Journal of Vascular Surgery
Volume 49, Issue 3 , Pages 552-560, March 2009

Contrast-enhanced ultrasound versus color duplex ultrasound imaging in the follow-up of patients after endovascular abdominal aortic aneurysm repair

Department of Clinical Science and Bioimaging, Section of Radiology, University “G. D'Annunzio”, Chieti, Italy

Received 21 July 2008; accepted 3 October 2008. published online 12 January 2009.

Article Outline

Purpose

This study assessed the negative predictive value, sensitivity, specificity, and diagnostic accuracy of real-time contrast-enhanced ultrasound imaging (CEUS) in the detection of endoleaks in patients with abdominal aortic aneurysm (AAA) who underwent endovascular repair (EVAR) compared with unenhanced ultrasound imaging. Computed tomography angiography (CTA) was the gold standard. The secondary objective was to define the optimal dose of the second-generation contrast agent to routinely use in the CEUS examinations for endoleak detection.

Methods

The study enrolled 84 patients with unruptured AAA who were treated with EVAR and underwent CTA follow-up. In the same day, CTA (4- × 1-mm collimation, 1.25-mm slice width), unenhanced US imaging and CEUS imaging was performed in all patients. The CEUS studies were performed after an intravenous bolus injection of 1.2 mL and 2.4 mL of a second-generation contrast agent with continuous low-mechanical index (range, 0.01-0.04) real-time tissue harmonic imaging. The unenhanced US and CEUS studies were interpreted separately by two independent experienced readers to detect the presence of endoleaks by viewing recorded videotapes according to a five-point confidence scale. The standard of reference was represented by the consensus reading of CTA performed by two experienced radiologists not involved in the image analysis. Qualitative analysis as well as sensitivity, specificity, negative predictive value, and diagnostic accuracy in detecting endoleaks of each reading session were compared.

Results

CEUS imaging significantly improved the diagnostic performance of unenhanced US studies in the detection of endoleaks in terms of sensitivity (97.5% vs 62.5%), negative predictive value (97.3% vs 65.1%), accuracy (89.3% vs 63.1%), and specificity (81.8% vs 63.6%). The optimal dose of contrast agent to detect and characterize endoleaks was 2.4 mL. No adverse events were recorded during the study.

Conclusions

The results showed CEUS imaging is a fast, noninvasive, reliable, and valid alternative to multislice CTA for endoleak detection in endovascular aortic stent graft patients, and is superior to unenhanced US imaging. Contrast-enhanced ultrasound imaging should be performed using a recommended contrast medium dose of 2.4 mL.

 

Endovascular stent graft placement is a successful alternative to open vascular surgery for abdominal aortic aneurysm (AAA) treatment.1, 2, 3 Although endovascular aneurysm repair (EVAR) can reduce perioperative mortality, the rate of complications and hospitalization may necessitate reintervention during follow-up. Patient surveillance and early complication detection are crucial to determine the long-term performance of these devices.

The most common complication (2% to 45%) is endoleak, which is the persistent perigraft flow within the aneurysmal sac excluded by the stent graft.4, 5, 6, 7, 8, 9 Persistent endoleak is considered a procedural failure because it may cause enlargement and rupture of the aneurysm, representing the main indication for surgical late conversion.10

Contrast-enhanced helical computed tomography (CT) has shown high sensitivity for detecting endoleaks and is considered the gold standard in the follow-up of patients with stent graft implantation.8, 9 The European Collaborators on Stent/graft Techniques for Aortic Aneurysm Repair (EUROSTAR) Registry10 recommends follow-up consisting of CT angiography (CTA) examinations performed 1, 6, and 12 months after the procedure and subsequent annual investigations, unless complications develop.

CTA has limitations, however. The investigation is repeated several times, making radiation exposure a necessary concern. As recently reported,11 and also suggested by the American College of Radiology, the dose of unnecessary radiation needs to be reduced in diagnostic imaging. Therefore, another reliable diagnostic examination during follow-up would be useful.

Color duplex ultrasound (CDUS) imaging is noninvasive, does not use radiation or contrast medium, is less expensive, easy to perform, and widely available. However, the sensitivity of CDUS in the detection of endoleak varies considerably, from 42.9%12 to 97%,13 presumably because results depend not only on the skills and experience of the operator but also on the habitus and level of cooperation of the patient.

Early published experience indicates that a contrast US examination seems to improve the diagnostic sensitivity of US imaging and its accuracy in endoleak detection.14, 15 The primary objective of this prospective, single-center study was to assess the negative predictive value, sensitivity, specificity, and diagnostic accuracy of real-time contrast-enhanced ultrasound (CEUS) imaging in the detection of endoleaks in patients who underwent EVAR compared with unenhanced US imaging using CTA as the gold standard. The secondary objective was to define the optimal dose of the second-generation contrast agent to routinely use in CEUS examinations for endoleak detection.

