| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Vascular and Interventional Radiology |
1 From the Department of Clinical Science and Bioimaging, Section of Radiology, G. D'Annunzio University, SS Annunziata Hospital, Via dei Vestini, 66013 Chieti, Italy (R.I., A.R.C., A.F., F.D.F., F.Q., C.C.); and Department of Bioimaging and Radiological Science, Institute of Radiology, Catholic University, Rome, Italy (L.B.). Received June 8, 2005; revision requested August 2; revision received October 13; accepted November 14; final version accepted February 2, 2006. Address correspondence to R.I. (e-mail: r.iezzi{at}rad.unich.it).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Materials and Methods: Fifty patients (two women, 48 men; mean age, 72 years) underwent follow-up multi–detector row CT 1, 6, and 12 months after endovascular aneurysm repair. Unenhanced CT was performed with 2.5-mm collimation; 1-mm collimation was used with AEP and DEP examinations. Two independent readers assessed the presence of endoleak in three reading sessions: AEP (session A), unenhanced and AEP (session B), and AEP and DEP (session C). At 6- and 12-month follow-up, a fourth set was included: 1-month unenhanced and AEP (session D). Sensitivity, specificity, and positive predictive value of each session were calculated. Triple-phase multi–detector row CT was the reference standard.
Results: At 1 month, sensitivity, specificity, and positive predictive value, respectively, were 79%, 75%, and 55% for session A; 93%, 97%, and 93% for session B; and 93%, 78%, and 62% for session C. At 6 months, sensitivity, specificity, and positive predictive value, respectively, were 92%, 68%, and 48% for session A; 92%, 100%, and 100% for session B; and 100%, 84%, and 67% for session C. At 12 months, sensitivity, specificity, and positive predictive value, respectively, were 80%, 80%, and 50% for session A; 90%, 98%, and 90% for session B; and 100%, 80%, and 56% for session C. Sensitivity did not significantly differ (P > .05) among reading sessions A, B, and C, whereas specificity and positive predictive values in session B were significantly higher (P < .001). For 6- and 12-month follow-up, no significant differences (P > .05) were found between sessions D and B.
Conclusion: The combination of AEP and unenhanced imaging performed at 1-month follow-up offers improved specificity and positive predictive values compared with AEP alone. DEP imaging does not significantly increase sensitivity for detection of endoleaks, but it does depict low-flow endoleaks not seen at AEP.
© RSNA, 2006
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
It is generally accepted that endoleaks are better depicted during a phase of maximal arterial enhancement known as the arterial phase. Several authors have investigated the role of unenhanced and delayed phase CT in the detection of endoleaks and have advocated various combinations of these three CT acquisitions (unenhanced, arterial phase, and delayed phase) for achieving an optimal protocol (9–12).
Because the acquisition of images at an additional phase leads to increased patient exposure to ionizing radiation and an increase in the time for CT examination and image interpretation, it is important to clarify the advantages of unenhanced and delayed phase CT to warrant their use. Thus, the aim of our study was to determine the sensitivity and specificity of unenhanced, delayed enhanced phase, and arterial enhanced phase multi–detector row CT for depicting endoleaks during follow-up of endovascular aneurysm repair, with the triple-phase examination findings serving as the reference standard.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
From June 2001 to July 2002, 50 consecutive patients (two women, 48 men; mean age, 72 years; age range, 63–89 years) with infrarenal abdominal aortic aneurysm underwent elective endovascular aneurysm repair. The endovascular grafts used in these patients consisted of bifurcated grafts (n = 48) (Talent, Medtronic, Minneapolis, Minn [n = 37]; Excluder, WL Gore, Newark, De [n = 10]; AneuRx, Medtronic [n = 1]) and aortomonoiliac grafts with contralateral iliac occlusion and crossed femoro-femoral bypass grafting (n = 2) (Talent; Medtronic). All patients underwent follow-up CT 1, 6, and 12 months after stent-graft implantation.
CT Scanning
Triple-phase CT (unenhanced and contrast-enhanced transverse scanning in the arterial and delayed phases) was performed with a multi–detector row helical scanner (Somatom Plus 4 Volume Zoom; Siemens Medical Systems, Forchheim, Germany) at each follow-up CT study.
