| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Vascular and Interventional Radiology |
1 From the Department of Radiology, Stanford University School of Medicine, Stanford University Medical Center, Rm S-072B, Stanford, CA 94305-5105. From the 1997 RSNA scientific assembly. Received December 16, 1998; revision requested February 11, 1999; final revision received August 3; accepted August 5. Address reprint requests to G.D.R. (e-mail: grubin@stanford.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: Forty patients underwent conventional angiography and CT angiography following treatment of aortoiliac aneurysms with endoluminal stent-grafts. Six additional sets of conventional angiographic–CT angiographic examinations were performed in five patients after placement of additional stent-grafts or coil embolization to treat perigraft leakage. Three faculty CT radiologists who were blinded to patient clinical data and outcome independently interpreted the CT angiograms, and three faculty angiographers, who were not involved in the stent-graft deployment, interpreted the conventional angiograms. Images were assessed for the presence of postdeployment complications. A reference standard was developed by experienced radiologists using all available images and clinical data. Sensitivities, specificities, and 
values were calculated.
RESULTS: Perigraft leakage was the most commonly identified complication. Twenty perigraft leaks were detected in the results of 46 examinations. Sensitivities and specificities for detecting perigraft leakage were 63% and 77% for conventional angiography and 92% and 90% for CT angiography, respectively. The value was 0.41 for conventional angiography and 0.81 for CT angiography.
CONCLUSION: CT angiography is the preferred method for establishing the presence of perigraft leakage following treatment of aortoiliac aneurysms with stent-grafts.
Index terms: Aneurysm, aortic, 94.73, 981.73 • Aneurysm, CT, 94.12916, 981.12916 • Angiography, comparative studies, 94.1211, 981.1211 • Computed tomography (CT), angiography, 94.12916, 981.12916 • Computed tomography (CT), comparative studies, 94.12916, 981.12916 • Interventional procedures, complications, 94.458, 981.458 • Stents and prostheses, 94.456, 981.456
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Results of preliminary investigations suggest that CT angiography may be accurate for assessing the success of endoluminal aortic aneurysm repair; however, to our knowledge, an assessment of blinded CT and conventional angiographic interpretations by multiple independent reviewers has not been performed (4,5). We therefore sought to rigorously determine the accuracy of conventional angiography and CT angiography for assessing complications following stent-graft deployment by using blinded independent interpretations of CT angiographic and conventional angiographic examinations from a database of patients who previously had undergone repair of aortic aneurysms with stent-grafts.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The remaining 40 patients (31 men, nine women; age range, 40–89 years; mean age range ± SD, 70.6 years ± 9.5) had both CT angiographic and conventional angiographic examination results available for review and were thus identified for inclusion into this study. Thirty-six patients had thoracic aortic aneurysms and four had abdominal aortic or iliac arterial aneurysms. Because of placement of additional stent-grafts or coil embolization to treat perigraft leakage in some patients after primary stent-graft repair, five patients had one additional set of conventional angiograms and CT angiograms available for comparison, and one patient had two additional sets of conventional angiograms and CT angiograms available for comparison. Therefore, results from 46 (40 thoracic and six abdominal aortoiliac) conventional angiographic and CT angiographic examinations were available for review.
Thirty-one pairs of conventional angiographic–CT angiographic examinations were completed within 24 hours of each other; eight, in 1–3 days; three, in 4–6 days; and four, in 7 days or more. CT angiography was performed before conventional angiography in 19 cases, and conventional angiography was performed before CT angiography in 27 cases. All imaging studies were performed as part of the patient's clinical care; therefore, specific institutional review board approval was not required.
All stent-grafts used to treat the patients in this study were made in-house and were composed of a stainless steel two-stent endoskeleton with woven polyester graft material sutured to its exterior at each end by interrupted sutures of 5-0 polypropylene. Between the ends, the graft material was not secured to the stent.