Back to Article Outline

Materials and methods 

Study design 

The study received local ethics board approval. Patients provided written informed consent, waived by the ethics board, for CEUS and multidetector CT examinations.

The study enrolled all patients treated with EVAR who underwent CTA as part of a routine surveillance program at 1, 6, and 12 months after the procedure and annually thereafter. They underwent CTA and CDUS and CEUS imaging on the same day. Patients with unstable general conditions, such as heart failure (New York Heart Association class IV), severe chronic bronchopulmonary disorders, severe pulmonary hypertension, or uncontrolled hypertension were excluded.

To avoid selection bias in favor of patients who were “easy to scan,” patients were recruited before undergoing a baseline US scan. No patient was excluded on the basis of poor technical quality of the baseline US study.

Ultrasound imaging 

The precontrast and postcontrast US scans were performed by a single radiologist who was a specialist in vascular radiology and experienced in the use of ultrasound contrast material (R. I. with 8 years of experience). The radiologist was blinded to all other imaging findings at the time of examinations.

All US scans were performed with a Philips HDI 5000 scanner (Philips Medical Systems, Bothell, Wash), equipped with 10.4 software, with a convex multifrequency 5- to 2-MHz probe. CEUS was performed after administration of a second-generation contrast agent (SonoVue, Bracco, Milan, Italy) made of sulfur hexafluoride-filled microbubbles with flexible shells that allow real-time imaging at low acoustic pressure (mechanical index range, 0.12-0.14). Axial and longitudinal and acquisition scans were used for the US imaging.

The CEUS scans were performed after the administration of a bolus of two different doses of contrast agent dissolved in 0.9% saline solution (1.2 mL and 2.4 mL), each followed by flushing with an injection of a 5-mL bolus of saline solution through an 18- to 20-gauge cannula placed in an arm vein. A minimum interval of 10 minutes and complete bubble destruction, which was achieved by scanning the entire abdominal aorta at a high mechanical index, were required between the two injections to avoid carryover effects. Scanning was started at the beginning of contrast agent injection and the sweep was usually completed within 5 minutes. The phases of CEUS were defined as arterial (10 to 40 seconds after contrast agent injection) and late (90 to 300 seconds after injection).

The CDUS and CEUS image data were recorded on videotapes in digital format for subsequent analysis.

Computed tomography angiography 

Triple-phase acquisition (unenhanced and contrast-enhanced transverse imaging, in arterial and delayed phases) was performed using a multidetector row helical scanner (Somatom Plus 4 Volume Zoom; Siemens, Forccheim, Germany) at each follow-up CT study. Unenhanced images were obtained with a slice collimation of 2.5 mm, whereas a 1-mm slice collimation was used for contrast-enhanced acquisitions, obtained after bolus intravenous injection of 120 mL of iodinated nonionic contrast medium (Iomeprol 300 mgI/mL, Iomeron; Bracco) at a flow rate of 3 mL/s through an antecubital vein. Delayed-phase acquisition, focused on the endovascular graft, was performed 60 seconds after contrast medium injection.

Analysis of US images 

Cine loop sweeps from the US examinations were randomly reviewed independently by two radiologists not involved in the imaging, one radiologist specialized in vascular radiology (D. P. with 10 years of experience) and the other in CEUS (R. B. with 15 years of experience), and neither was aware of the CTA outcomes or dose of contrast used for CEUS. They reviewed videotapes of each patient during three different sessions:

1.The baseline unenhanced US scan—session A (CDUS),

2.CEUS after the administration of 1.2 mL of the contrast agent—session B, low-dose contrast-enhanced (LDCE) US imaging, and

3.CEUS after the administration of 2.4 mL contrast medium—session C, high-dose contrast-enhanced (HDCE) US imaging.

The tapes were viewed at an interval of at least 1 week to reduce their memory of previous images. The readers were blinded to image session sequence; furthermore, names, ages, and identification numbers of patients, as well as imaging parameters were always hidden during the review.

The readers independently assigned a confidence level for endoleak diagnosis using a 5-point scale: 1, certainly absent; 2, probably absent; 3, possibly present; 4, probably present; and 5, certainly present. The readers were informed that a confidence level of 3 or higher represented a positive diagnosis of endoleak. In case of significant disagreement between the two readers, the final decision to modify the diagnosis was based on a second evaluation performed in “consensus.”

