly depicted a hematoma around the AVF in six cases and the
presence of two pathologically proven mycotic aneurysms with perianeurysmal
inflammatory soft tissue and stranding. MDCT angiography also offered
additional important information (4 central vein stenoses and 4 brachial
arterial aneurysms) that was overlooked on the initial clinical
examination.
|
|
|
|
|
Image Quality and Diagnostic Value
Among 36 patients, only one quality score of 2 was given to the draining
veins of a native AVF on MDCT angiography by a vascular surgeon because
intermittent trembling of the patient resulted in a focal shaggy venous
outline mimicking venous stenosis. Otherwise, every section of AVFs was
considered diagnostic on MDCT angiography with excellent quality scores
(median score = 4) for assessing the supply arteries and AVF anastomosis
sites, and good quality scores (median score = 3) for assessing the graft
loops, draining veins, and central veins. Excellent interobserver correlations
( = 0.809; p > 0.999) were obtained between two reviewers
in the validation of clinical feasibility of MDCT angiography for assessing
various segments of AVF.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This study has shown that MDCT angiography is a reliable and reproducible noninvasive method for assessing failing AVFs. A diameter reduction of 50% or more in an AVF is considered hemodynamically significant, whereas an exact differentiation of flow-limiting grade 2 from grade 3 stenosis, from a practical standpoint, is not absolutely necessary because both conditions require interventional angioplasty. Among 14 patients who underwent both MDCT angiography and DSA that showed a total of 37 stenoses, MDCT angiography underestimated only one stenosis as grade 1 (48% diameter reduction) that turned out to be a significant stenosis (grade 2, 52% diameter reduction) on DSA at the draining vein. MDCT angiography also overestimated only one stenosis as significant (grade 2, 53% diameter reduction) that turned out to be a grade 1 stenosis (48% diameter reduction) on DSA at the graft loop. In addition, on MDCT angiography, only one false-positive stenosis was noted (8% diameter reduction), which looked normal on DSA. Overall, analysis of the 14 patients having both MDCT angiography and DSA studies in this series revealed no significant difference in the detection and grading of stenoses at the anastomosis sites, graft loops, and draining veins of failing hemodialysis AVFs. In addition, the 3D capability of MDCT angiography can offer freely rotated projection angiograms to show vascular lesions from the appropriate perspective. Therefore, MDCT angiography before DSA is helpful in identifying lesions and stratifying patients for a more targeted procedure such as percutaneous angioplasty or surgery.
The advantages of MR angiography in the assessment of hemodialysis AVFs include its lack of radiation and no need for iodinated contrast material, but time-of-flight and phase-contrast MR angiography may be limited by tortuosity and flow turbulence of the vessels, leading to overestimation of stenosis, long examination times, high susceptibility to motion artifacts, and high cost [15–19]. Contrast-enhanced MR angiography has been reported to be well suited for detecting AVF stenoses, but it may still be limited by its spatial resolution and the fact that vascular clips may mimic graft stenosis [21, 30]. Conversely, vascular clips adjacent to the area of interest, if present, could be segmented during postprocessing of MDCT angiography, and no image degradation was noted among our patients.
Two single-detector CT angiography studies have reported that CT angiography has the potential to serve as an alternative imaging technique for impaired AVFs. However, these studies were hindered by a short scanning span adjacent to the targeted area and limited resolution, leading to a zigzag outline of the vascular contour [13, 22]. Vascular disorders in the upper arm, axilla, or central veins may be overlooked. MDCT angiography provides the advantages of speed, improved temporal and spatial resolution, greater anatomic coverage, and higher-quality reconstructions [24–29]. In addition to providing axial CT images, as shown in this study, retrospective data reconstructions facilitate the use of multiplanar and 3D delineation [26–29] of the complete vascular tree, including the supplying artery and central veins. As in our study, central vein stenosis in four of six patients and all four brachial artery aneurysms revealed on MDCT angiography were not recognized on the initial clinical examination. In our study, MDCT angiography had 100% accuracy in diagnosing brachial aneurysms, stenoses at the anastomosis site, venous aneurysms, thromboses of failing AVFs, and central vein stenoses. In the assessment of various lesions affecting different segments of the vascular tree, our study confirmed the excellent diagnostic capability of MDCT angiography, including an overall sensitivity, specificity, PPV, NPV, and accuracy of 98.7%, 97.5%, 98.8%, 97.2%, and 98.3%, respectively.
To obtain optimal vascular enhancement, comprehensive adjustment of the injection flow rate and precise timing of scanning are essential [13, 22, 26–29]. For single-detector CT angiography, a large volume of contrast material and a high flow rate (3–5 mL/sec) are required to maintain high and uniform opacification of vessels [13, 22, 26]. In our study, the contrast injection rate (2.5 mL/sec) was based on a preliminary trial in five patients. Jacobs et al. [31] have reported that no correlation exists between injection rate and extravasation rate; nevertheless, four of five patients suffered from peripheral venous rupture at an injection rate of 3 mL or more per second in the initial trial. This might be because, in contrast to American patients, our patients were Chinese patients with chronic illnesses, a small body size, small peripheral veins, and great vascular fragility. With the help of an automated tracking system of contrast enhancement and a predefined trigger threshold at the appropriately selected region of interest [32, 33], high-quality MDCT angiography images can be obtained. In addition, the total amount of contrast material can be reduced to less than 100 mL compared with the usual dose of more than 200 mL for single-detector CT angiography [26]. MDCT angiography examinations can be accomplished quickly (8–10 min for patient preparation and CT), and the subsequent postprocessing usually takes about 20 min. Excellent interobserver correlation between the radiologist and surgeon was obtained in the evaluation of the imaging quality and diagnostic value of MDCT angiography. Our results show good to excellent image quality scores for all segments of the vascular tree of both native arteriovenous and PTFE graft fistulas, which results in highly reliable evaluations for treatment planning.
MDCT angiography has several shortcomings. First, the use of IV contrast material is associated with a risk of anaphylactic reaction. Second, as with DSA, this method uses ionizing radiation. Third, in our experience, because of the presence of failing AVF or AVF-related complications, patients find it quite difficult to hold their upper extremities absolutely still, and mild quivering may lead to a minimally blurred outline of the reformatted images. Fourth, the spatial resolution of 4-MDCT is still limited; the use of 16 (or more)-MDCT can further improve image quality and shorten the data acquisition time. Fifth, although vascular clips did not cause any image degradation in our patients, they are a possible limitation of MDCT angiography if they are abutting the AVF vascular tree. Finally, MDCT angiography cannot provide hemodynamic details, including the flow velocity and pressure gradient across the stenosis, and sonography of the targeted area is recommended if this information is critical. On the other hand, unlike DSA and MR angiography, MDCT angiography can provide supplementary information, including identification of the thrombosed or occluded portion of the vascular lesion, and can reveal associated conditions, such as the presence of calcification, hemorrhage, or infection.
In conclusion, MDCT angiography may be considered as a noninvasive alternative to DSA for evaluation of the complete vascular tree of failing AVFs. In addition, MDCT angiography is useful in elucidating uncommon AVF-related complications such as supplying artery aneurysms and central vein lesions.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S.-F. Ko, C.-C. Huang, S.-H. Ng, M.-J. Hsieh, C.-C. Lee, Y.-L. Wan, and C.-D. Liang Imaging of the Brachiocephalic Vein Am. J. Roentgenol., September 1, 2008; 191(3): 897 - 907. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