CT Angiographic Technique
CT scans were obtained with either a Somatom Plus S (Siemens Medical Systems, Iselin, NJ) or a HiSpeed Advantage (GE Medical Systems, Milwaukee, Wis) CT scanner. The imaging protocol was the same for each scanner. An anteroposterior localizing projectional radiograph was obtained from which a low-resolution localizing CT scan was prescribed. Sections were acquired with 10-mm collimation, 80 kV, 80 mA, a pitch of 2.0, 180° linear interpolation, and 10-mm reconstruction intervals from 5 cm above the stent-graft to 5 cm below the stent-graft. From the resultant transverse sections, CT angiography was prescribed to include the entirety of the stent-graft plus any major aortic branches within 5 cm of the stent-graft.
A 20-gauge intravenous catheter was placed into an antecubital vein, and the time required for a 15–20-mL bolus of nonionic iodinated contrast medium (Omnipaque 300, Nycomed Amersham, Princeton, NJ, or Isovue 300, Bracco, New Brunswick, NJ) to travel from the injection site to the aorta was determined. Eight seconds after the contrast medium was injected, 5-mm collimated CT sections were acquired every 2 seconds at the anticipated initiation site of the CT angiography. A time-attenuation curve was generated from a circular region of interest placed on the aorta on each of the sections. The peak of the curve was selected as the delay time for the CT angiography.
CT angiography was performed with 120–160 mL of contrast medium injected at a rate of 4–5 mL/sec by using 3-mm collimation, a pitch of 2.0, 120 kV, 180–280 mA, 180° linear interpolation, and 2-mm reconstruction intervals. Although multiplanar reformations, maximum intensity projections, and shaded surface displays were created in the majority of cases, they were not used in this study.
Conventional Angiographic Technique
Screen-film or digital subtraction angiography was performed in a dedicated angiography suite; aortograms obtained in the operating room by using portable equipment were excluded from this study. A standard Seldinger technique was used to access the femoral artery with an 18-gauge needle through which a guide wire was passed into the aorta. A 5-F angiographic pigtail catheter was then passed over the guide wire, and aortography was performed. In some cases, if a leak was suspected at initial aortic injection or from a prior CT scan, then selective injections were performed with occasional transcatheter deployment of coils or additional stent-graft bodies.
Digital subtraction angiography was performed in 43 of the 46 cases, with a mean of 2.63 runs each (range, 1–5). Screen-film conventional angiography was performed in five cases, with a mean of 2.00 runs each (range, 1–3). In two of these cases, a combination of digital subtraction angiography and screen-film conventional angiography was performed.
Selective injections were performed in six cases, with a mean of 2.33 additional runs each (range, 1–4). Overall, the mean number of runs per case, including selective injections, was 2.98 (range, 1–8). All six cases, in which selective injections were performed, were acquired after a perigraft leak first had been identified at CT angiography (five of six) or aortic injection at conventional angiography (one of six).
Data Collection
Three faculty radiologists (S.M.S., E.W.O., S.L.S.) with 6–10 years of experience performing conventional angiography independently interpreted the conventional angiographic studies, and three different radiologists (G.D.R., C.F.B., R.B.J.) with 6–15 years of experience interpreting CT images independently interpreted the CT angiographic studies. None of the six readers were involved in the clinical care of the patients, which eliminated the possibility that a patient might be recognized from a characteristic image, which would potentially bias the results. Each reader was blinded to the patient's clinical data and outcome. Each reader was shown all of the contrast medium–enhanced CT angiographic or conventional angiographic images, including all projections and selective injections.
A single scoring form was presented to each reviewer for each of the 46 cases reviewed. The scoring sheet included anticipated complications associated with aortic stent-graft placement: perigraft leakage, occlusion of major aortic branches, graft thrombosis (presence of any thrombus in the stent-graft), false aneurysm development due to aortic perforation, stent-graft collapse or incomplete deployment, and opacification of intercostal or lumbar arteries covered by the stent-graft. Each reader was asked to indicate if each of the complications was present or not present and to assign a confidence level of 1–5, with 5 being most confident, to each observation. If perigraft leakage was detected, then the reviewers were asked additionally to determine the site of leakage as proximal, distal, or the middle of the graft or via a patent aortic branch.