For image analysis objectivity and reproducibility, standard criteria for endoleak diagnosis were provided. During the reading session that included the baseline unenhanced US images, the presence of endoleak was considered probable or certain if a color duplex signal was present beyond the graft. During the reading session of the CEUS imaging, the presence of endoleak was considered probable or certain if a high attenuation area, absent on the baseline unenhanced-phase images, due to the presence of contrast enhancement, was present beyond the graft but within the aneurysm sac. The evaluation was based on visual assessment, without attenuation measurements. The baseline and CEUS images were not evaluated side by side with a split screen mode so that the evaluation could be performed as it would in a real practice. Readers were asked to classify the endoleak in cases with scores of 4 or 5.

The radiologists were also asked to provide a qualitative assessment of the visualization of various parts of the endovascular stent graft at the levels of the proximal anastomosis, main body of the prosthesis, right and left branch, and right and left distal anastomosis. They used a 5-point scale: 0, not visible; 1, poor visualization; 2, sufficient visualization; 3, adequate visualization; 4, excellent visualization. They were also asked to provide a qualitative assessment of the duration of contrast enhancement achieved, specifying whether it was too brief to enable an adequate assessment of the whole prosthesis and detection of endoleak (score 1) or whether its duration was adequate to enable the detection of endoleak (score 2). They rated the degree of intensity of vascular contrast enhancement achieved as 0, insufficient; 1, poor; 2, sufficient; 3, good; and 4, optimal.

Standard of reference 

Triple-phase CT acquisition (unenhanced, arterial and delayed phase images), assessed in consensus by two experienced vascular radiologists (A. R. C. and D. G., with 18 and 3 years of experience in body CT, respectively) not involved in US image analysis, who knew previous CT imaging findings, represented our standard of reference for both endoleak detection and exclusion. They established the presence/absence of endoleak according to a 3-point scale: 1, no leak; 2, uncertain; and 3, presence of a leak. Patients with a score of 2 were not considered in the evaluation.

They were asked to classify a detected leak according to its etiology as described by White et al.16, 17, 18 All endoleaks detected only on delayed phase were classified as low-flow leaks.7 If a CTA diagnosis of endotension was performed or if CTA was not able to classify the endoleak, the standard of reference was represented by a selective catheter angiography performed by a senior vascular radiologist (A. R. C.).

The size of each endoleak was categorized as small (≤3%), medium (>3% to <10%), or large (≥10%) by comparing the area of the endoleak (A(e)) with the maximum cross-sectional area of the aneurysm sac (A(a)) evaluated on axial images by using an electronic cursor (percentage size of endoleak = A(e)/A(a)).

Readers assessed changes in aneurysmal sac size (increment, stability or reduction) compared with previous CT studies by measuring the largest diameter of the aneurysm perpendicular to the aortic axis on the axial images using an electronic cursor. A change was recorded if it was ≥5 mm.

Statistical analysis 

Data were reported as mean ± standard error for continuous variables, whereas categoric and ordinal data were reported as frequencies and percentages. The qualitative assessment of duration and degree of intensity of contrast enhancement was used to establish the correct dose; the three sessions were also qualitatively compared in terms of degree of visualization of various parts of the endovascular stent graft (5-point scale). Differences among the three sessions in image quality score were evaluated by analysis of variance with generalized linear model.

Interobserver agreement for US image evaluation was assessed with intraclass correlation, rating reliability by comparing the variability of different ratings of the same subject with the total variation across all ratings and subjects.19

The diagnostic accuracy of each set was also estimated by calculating the area under the receiver operating characteristic curve (AUC), representing a combined measure of sensitivity and specificity. Differences in diagnostic tests can be evaluated by comparing AUCs,20, 21, 22 because they measure overall performance. We also evaluated differences in AUC values between readers in the image sets using U statistics, according to DeLong.23 Negative predictive values, sensitivity, and specificity to detect endoleaks for each set were calculated and compared using the McNemar test.24 Two-tailed values of P < .05 were considered significant. Statistical analyses were done using SAS 8.2 (SAS Institute, Cary, NC). When the CTA outcome was uncertain, the US examination result was not considered.

Back to Article Outline

Results 

Patients 

The study enrolled 84 consecutive patients (69 men, 15 women) with a mean age of 79.6 ± 5.2 years, (range, 62-89 years) and a mean body mass index (BMI) of 27.4 ± 3.5 kg/m2 (range, 22-34.2 kg/m2) who underwent endovascular repair of an unruptured infrarenal AAA. Devices used were 81 aortobiiliac stent grafts, consisting of 43 Talent (Medtronic AVE), 28 Excluder (WL Gore), 8 Zenith (Cook), 1Vanguard, and 1 AneuRx (Medtronic AVE); and 3 aortomonoiliac stent grafts (Talent, Medtronic, AVE). The mean follow-up after EVAR was 8.6 ± 5.4 months (range: 1-24 months). All patients completed the protocol, and no adverse events were recorded during CEUS or multidetector CT examinations.