Unenhanced CT images were obtained from the level of the diaphragm to the symphysis pubis, with collimation of 2.5 mm, pitch of 6, section width of 5 mm, reconstruction increment of 5 mm, table speed of 15 mm per rotation, and gantry rotation time of 0.5 second. Contrast-enhanced images were obtained after bolus intravenous injection of 120 mL of iodinated nonionic contrast medium (iomeprol [300 mg of iodine per milliliter], Iomeron; Bracco, Milan, Italy) at a flow rate of 3 mL/sec via an antecubital vein, with collimation of 1 mm, pitch of 6, section width of 1.25 mm, reconstruction increment of 1 mm, table speed of 15 mm per rotation, and rotation time of 0.5 second. Arterial phase images were acquired from the suprarenal abdominal aorta to the common femoral artery. Scan delay was individualized for each patient by using proprietary bolus-tracking software (CARE Bolus; Siemens Medical Systems) that measures the inflow of contrast material in a region of interest placed on the abdominal aorta at the level of the celiac trunk and automatically triggers the start of the examination when it registers a density of 100 HU. Delayed phase acquisition, focused on the endovascular graft, was performed 60 seconds after contrast material injection.
Image Analysis
The 1-, 6-, and 12-month follow-up CT images were divided into different reading sessions to be evaluated independently by two experienced blinded readers (A.F., R.I.) with 10 and 5 years of experience in body CT. These readers were unaware of imaging findings (presence or absence of endoleak) and aneurysmal sac evolution (in comparison with the diameter of the aneurysmal sac before treatment or at follow-up CT).
The different sets of images consisted of arterial phase images in reading session A, unenhanced and arterial phase images in session B, and arterial and delayed phase images in session C. For the analysis of 6- and 12-month follow-up CT images, we added another evaluation (session D) that consisted of 6- or 12-month follow-up arterial phase and 1-month follow-up unenhanced images.
During each reading session, the readers separately and independently assigned a confidence level for the diagnosis of endoleak by using a five-point scale: 1, certainly absent; 2, probably absent; 3, possibly present; 4, probably present; and 5, certainly present. Before readers evaluated images, they were informed that a confidence level of 3 or higher represented a positive diagnosis of endoleak.
For objectivity and reproducibility of the image analysis performed in this study, standard criteria for endoleak diagnosis were provided. During reading session B, which included unenhanced and arterial phase images, the presence of endoleak was considered probable or certain if a high-attenuation area was present beyond the graft but within the aneurysmal sac on the arterial phase images but absent on the unenhanced images. On the other hand, during the reading sessions that included contrast-enhanced phases without unenhanced images (sessions A and C), the presence of endoleak was considered probable or certain if a high-attenuation area, different from calcifications, was present beyond the graft but within the aneurysmal sac. The evaluation was based on visual assessment, without measurement of the attenuation.
The readers reviewed entire sets of images in random order at a dedicated computer workstation (Leonardo; Siemens Medical Systems). To minimize learning bias, the names, ages, and identification numbers of patients and the imaging parameters were always hidden during the review. The specific image sets were randomly selected by an author (F.Q.) with 3 years of experience in body CT, who did not participate in the image review. The interval between the reading sessions was at least 1 month.
Reference Standard
Our reference standard both for detection and for exclusion of endoleak was evaluation of triple-phase (unenhanced, arterial phase, and delayed phase) CT images performed in consensus by two experienced readers (A.R.C., F.D.F.) with 15 and 5 years of experience in body CT who were not involved in image analysis and knew the previous CT findings. These two readers were asked to classify the endoleak, as follows: type I, leak due to incomplete attachment of the proximal and the distal tract of the prosthesis caused by technical or anatomic problems; type II, leak due to regressed flow in the aneurysmal sac caused by secondary arteries; type III, leak due to holes in the prosthesis or detachment of its various sections; type IV, leak due to porosity of the prosthesis; and type V or endotension, no actual leak seen but evidence of persistent pressure in the aneurysmal sac, demonstrated by an increase in the diameter of the aneurysm (4). All endoleaks detected only at delayed phase CT were classified as low-flow leaks (2).
The readers also compared changes in the aneurysmal sac (increment, stability, or reduction) with previous CT findings by measuring the largest diameter of the aneurysm perpendicular to the aortic axis on the transverse images by using an electronic cursor. A change in the size of the aneurysmal sac was recorded if this change was at least 2 mm.