Statistical Methods
To assess the accuracy of CT angiographic and conventional angiographic interpretations for perigraft leakage and branch occlusion, a reference standard interpretation was developed by consensus of a CT radiologist (G.D.R.) and the primary angiographer (M.D.D.) who had placed the stent-grafts and was familiar with each patient's case and follow-up. These two radiologists together reviewed all available imaging studies prior to determining the presence of perigraft leakage and aortic branch occlusion. A consensus interpretation was achieved in all 46 cases. In addition, selective and superselective arteriographic injections around the stent-graft were performed in six patients within 24 hours of suspicious CT or conventional angiographic findings. These images were additionally available for the consensus review.
The scoring sheet results were tabulated for each reader. For all patients in whom perigraft leakage had been determined, the site of leakage was determined as the majority opinion of all readers who identified the leakage. If there was no majority opinion, then the site of leakage was considered indeterminate.
The maximum dimensions of all perigraft leaks were measured as Dx, Dy, and Dz. The size of each perigraft leak was estimated as the volume of an ellipsoid given as 1/6
(Dx x Dy x Dz). Leaks were categorized as small (<1 mL), moderate (1–10 mL), or large (>10 mL).
The sensitivities, specificities, and mean confidence levels were calculated for perigraft leakage and branch occlusion. The value was calculated for perigraft leak detection by means of conventional angiography and CT angiography by using the method described by Fleiss (6). Statistical significance was tested with a two-tailed Student t test.
Because we did not know the prevalence of various complications, prospectively, the reviewers analyzed all potential complications. However, statistical analysis was performed for only those complications with sufficient prevalence to provide statistically significant results: perigraft leakage and branch occlusion. Less prevalent complications are reported descriptively.
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
values for each reader and overall for both conventional angiography and CT angiography are also presented in Table 1. The overall value was 0.41 for conventional angiography (range, 0.27–0.63) and 0.81 for CT angiography (range, 0.73–0.91). Using the terms suggested by Landis and Koch (7), we found the strength of agreement for conventional angiography was "moderate," and for CT angiography it was "almost perfect." Confidence scores for both correct and incorrect diagnoses are listed in Table 2. All readers gave relatively high confidence levels when scoring leakage correctly (conventional angiography mean, 4.6; CT angiography mean, 4.4). CT angiography readers were much less confident than conventional angiography readers when reporting incorrect scores (conventional angiography mean, 4.1; CT angiography mean, 2.9). These values were all significantly different from each other, with a P value less than .01 as measured by means of a two-tailed t test.
|
|
|
|
|
|
|
|
|
The diagnostic accuracy of digital subtraction angiography or screen-film conventional angiography with selective injections was 100% (six of six), with a mean confidence level of 4.8; sensitivity was 100%, but specificity could not be calculated, because selective injections were not performed in any cases without a leak according to the reference standard. In five of the six cases, selective injection was performed after a leak had been identified at CT. In only one case was selective injection performed without the benefit of prior CT results.
The official radiology reports were reviewed for all of the 64 patients excluded from our study. This review revealed no evidence that inclusion of these patients would have substantially altered our study cohort. For patients who underwent both postdeployment conventional angiography and CT, and for whom radiology reports were available, the percentage with reported leaks was 30% (13 of 44) for conventional angiography and 41% (18 of 44) for CT. A similar review of the official radiology reports for the study cohort revealed a reported leak rate of 28% (13 of 46) for conventional angiography and 41% (19 of 46) for CT.
Occlusion of Major Aortic Branches
Ten of 46 patient studies showed at least one aortic branch origin occluded. These included five left subclavian arteries and five hypogastric arteries. In all cases, these branches were purposely occluded prior to stent-graft deployment by means of transplantation of the left subclavian artery onto the middle of the left common carotid artery or by means of coil embolization of the hypogastric artery. In this series, there were no cases in which aortic branches were inadvertently occluded by the stent-graft. Because there were substantial variations in the scoring of coil embolization of the hypogastric arteries by some readers, in spite of obvious findings on the corresponding images, cases of hypogastric arterial occlusion were excluded from our analysis.