Image analysis 

Gold standard 

Multislice CTA detected endoleaks in 40 of 84 patients (47.6%). None of the CTAs resulted in an uncertain diagnosis (score 2) on the presence or absence of endoleak, so all US examinations were evaluated.

Endoleaks were classified according to the size and etiology as summarized in Table I. In detail, two large endoleaks were not clearly classified by CTA (differential diagnosis between type II and type III endoleak). These two patients underwent selective conventional angiography that detected two type II endoleaks due to retrograde flow into the aneurysm sac through lumbar arteries. Five small type II endoleaks were detected only on delayed phase and were classified as low-flow leaks.

Table I. Classification of detected endoleaks
EndoleakEtiology
IIIIIIIVVTotal
IMALA
Small…617………23
Medium…57………12
Large2...21……5
Total211261……40

IMA, Inferior mesenteric artery; LA, lumbar artery.

An increase in size of the aneurysm sac associated with endoleak was observed in all type I (1.6 cm and 1.2 cm, respectively) and type III endoleaks (1.5 cm) and in two large (0.8 cm and 1 cm, respectively) and two medium type II endoleaks (0.5 cm and 0.8 cm, respectively); in the remaining 33 type II endoleaks, 16 aneurysm sacs were stable and 17 decreased. In all patients diagnosed as negative for endoleak at the standard of reference, the aneurysm sac always decreased or remained unchanged, without any complication. Therefore, no diagnosis of endotension was made.

Reading session 

No US examination was considered technically inadequate (no visualization of endovascular stent graft at all different levels, score 0) due to body habitus or collaboration of patient or other difficult scanning conditions. A score of at least of 2 (sufficient visualization) was obtained in all patients, except for three patients studied at the 1-month follow-up who underwent EVAR by using a low-permeability design Gore Excluder endoprosthesis. In these patients a poor visualization of all studied segments was registered due to significant artifacts with echo reflection.

The κ test analysis showed excellent interobserver agreement (κ analysis value ≥0.89) in all reading sessions for endoleak detection. Furthermore, there was no significant disagreement between readers in negative vs positive endoleak diagnosis and no second evaluation was needed to achieve consensus. Data for the two readers were pooled based on statistical results.

When considering true positive cases, readers detected 25 endoleaks (62.5%) in the CDUS imaging session A, 37 (92.5%) in the LDCE US session B and 39 (97.5%) in the HDCE US session C. In detail, all 17 medium and large leaks were correctly detected in the three sessions (Fig 1), whereas eight of 23 small leaks were correctly detected in the CDUS session A, 20 in the LDCE US session B, and 22 in the HDCE US session C. In detail, the small type II endoleak undetected at the 1-month follow-up in the LDCE and HDCE US sessions was in one patient who underwent EVAR with a low-permeability design Gore Excluder endoprosthesis, with significant periprosthetic artifacts with echo reflection. The other two leaks not detected with LDCE US imaging were small type II endoleaks from a stable aneurysm sac. All endoleaks detected were correctly classified on both LDCE and HDCE US imaging.

  • View full-size image.
  • Fig 1. 

    A 74-year-old man treated with endovascular aneurysm repair (6-month follow-up). a, A large endoleak was correctly detected on color duplex ultrasound imaging and (b) contrast-enhanced ultrasound images, and (c) was confirmed by standard of reference.

All five small low-flow endoleaks were correctly detected in US sessions B (LDCE) and C (HDCE), whereas none were detected on CDUS imaging (false-negative cases; Fig 2).

  • View full-size image.
  • Fig 2. 

    An 82-year-old man treated with endovascular aneurysm repair at 1-month follow-up. a, The baseline color duplex ultrasound image did not demonstrate any color duplex signal beyond the graft, with a consequent negative diagnosis for endoleak. b, The contrast-enhanced ultrasound image showed a small endoleak (arrows) at 150 seconds after contrast injection which was regarded as a low-flow leak. c and d, Standard of reference confirmed the presence of a small endoleak on the posterolateral side of the aneurysm, detected only on delayed phase axial CT image (low-flow leak) (arrows in panel d).