Statistical Analysis
Interobserver agreement for the evaluation of CT images was assessed with the 
statistic. Degrees of agreement were categorized as follows: values of 0.00–0.20 were considered to indicate poor agreement; values of 0.21–0.40, fair agreement; values of 0.41–0.60, moderate agreement; values of 0.61–0.80, high agreement; and values of 0.81–1.00, excellent agreement (13).
Sensitivity, specificity, and positive predictive values for detection of endoleaks in each set of images at 1-, 6-, and 12-month follow-up were calculated and compared by using a two-sample test of proportion. All two-tailed P values less than .05 were considered to indicate a significant difference. Statistical analyses were performed with Stata statistical software, release 8.2 (StataCorp, College Station, Tex).
CT images with a score of 3 or higher that were confirmed as positive for endoleak with the reference standard were considered true-positive diagnoses, whereas CT images with a score of 1 or 2 that were confirmed as negative for endoleak with the reference standard were considered true-negative diagnoses. False-negative diagnoses were represented by CT images with a score of 1 or 2 that were confirmed as positive for endoleak with the reference standard, whereas false-positive diagnoses were represented by CT images with a score of 3 or higher that were confirmed as negative for endoleak with the reference standard.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
|
An increase in the size of the aneurysmal sac associated with endoleak was also detected in four of 32 (12%) type II endoleaks (two of 15 patients). In the first patient, type II endoleak detected at 1-month and 6-month follow-up CT spontaneously disappeared at 12-month follow-up; in the other patient, endoleak persisted at all CT follow-up examinations but no further treatment was performed because of the decrease in the size of the aneurysmal sac at 12-month follow-up. Of the other 28 type II endoleaks, 21 aneurysmal sacs were stable and seven decreased in size.
Readers classified endoleaks as low-flow type II leaks in two of 18 patients (11%) at 1-month and 6-month follow-up CT, for a total of four of 36 (11%) leaks. In these four type II low-flow leaks, three aneurysmal sacs were stable and one decreased in size.
In all subsequent CT examinations with findings that were classified as negative for endoleak with the reference standard, the aneurysmal sac always decreased or remained unchanged, without any complication.
Reading Sessions
The analysis showed excellent interobserver agreement ( 
0.86) in all reading sessions for endoleak detection; thus, the data for the two readers were pooled.
When considering true-positive cases, reading session A (arterial phase images) revealed 11 of 14 endoleaks (79%) at 1 month, 11 of 12 endoleaks (92%) at 6 months, and eight of 10 endoleaks (80%) at 12 months; reading session B (unenhanced and arterial phase images) depicted 13 of 14 endoleaks (93%) at 1 month, 11 of 12 endoleaks (92%) at 6 months, and nine of 10 endoleaks (90%) at 12 months; reading session C (arterial plus delayed phase images) depicted 13 of 14 endoleaks (93%) at 1 month, 12 of 12 endoleaks (100%) at 6 months, and 10 of 10 endoleaks (100%) at 12 months (Fig 1).
|
|
|
|
|
|
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
In our institution, follow-up CT images were acquired 1, 6, and 12 months after stent-graft implantation and every year thereafter until reduction of the aneurysmal sac was complete (if no complications occur). It is generally accepted that arterial phase CT easily allows detection of relatively large perigraft endoleaks located away from the metallic component of the stent-graft or calcified aortic wall. However, detection of small endoleaks proximal to high-attenuation components (metallic portions of the stent-graft and calcifications) could be difficult and requires a thorough analysis of CT images.
Different authors advocate various combinations of unenhanced, arterial phase, and delayed phase CT for achieving optimal and clinically important detection of endoleaks. To the best of our knowledge, however, there has been no published work regarding the specificity and sensitivity of these different combinations in detecting endoleak.
In our series, no significant differences were found between reading session B and sessions A and C in terms of sensitivity; however, session B was significantly superior to sessions A and C in terms of specificity and positive predictive value.
Reading session B enabled a correct diagnosis in 96.7% (145 of 150 CT examinations) of cases. In comparing unenhanced images and arterial phase images with arterial phase images alone, the rate of endoleak detection increased and endoleaks could be excluded with more confidence. According to the literature, unenhanced images are mainly required to help in the evaluation of postcontrast data sets to detect small endoleaks and differentiate them from calcified aortic wall, intrathrombus calcifications, or metallic portion of the stent-graft (10,12,19,20). We agree with the findings of Rozenblit et al (11), who affirmed the clinical impossibility of Golzarian's (21) suggestion to identify, during CT scanning, patients who need to undergo additional delayed phase scanning to differentiate endoleaks and calcifications instead of systematically unenhanced scanning.