Five, four, and one of the five left subclavian arterial occlusions were detected by each of the three CT angiography readers, whereas three, five, and one of the left subclavian arterial occlusions were detected by each of the three conventional angiography readers. Overall, the sensitivities of CT angiography and conventional angiography for depicting left subclavian arterial occlusion after stent-graft deployment were 67% (10 of 15) and 60% (nine of 15), respectively, and were not significantly different. The mean specificities of CT angiography and conventional angiography were 93% (100 of 108) and 87% (94 of 108), respectively, which were also not significantly different. Therefore, although there was substantial variability among readers for the detection of left subclavian arterial occlusions, the overall performances of CT angiography and conventional angiography were similar.
Opacification of Intercostal or Lumbar Arteries
Opacified intercostal arteries were observed over the stent-graft more frequently with CT angiography than with conventional angiography (41% [19 of 46] and 24% [11 of 46], respectively). We hypothesized that perigraft leakage might be associated with the opacification of intercostal or lumbar arteries. To evaluate this possibility, perigraft leakage was correlated with opacification for each case. On CT angiograms, the prevalence of leakage in cases with opacification was 37% (seven of 19); the prevalence of leakage in cases without opacification was 48% (13 of 27). On conventional angiograms, the prevalence of leakage in cases with opacification was 45% (five of 11); the prevalence of leakage in cases without opacification was 43% (15 of 35). Thus, perigraft leakage did not correlate with a higher prevalence of intercostal or lumbar arterial opacification.
Graft Thrombosis
Graft thrombosis was an infrequently reported complication. All three CT angiography readers identified thrombosis in two patients. At least two of three conventional angiography readers also identified thrombus in these same patients.
Stent-Graft Collapse or Incomplete Deployment
At least two of three CT readers rated the stent-graft as incompletely deployed or partially collapsed in seven patients (mean confidence level, 4.2), and all three agreed in four patients (mean confidence level, 4.4) (Fig 5). There were no patients in whom conventional angiograms were interpreted as demonstrating incomplete deployment or partial collapse by more than one reader. There were four instances in which collapse was scored by one of the three conventional angiography readers (one, two, and one for each reader) with a mean confidence level of 3.5; however, none of the CT readers identified collapse in these patients (mean confidence level, 4.4).
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
The ultimate determinants of successful stent-graft deployment are an arrest or regression in the growth of the aneurysm and ultimately the prevention of aneurysm rupture. Once a stent-graft has been successfully deployed, the best early predictor of its long-term success is the complete exclusion of flow into the aneurysm sac. Although we do not know which perigraft leaks ultimately will seal spontaneously and which will require secondary intervention, the identification of these leaks is critical to our understanding of the rapidly evolving field of endovascular repair of aortic aneurysms.
In our study population of patients predominantly with thoracic aortic aneurysms, we established that CT angiography is significantly more accurate for diagnosing perigraft leakage than is conventional angiography. Because of the implications for continued aneurysm growth, false-negative diagnoses warrant particular concern. The prevalence of false-negative diagnoses at conventional angiography was reflected in an overall sensitivity of only 63%. Two examples of large perigraft leaks missed by all three conventional angiography interpretations are illustrated in Figures 1 and 2. Although motion can be an important limitation of image quality at thoracic conventional angiography, these two examples are not substantially limited by image quality.
We believe that the most likely explanation for the lack of perigraft leak visualization relates to the high flow that was present within the thoracic aorta and the relatively short duration (typically 2 seconds) of contrast medium injection that was used. The flow through the narrow channels supplying the patent portions of the aneurysm sac is likely to be very slow, thus virtually all of the contrast medium injected locally into the thoracic aorta passes over the perigraft channel, which prevents its opacification. CT angiography relies on a peripheral venous injection of at least 25–30 seconds. As a result, the entirety of the vascular system will be opacified for 10–15 times longer than the thoracic aorta is opacified at conventional angiography, which allows time for nonenhanced blood within the patent aneurysm sac to be replaced by enhanced blood. An additional explanation for the lack of perigraft leak detection at conventional angiography in some cases may be attributed to projections that do not allow a leak to be viewed in profile.