Twenty-eight (63.6%) of 44 patients devoid of endoleak according to multislice CTA were classified as such by CDUS imaging vs 35 (79.5%) by LDCE and 36 (81.8%) by HDCE US imaging; thus, nine and eight false-positive judgments were made with CEUS imaging. Five and four false positive endoleaks were detected by LDCE and HDCE US imaging, respectively, in the arterial phase (<30 seconds after contrast medium injection; Fig 3), whereas the last four false-positive judgments on both CEUS sessions were detected at >150 seconds after the contrast medium injection and were regarded as low-flow leaks. Two arterial-phase false-positive judgements were performed at the 1-month follow-up in two patients who underwent EVAR by using a low-permeability design Gore Excluder endoprosthesis. The presence of significant periprosthetic artifacts resulted in both readers providing an uncertain diagnosis (score 3). The diagnostic performance of CDUS and CEUS imaging is compared in Table II.

  • View full-size image.
  • Fig 3. 

    A 78-year-old woman treated with endovascular aneurysm repair at the 12-month follow-up. a, A contrast-enhanced ultrasound image shows a high attenuation area outside the graft (arrows) but within the aneurysm sac, and an endoleak was diagnosed. b, However, no endoleak was detected on axial arterial and (c) delayed-phase computed tomography images. d, An accurate evaluation of the baseline color duplex ultrasound image allows the recognition of a high attenuation of the thrombus outside the stent graft lumen (arrows) excluding the previous false-positive endoleak.

Table II. Diagnostic performance of color duplex and contrast-enhanced ultrasound imaging in endoleak detection using multislice computed tomography as a reference
Diagnostic parameterType of ultrasound imaging
CDLDCEHDCE
Negative predictive value, %65.192.1a97.3a
Sensitivity, %62.592.5a97.5a
Specificity, %63.679.5a81.8a
Accuracy, %63.185.7a89.3a
Az index0.7370.921a0.971a

CD, color duplex; LDCE, low-dose contrast-enhanced; HDCE, high-dose contrast-enhanced; Az index, area under the receiver operating characteristic curve.

aSignificantly higher than rate achieved with color duplex ultrasound imaging (P < .05).

Contrast enhancement with 1.2 mL of contrast medium (LDCE) significantly improved the endoprosthesis visualization score from 2.57 to 3 (P < .05) compared with baseline CDUS imaging. Doubling the contrast medium to 2.4 mL (HDCE) provided further significant improvement up to 3.39 (P < .05). The improvement included all parts of the endoprosthesis (Fig 4).

  • View full-size image.
  • Fig 4. 

    Qualitative evaluation of endoprosthesis visualization scores for color duplex (gray bars), and contrast-enhanced ultrasound imaging (CEUS) with 1.2 mL (diagonal-patterned bars) and 2.4 mL (dark gray). An, Anastomosis; Br, branch; Prox, proximal.

HDCE provided significantly longer contrast enhancement duration than LDCE (3 minutes 42 seconds ± 18 seconds vs 2 minutes 47 seconds ± 23 seconds p < .05), which ensured adequate enhancement for endoleak detection in 90.5% of cases vs 58.3% with the low dose. Contrast enhancement intensity was good to optimal in 89.3% of cases with HDCE vs 64.3% with LDCE (P < .05).

Back to Article Outline

Discussion 

The most reliable diagnostic alternative to CTA in post-EVAR life-long surveillance is still heavily debated. CDUS imaging is routinely used in vascular screenings because it is easy to perform, inexpensive, portable, safe, and widely available. In our experience and as reported in the literature, however, this technique performs poorly in endoleak detection, with high false-negative and false-positive results, principally due to echo reflection by the metallic portion of stent graft, presence of calcifications, meteorism, obesity, and slow endoleak flow, which does not allow distinction of color signals coming from vessel walls and surrounding tissue from those derived from corpuscular hematic components.

Conversely, this study shows CEUS imaging significantly improves the diagnostic performance of CDUS imaging in endoleak detection in patients with endovascular aortic stent grafts. Its sensitivity and negative predictive value are similar to multislice CTA (97.5% and 97.3%, respectively), and its specificity and accuracy are satisfactory (81.8% and 89.3%) but not ideal because the false-positive rate is nearly 10%. These findings support previous studies evaluating aortic stent grafts by CEUS imaging vs CTA, where sensitivity for endoleak detection was 50% to 100%, with many false-positive results.25, 26, 27, 28, 29

Napoli et al,30 however, reported 10 men with aneurysm enlargement and no evidence of endoleak during color-coded DUS imaging or CTA in whom CEUS scans performed with SonoVue detected endoleaks confirmed by conventional angiography. The authors suggest that CTA failure may have resulted from shorter imaging duration than with CEUS imaging; other published data have also confirmed14 that CEUS imaging seems to be more sensitive than CTA in diagnosing low-flow endoleaks.