The remaining five cases that were not correctly diagnosed in reading session B were represented by three false-negative and two false-positive cases. Although all type I and type III endoleaks—as well as all type II endoleaks associated with an increase in the size of the aneurysmal sac—were correctly diagnosed, three type II low-flow leaks (false-negative cases) not associated with an increase in the aneurysmal sac were not diagnosed in session B. Nevertheless, low-flow leaks have uncertain clinical importance; their importance, according to the literature, is secondary to an eventual increase in the size of the aneurysmal sac. Because untreated type II endoleaks not accompanied by an increase in the size of the aneurysmal sac may seal spontaneously because of thrombosis, the majority of interventionists are prepared to observe them (22). In our series, low-flow type II leaks were not associated with any interrelated complications; no treatment has been required, and sac puncture and glue application or embolization have not been performed.
In reading session B, the presence of an endoleak was not correctly excluded in two patients in the negative group (false-positive cases). However, these suspected endoleaks were classified as type II leaks and were associated with a stable or smaller-sized aneurysmal sac. Therefore, images in reading session B led to an incorrect diagnosis in five cases; however, no patient had changes in his or her care or outcome based on these uncorrected diagnoses. Furthermore, in considering 6- and 12-month follow-up CT findings, significant differences in sensitivity and specificity were found between reading sessions B and D; therefore, we suggest that unenhanced images be acquired only at 1-month follow-up.
The main limitation of our study was the relatively small number of patients examined; further investigations with a large series of patients are needed to confirm our findings.
In conclusion, our results indicate that follow-up CT angiography in patients who have undergone endovascular aneurysm repair could include an initial precontrast acquisition and an arterial phase acquisition at 1-month follow-up, whereas only arterial phase acquisition can be performed at 6- and 12-month follow-up. Since this protocol does not include the delayed phase, there is a risk of missing low-flow endoleaks and of visualizing only the increase in the size of the aneurysmal sac that could eventually be associated with an erroneous diagnosis of endotension (2). In these patients, we suggest performing close CT follow-up, including all contrast-enhanced acquisition phases; contrast-enhanced ultrasonography, as suggested by Napoli et al (23); or angiography as a preliminary step before eventual endovascular treatment.
|
ADVANCES IN KNOWLEDGE |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, L.B.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, R.I., F.Q.; clinical studies, R.I., A.R.C., F.D.F., F.Q.; statistical analysis, R.I., F.Q.; and manuscript editing, R.I., A.R.C., A.F., F.D.F.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONADVANCES IN KNOWLEDGEReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  H. Chandarana, M. Godoy, I. Vlahos, A. Grasser, J. Babb, C. Leidecker, and M. Macari Abdominal Aorta: Evaluation with Dual-Source Dual-Energy Multidetector CT after Endovascular Repair of Aneurysms--Initial Observations Radiology, September 23, 2008; (2008) 2492080359. [Abstract] [Full Text]
|
|
|
|
|
|
P. Stolzmann, T. Frauenfelder, T. Pfammatter, N. Peter, H. Scheffel, M. Lachat, B. Schmidt, B. Marincek, H. Alkadhi, and T. Schertler Endoleaks after Endovascular Abdominal Aortic Aneurysm Repair: Detection with Dual-Energy Dual-Source CT Radiology, September 9, 2008; (2008) 2483080193. [Abstract] [Full Text]
|
|
|
|
 |
|
 C. Hong, J. P. Heiken, G. A. Sicard, T. K. Pilgram, and K. T. Bae Clinical Significance of Endoleak Detected on Follow-Up CT After Endovascular Repair of Abdominal Aortic Aneurysm Am. J. Roentgenol., September 1, 2008; 191(3): 808 - 813. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 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]
|
|
|
|
|
|
H. T. Abada, J. Golzarian, R. Iezzi, A. R. Cotroneo, M. L. Storto, and L. Bonomo Multidetector CT in Abdominal Aortic Aneurysm Following Endovascular Repair: How to Consider the Value of a Delayed Phase Radiology, November 1, 2007; 245(2): 610 - 611. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