False-negative diagnoses were substantially less common with CT angiography. The leaks missed with CT angiography tended to be very small and were likely overlooked. This suggests that proper training and exposure to the CT appearance of small perigraft leaks should precede routine interpretation of CT angiograms for detecting perigraft leakage.
False-positive diagnoses occurred with greater frequency with conventional angiography than with CT angiography (77% versus 90% overall specificity). Although conventional angiographic sensitivities were similar for all three reviewers, specificity values were substantially more variable. Although it was difficult to determine retrospectively the source of all false-positive diagnoses, many related to regions where the polyester covering billowed out from the metallic endoskeleton between the sutures that secured the graft to the stent at each end, which simulated perigraft leakage (Fig 3). One reviewer of conventional angiograms was not fooled by this finding and scored a specificity of 100%. This result suggests that the limited specificity of conventional angiography may be substantially eliminated through experience and education about this pitfall. Furthermore, this pitfall may be specific to our device and might not be a problem with devices in which the graft material is tethered to the stent over the entire length of the device. All false-positive CT angiographic diagnoses were also because of graft billowing (Fig 4).
Another interesting result of our analysis was the confidence with which correct and incorrect diagnoses were made by the two groups of reviewers. These data are summarized in Table 2 and indicate that there are cues on CT angiograms that reduce an interpreter's confidence in the diagnosis when it is incorrect. These cues were significantly less prevalent on conventional angiograms.
The prevalence of perigraft leakage, or "endoleak," has varied widely in published reports, from a low of 4% to a high of 44% (8–20). Development and progression of these leaks are highly variable. Leaks may occur immediately after stent-graft deployment or may develop any time later. Whereas some leaks require repair, many spontaneously resolve. Leaks may occur at the ends of the stent-graft or may result from retrograde filling from an aortic branch. The potential consequences of such leaks are quite serious, as rapid aneurysm enlargement has been reported in the setting of perigraft leakage (21). Therefore, effective postoperative surveillance is essential.
Alternative noninvasive methods for assessing the aorta after stent-graft deployment include ultrasonography (US) and magnetic resonance (MR) imaging. In a study of abdominal aortic aneurysms after endovascular repair, both CT angiography and US depicted four perigraft leaks (22). Although the relative ability of CT angiography and US to depict perigraft leaks has not been assessed rigorously, for evaluation of the thoracic aorta US is not an option.
There is limited experience in the use of MR imaging to assess aortic aneurysms after endoluminal repair. Authors of one article (23) describe the finding of five of 15 patients with major and 10 of 15 patients with minor perigraft leakage detected at MR imaging after endovascular treatment of abdominal aortic aneurysms. The authors did not describe criteria for differentiating major from minor leaks. Although neither a formal protocol for image evaluation nor a reference standard for the presence of perigraft leakage was described, the authors stated that MR imaging was superior to CT scanning for detecting the minor leaks because of "pronounced metal artifacts" at CT. The CT examinations in this study were limited by the use of thick (7–10-mm) sections and slow (1.5–2.0 mL/sec) administration of a low-volume (90-mL) bolus of iodinated contrast medium. Nevertheless, these technical limitations do not explain the genesis of "pronounced metal artifacts," which we did not observe in our study.
One likely explanation might relate to the selection of window width and level for CT display. Just as standard mediastinal or abdominal display settings are insufficient for assessing bony details, they are also insufficient for assessing the metallic components of stent-grafts, which typically require window levels of 200–600 HU and widths of 600–1,200 HU.
Engellau and colleagues (23) assessed stent-grafts in which nitinol endoskeletons were used; however, stent-grafts are manufactured with both nitinol and stainless steel. Although both of these metals have similar CT attenuation, nitinol results in substantially less magnetic susceptibility artifact on MR images when compared with stainless steel, which is ferromagnetic. The role of MR imaging for the surveillance of endovascular stent-grafts will require a thorough study of the influence of magnetic susceptibility artifacts associated with a variety of endovascular stent-grafts, pulse sequences, and echo times.