Consequently, the four false-positive endoleaks detected in our experience at >150 seconds after contrast medium injection and regarded as low-flow leaks could be false-negative multidetector CT diagnoses, with a consequent potential increase of CEUS diagnostic accuracy.

On the other hand, three false-positive arterial endoleaks detected on US imaging with LDCE and two with HDCE were due to a baseline high attenuation of the thrombus not completely recognized on the baseline US scan (Fig 3). These uncorrected diagnoses could be avoided with an evaluation of baseline and contrast-enhanced US images performed side by side with a split screen mode, but this cannot be easily implemented in clinical practice. We would like to emphasize that to reduce false-positive diagnoses, an accurate baseline US examination before contrast medium injection must be performed, mainly to assess the morphology of the aneurysmal sac.

In our experience, the last two false-positive arterial endoleaks on CEUS imaging, as well as one false-negative undetected small endoleak, were found at the 1-month follow-up in three patients who underwent EVAR by using a low-permeability design Gore Excluder endoprosthesis. This innovative device, introduced in 2002, is composed of a durable, reinforced expanded polytetrafluoroethylene (ePTFE) graft, low permeability material layer, electropolished nitinol stent, and bonding film for stent to graft attachment. Its unique graft design reduces the potential for serous fluid movement through the graft wall, with consequent endotension. However, the ePTFE graft material produces significant artifacts with echo reflection at 1-month follow-up, explaining our erroneous diagnosis; these artifacts usually disappear at the 6-month follow-up.

Recent preliminary data31 suggest that CEUS imaging is more specific than CTA in endoleak classification thanks to longer duration of enhancement, lack of metallic artifacts, and angio-dynamic evaluation of the leak during the dynamic phase, as also demonstrated in our experience (Fig 5). CEUS advantages include minimal invasiveness, rapidity, and good tolerability: no adverse events were registered in our study.

  • View full-size image.
  • Fig 5. 

    A 71-year-old man treated with endovascular aneurysm repair at the 1-month follow-up. A large endoleak located in a posterolateral position was shown on the (a, c) three-dimensional and (b) axial computed tomography images, associated with (c) opacification of a lumbar artery, classified as a type II endoleak. However, the leak was also strictly adjacent to the prosthesis, with a consequent possible diagnosis of a concomitant type III endoleak. A classification of the endoleak was not clearly performed on the basis of the computed tomography images. d, An evaluation of dynamic contrast-enhanced ultrasound images demonstrated the back-filling of the excluded aneurysmal sac via lumbar artery, excluding a concomitant type III endoleak, as confirmed by digital subtraction angiography (e-g, arrows in f and asterisk in g).

On the other hand, CEUS imaging also has some limitations. Patient habitus (obesity) and bowel gas can interfere with imaging, and the patient must cooperate. The results of the US are operator-dependent, and obtaining quality images requires training and specific skills. Furthermore, CTA provides superior information related to graft anchoring and integrity, aneurysm morphologic changes, or visceral vessels patency (renal arteries).32

Therefore, CEUS should replace CTA at the 6-month follow-up and annually thereafter. In fact, in our opinion, the rationale of post-EVAR follow-up at 1-month and 12-month follow-up should be to detect endoleaks as well as to evaluate intraprocedural and periprocedural complications related to stent graft anchoring and visceral vessels patency and postprocedural complications related to stent graft migration and integrity, visceral vessels patency, and aneurysm morphologic changes, respectively. On the basis of this opinion, our suggested follow-up is based on CTA at 1 and 12 months after EVAR, with CEUS imaging performed at 6 months and annually thereafter, if no complications are detected.

The main strength of this study is that a large number of consecutive patients have been included. Our study also differs from others in literature for the attempt to define the optimal US contrast agent dose, which is not yet clearly defined for vascular examinations and stent-graft follow-up treatment.30, 33, 34 Our findings show that 2.4 mL is preferred to 1.2 mL because it provides significantly better results in intensity and duration of contrast enhancement and, consequently, in visualization than the low dose.

A potential limitation of our study could be the lack of a proper gold standard. However, as also reported in the literature, triple-phase CT acquisition—including unenhanced, 1-mm-slice arterial and delayed-phase images, with added clinical data (aneurysmal sac size change compared with previous CT exams)—seems to be the best gold standard in patient follow-up of patients who underwent EVAR. On the other hand, it could be interesting to check with selective digital angiography the patients that were negative by CTA but positive by CEUS imaging to confirm or deny the presence of an endoleak in order to define the real gold standard. However, in all patients in our study diagnosed as negative for endoleak at CTA, the aneurysm sac always decreased or remained unchanged, without any complication. These findings did not clinically justify the potential procedural risk of performing angiography.

Another potential limitation is based on the performed measurement of the size of the endoleak; in detail, a volumetric classification of size of the endoleaks could be more accurate than the comparison of the area with the maximum cross-sectional area of the aneurysm sac. Volume assessment is time-consuming, however, and requires advanced processing, dedicated equipment, and skilled operators; furthermore, as reported in the literature, the indication for treatment is based on the etiology of endoleaks and on the aneurysm diameter changes: as a matter of fact, an accurate classification of size of the endoleaks seems not to be clinically relevant.

In conclusion, CEUS is a fast, minimally invasive, reliable, and valid alternative to multislice CTA for endoleak detection in patients with endovascular aortic stent graft, and is superior to CDUS imaging. On the basis of our study, CEUS should be performed using a recommended contrast medium dose of 2.4 mL.

Back to Article Outline

Author contributions 


Conception and design: RI

Analysis and interpretation: AC, RB, DG, DP

Data collection: DG

Writing the article: RI

Critical revision of the article: RI, RB, AC

Final approval of the article: AC, RI

Statistical analysis: RI

Obtained funding: Not applicable

Overall responsibility: RI, MS

Back to Article Outline

References 

  1. Bush RL, Lumsden AB, Dodson TF, Salam AA, Weiss VJ, Smith RB, et al. Mid-term results after endovascular repair of abdominal aortic aneurysm. J Vasc Surg. 2001;33(2 suppl):S70–S76
  2. May J, White GH, Waugh R, Stephen MS, Chaufour X, Arulchelvam M, et al. Comparison of first- and second-generation prostheses for endoluminal repair of abdominal aortic aneurysms: a 6-year study with life table analysis. J Vasc Surg. 2000;32:124–129
  3. Zarins CK, White RA, Hodgson KJ, Schwarten D, Fogarty TJ. Endoleak as a predictor of outcome after endovascular aneurysm repair: AneuRx multicenter clinical trial. J Vasc Surg. 2000;32:90–107
  4. Golzarian J, Struyven J, Abada HT, Wery D, Dussaussois L, Madani A, et al. Endovascular aortic stent-grafts: transcatheter embolization of persistent perigraft leaks. Radiology. 1997;202:731–734
  5. Golzarian J, Dussaussois L, Abada HT, Gevenois PA, Van Gansbeke D, Ferreira J, et al. Helical CT of aorta after endoluminal stent-graft therapy: value of biphasic acquisition. AJR Am J Roentgenol. 1998;171:329–331
  6. Cuypers P, Buth J, Harris PL, Gevers E, Lahey R. Realistic expectations for patients with stent-graft treatment of abdominal aortic aneurysms (Results of a European multicentre registry). Eur J Vasc Endovasc Surg. 1999;17:507–516
  7. Gilling-Smith G, Brennan J, Harris P, Bakran A, Gould D, McWilliams R. Endotension after endovascular aneurysm repair: definition, classification, and strategies for surveillance and intervention. J Endovasc Surg. 1999;6:305–307
  8. Gorich J, Rilinger N, Sokiranski R, Orend KH, Ermis C, Krämer SC, et al. Leakages after endovascular repair or aortic aneurysms: classification based on findings at CT, angiography, and radiography. Radiology. 1999;213:767–772
  9. Makaroun MS, Deaton DH. Is proximal aortic neck dilatation after endovascular aneurysm exclusion a cause for concern?. J Vasc Surg. 2001;33(2 suppl):S39–S45
  10. Vallabhaneni SR, Harris PL. Lessons learnt from the EUROSTAR registry on endovascular repair of abdominal aneurysm repair. Eur J Radiol. 2001;39:34–41
  11. Brenner DJ, Hall EJ. Computer Tomography–An increasing Source of Radiation Exposure. N Engl J Med. 2007;357(22):2277–2284
  12. Raman KG, Missig-Carrol N, Richardson T, Muluk SC, Makaroun MS. Color-flow duplex ultrasound versus computed tomographic scan in the surveillance of endovascular aneurysm repair. J Vasc Surg. 2003;38:645–651
  13. Sato DT, Goff CD, Gregory RT, Walter BF, Gayle RG, Parent FN, et al. Endoleak after aortic stent graft repair: diagnosis by color duplex ultrasound scan versus computed tomography scan. J Vasc Surg. 1998;28:657–663