Although the presence of a perigraft leak is the main indicator of unsuccessful aneurysm exclusion after stent-graft deployment, an understanding of which perigraft leaks will ultimately resolve and which will lead to continued aneurysm enlargement is critical in caring for patients after stent-graft deployment and in assessing the adequacy of deployment techniques and stent-graft devices in this rapidly evolving field. This understanding can come only after prolonged periods of patient follow-up. Until these data are available, it is impossible to know which characteristics of a perigraft leak are most relevant to ultimate outcome.
We chose to quantify the size of the leaks and characterize the location by means of readily available data from CT angiographic and conventional angiographic examinations. The sizes of the perigraft leaks in our study were objectively measured; and although the absolute values of the measures were estimated as an ellipsoid of the maximal diameters in each of the three dimensions, their relative sizes are valid and serve as an index of perigraft leak size.
In addition to improved leak detectability, an advantage that CT holds over conventional angiography for assessing perigraft leaks is that the size of the leaks always can be measured in three dimensions, which allows greater characterization. The location of perigraft leakage did not correlate with missed conventional angiographic diagnoses; however, it is interesting to note that the location of the perigraft leakage was not consistently characterized in four (20%) of 20 cases.
Aortic branch occlusion is a potentially serious complication of endovascular stent-graft deployment. There were no patients with inadvertent aortic branch occlusions in our study population. However, there were five patients in whom left subclavian arterial occlusion was anticipated before deployment, because of the proximity of the proximal aneurysm neck to the origin of the left subclavian artery. In these patients, the left subclavian artery was transplanted to the left common carotid artery through a supraclavicular incision prior to stent-graft deployment. This procedure resulted in thrombosis of the left subclavian arterial stump proximal to the transplantation site in all patients after stent-graft deployment. The readers, who were blinded to patient surgical history, were asked to identify any aortic branch occlusions. Nevertheless, there was substantial variability among both CT and conventional angiography interpreters. Although the findings of subclavian occlusion were present in the results of all five CT angiographic and conventional angiographic examinations, only one CT angiography and one conventional angiography reader identified all five. This result underscores the importance of a careful inspection of all aortic branches near the stent-graft, as aortic branch occlusions may be subtle.
Three limitations can be identified in our study. First, our cohort was composed predominately of patients with thoracic aortic aneurysms, whereas most stent-grafts are placed in the abdominal aorta to treat abdominal aortic aneurysms. This reflects the patient population treated at our hospital between July 1992 and June 1996. Of the 104 patients who received a stent-graft during the study period, 80% had a thoracic aneurysm; and of the 40 patients who met our entry criteria, 90% had a thoracic aneurysm.
Second, another limitation was retrospective recruitment. Although prospective recruitment might have resulted in the inclusion of some patients whose imaging studies were irretrievable at the time of retrospective recruitment, a significant selection bias due to irretrievable studies is unlikely. Furthermore, interpretations were made prospectively by radiologists who were blinded. Since aortic stent-graft placement has been a relatively infrequent procedure, a retrospective design enabled us to maximize the number of cases available for inclusion in our study and to assess interobserver performance. A standardized scoring sheet was used to facilitate quantification and comparison of the results.
Third, because a true independent reference standard was unavailable in our cohort, we relied on a "surrogate truth" composed of a consensus reading in which all available imaging and clinical data were used to determine if a perigraft leak or aortic branch occlusion were present (24). The interventional radiologist involved in the stent-graft placement and follow-up of our cohort participated in the establishment of this "truth."
In summary, perigraft leakage was the most prevalent complication in our study cohort. Our data suggest that CT angiography is a more accurate technique for the detection of perigraft leakage than is conventional angiography and that there is substantially greater correlation between the confidence level of diagnosis and that diagnosis being correct for CT angiography than for conventional angiography. We found no advantage for the use of postprocedural conventional angiography over CT angiography for the detection of postdeployment complications.