  14. Bendick PJ, Bove PG, Long GW, Zelenock GB, Brown OW, Shanley CJ. Efficacy of ultrasound scan contrast agents in the noninvasive follow-up of aortic stent grafts. J Vasc Surg. 2003;37:381–385
  15. Martegani A, Aiani L, Borghi C. The use of contrast-enhanced ultrasound in large vessels. Eur Radiol. 2004;14(suppl 8):73–86
  16. White GH, Yu W, May J, Chaufour X, Stephen MS. Endoleak as a complication of endoluminal grafting of abdominal aortic aneurysms: classification, incidence, diagnosis, and management. J Endovasc Surg. 1997;4:152–168
  17. White GH, May J, Waugh RC, Chaufour X, Yu W. Type III and type IV endoleak: toward a complete definition of blood flow in the sac after endoluminal AAA repair. J Endovasc Surg. 1998;5:305–309
  18. White GH, May J, Petrasek P, Waugh R, Stephen M, Harris J. Endotension: an explanation for continued AAA growth after successful endoluminal repair. J Endovasc Surg. 1999;6:308–315
  19. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420–428
  20. Obuchowski NA. Receiver operating characteristic curves and their use in radiology. Radiology. 2003;229:3–8
  21. Park SH, Goo JMG, Jo CH. Receiver operating characteristic (ROC) curve: practical review for radiologists. Korean J Radiol. 2004;5:11–18
  22. Metz CE. ROC methodology in radiologic imaging. Invest Radiol. 1986;21:720–733
  23. DeLong ER, DeLong D, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44:837–845
  24. Rosner B. Fundamentals of biostatistics. New York, NY: Duxbury; 1995;p. 420-6
  25. McWilliams RG, Martin JM, White D, Gould DA, Harris PL, Fear SC, et al. Use of contrast-enhanced ultrasound in follow-up after endovascular aortic aneurysm repair. J Vasc Interv Radiol. 1999;10:1107–1114
  26. McWilliams RG, Martin JM, White D, Gould DA, Rowlands PC, Haycox A, et al. Detection of endoleak with enhanced ultrasound imaging: comparison with biphasic computed tomography. J Endovasc Ther. 2002;9:170–179
  27. Giannoni MF, Palombo G, Sbarigia E, Speziale F, Zaccaria A, Fiorani P. Contrast-enhanced ultrasound imaging for aortic stent-graft surveillance. J Endovasc Ther. 2003;10:208–217
  28. Giannoni MF, Fanelli F, Citone M, Acconcia CM, Speziale F, Gossetti B. Contrast ultrasound imaging: the best method to detect type II endoleak during endovascular aneurysm repair follow-up. Interact Cardiovasc Thorac Surg. 2007;6:359–362
  29. Dill-Macky MJ, Wilson SR, Sternbach Y, Kachura J, Lindsay T. Detecting endoleaks in aortic endografts using contrast-enhanced ultrasonography. Am J Roentgenol. 2007;188:W262–W268
  30. Napoli V, Bargellini I, Sardella SG, Petruzzi P, Cioni R, Vignali C, et al. Abdominal aortic aneurysm: contrast-enhanced US for missed endoleaks after endoluminal repair. Radiology. 2004;233:217–225
  31. Carrafiello G, Laganà D, Recaldini C, Mangini M, Bertolotti E, Caronno R, et al. Comparison of contrast-enhanced ultrasound and computed tomography in classifying endoleaks after endovascular treatment of abdominal aorta aneurysms: preliminary experience. Cardiovasc Intervent Radiol. 2006;29:969–974
  32. Thompson MM, Boyle JR, Hartshorn T, Maltezos C, Nasim A, Sayers RD, et al. Comparison of computed tomography and duplex imaging in assessing aortic morphology following endovascular aneurysm repair. Br J Surg. 1998;85:346–350
  33. Henao EA, Hodge MD, Felkai DD, McCollum CH, Noon GP, Lin PH, et al. Contrast-enhanced duplex surveillance after endovascular abdominal aortic aneurysm repair: improved efficacy using a continuous infusion technique. J Vasc Surg. 2006;43:259–264
  34. Bargellini I, Napoli V, Petruzzi P, Cioni R, Vignali C, Sardella SG, et al. Type II lumbar endoleaks: hemodynamic differentiation by contrast-enhanced ultrasound scanning and influence on aneurysm enlargement after endovascular aneurysm repair. J Vasc Surg. 2005;41:10–18

 Competition of interest: none.

 CME article

PII: S0741-5214(08)01677-7

doi:10.1016/j.jvs.2008.10.008

Journal of Vascular Surgery
Volume 49, Issue 3 , Pages 552-560, March 2009

Article Tools

* Image Usage & Resolution