|
Footnotes |
|---|
Author contributions: Guarantors of integrity of entire study, G.D.R., M.D.A.; study concepts, G.D.R.; study design, G.D.R., M.D.A., M.D.D.; definition of intellectual content, G.D.R., M.D.D.; literature research, M.D.A.; clinical studies, M.D.A., G.D.R., C.F.B., E.W.O., S.L.S., S.M.S., C.P.S., R.B.J., M.D.D.; data acquisition, M.D.A., G.D.R., C.F.B., E.W.O., S.L.S., S.M.S., M.J.J.; data analysis, M.D.A., G.D.R.; statistical analysis, M.D.A., G.D.R.; manuscript preparation, M.D.A., G.D.R.; manuscript editing, G.D.R.; manuscript review, C.F.B., S.M.S., C.P.S., R.B.J., M.D.D., E.W.O., S.L.S.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  P Sharma and C Kyriakides Surveillance of patients post-endovascular aneurysm repair Postgrad. Med. J., December 1, 2007; 83(986): 750 - 753. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Burrill, Z. Dabbagh, F. Gollub, and M. Hamady Multidetector computed tomographic angiography of the cardiovascular system Postgrad. Med. J., November 1, 2007; 83(985): 698 - 704. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. W. Stavropoulos and S. R. Charagundla Imaging Techniques for Detection and Management of Endoleaks after Endovascular Aortic Aneurysm Repair Radiology, June 1, 2007; 243(3): 641 - 655. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. Chernyak, A. M. Rozenblit, M. Patlas, J. Cynamon, Z. J. Ricci, M. P. Laks, and F. J. Veith Type II Endoleak after Endoaortic Graft Implantation: Diagnosis with Helical CT Arteriography Radiology, September 1, 2006; 240(3): 885 - 893. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Kubo, E. Tadamura, M. Yamamuro, R. Hosokawa, T. Kimura, T. Kita, M. Komeda, and K. Togashi Thoracoabdominal-aortoiliac MDCT angiography using reduced dose of contrast material. Am. J. Roentgenol., August 1, 2006; 187(2): 548 - 554. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Tolia, R. Landis, P. Lamparello, R. Rosen, and M. Macari Type II Endoleaks after Endovascular Repair of Abdominal Aortic Aneurysms: Natural History Radiology, May 1, 2005; 235(2): 683 - 686. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. M. Ho, R. C. Nelson, J. Thomas, E. I. Gimenez, and D. M. DeLong Abdominal Aortic Aneurysms at Multi-Detector Row Helical CT: Optimization with Interactive Determination of Scanning Delay and Contrast Medium Dose Radiology, September 1, 2004; 232(3): 854 - 859. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Olin, J. A. Kaufman, D. A. Bluemke, R. O. Bonow, M. D. Gerhard, M. R. Jaff, G. D. Rubin, and W. Hall Atherosclerotic Vascular Disease Conference: Writing Group IV: Imaging Circulation, June 1, 2004; 109(21): 2626 - 2633. [Full Text] [PDF]
|
|
|
|
|
|
A. L. Greenfield, E. J. Halpern, J. Bonn, R. J. Wechsler, and M. B. Kahn Application of Duplex US for Characterization of Endoleaks in Abdominal Aortic Stent-Grafts: Report of Five Cases Radiology, December 1, 2002; 225(3): 845 - 851. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Bosch, J. A. Kaufman, M. T. Beinfeld, M. E. A. P. M. Adriaensen, D. C. Brewster, and G. S. Gazelle Abdominal Aortic Aneurysms: Cost-effectiveness of Elective Endovascular and Open Surgical Repair Radiology, November 1, 2002; 225(2): 337 - 344. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. D. Rubin, A. J. Schmidt, L. J. Logan, and M. C. Sofilos Multi-Detector Row CT Angiography of Lower Extremity Arterial Inflow and Runoff: Initial Experience Radiology, October 1, 2001; 221(1): 146 - 158. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Macari, G. M. Israel, P. Berman, M. Lisi, A. J. Tolia, M. Adelman, and A. J. Megibow Infrarenal Abdominal Aortic Aneurysms at Multi-Detector Row CT Angiography: Intravascular Enhancement without a Timing Acquisition Radiology, August 1, 2001; 220(2): 519 - 523. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. D. Rubin, M. C. Shiau, A. N. Leung, S. T. Kee, L. J. Logan, and M. C. Sofilos Aorta and Iliac Arteries: Single versus Multiple Detector-Row Helical CT Angiography Radiology, June 1, 2000; 215(3): 670 - 676. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||
