McWilliams 2002.

Clinical features and settings	People who received endovascular repair of an unruptured infrarenal AAA. Type of stents received: not reported (endografts were all bifurcated with either a modular or 1‐piece design except for 1 aortic tube device). Aneurysm diameter: not reported. Setting: department of radiology.
Participants	53 participants; 44 men; mean age 70 years; mean height of group: 171 cm; range: 150‐183 cm); mean weight: 77 kg; range: 47‐107 kg). Comorbidities: not reported. Geography: UK.
Study design	Prospectively enrolled participants. "All patients seen at these follow‐up intervals were asked to participate unless there was a documented contraindication to the use of Levovist (e.g., galactosemia and a known or suspected right‐to‐left cardiac shunt)."
Target condition and reference standard(s)	Target condition: endoleak. Definition of endoleak: presence of intrasac flow outside stent‐graft; characterized by its relationship to endograft, aneurysm wall, and aortic side branches and categorized using White/May classification. Endoleak (absolute n): 7. Prevalence of endoleak: 9.2% (7/20). Reference standard: contrast‐enhanced biphasic (arterial and delayed) CT. Image acquisition: not reported. thickness: scanning parameters identical for both phases: 5 mm collimation, 1‐s tube rotation, pitch 2, reconstructions every 5 mm. Type of CT scanner: HiSpeed Advantage (IGE Medical Systems, Slough, UK). Use of contrast: yes (type not reported).
Index and comparator tests	Index test: CDUS. Image production: not reported. Type of US: 3.5‐MHz probe on either a Dyna‐View SSD‐1700 or a ProSound 5500 (Aloka Co Ltd, Tokyo, Japan). Use of contrast: yes, Levovist (Schering Health Care, Felbridge, UK). Operator: 1 experienced vascular sonographer.
Follow‐up	2 participants excluded because radiology staff failed to follow protocol during 5 imaging sessions.
Notes	Where there was evidence of endoleak on either the US or enhanced CT and sac diameter had increased or remained static ≥ 6 months, then protocol allowed for selective arteriography. Mean number of follow‐up evaluations was 1.8 per participant; maximum number was 4 over a follow‐up period of 1‐36 months. Mean and median intervals since endovascular repair at the time of imaging were 11 and 6 months, respectively. Endoleak detected in 20 (21%) of the 96 CT examinations; majority were isolated type II (18, 90%). The other 2 cases had graft‐related endoleaks from 1 iliac limb dislocation and 1 stump dislocation. In 7 of the CT examinations, endoleak detected on the delayed phase CT only.
Table of Methodological Quality
Item	Authors' judgement	Description
Representative spectrum? All tests	Yes	Participants represents average patients who after receiving EVAR are exposed to endoleak surveillance.
Acceptable reference standard? All tests	Yes	Reference standard was biphasic enhanced CT.
Acceptable delay between tests? All tests	Yes	"Biphasic enhanced CT was performed on the same day using the same protocol and imager."
Partial verification avoided? All tests	Yes	All study participants accounted for and results of reference standard reported for all.
Differential verification avoided? All tests	Yes	All participants who received index test subjected to same reference standard.
Incorporation avoided? All tests	Yes	Index test not part of reference standard.
Reference standard results blinded? All tests	Yes	"Either of two radiologists (R.M., D.A.G.), who were blinded to the ultrasound results, recorded all the CT studies."
Index test results blinded? All tests	Yes	US scan apparently performed before CT scan.
Relevant clinical information? All tests	Yes	Yes.
Uninterpretable results reported? All tests	Yes	"Angiography in the patient with a 'definite' endoleak on the CT study confirmed a lumbar endoleak, which was treated. One patient with a 'probable' endoleak on the CT study showed lumbar vessels perfusing the sac margin at 3 levels but no endoleak. The other patient with a 'probable' lumbar endoleak on biphasic CT had increasing sac diameter, but no endoleak was seen on the arteriogram. Comparing each of the 4 ultrasound techniques with biphasic CT in the detection of endoleak, the number of nondiagnostic studies (flow detection too poor to allow diagnosis) was highest (n 12) in the unenhanced ultrasound group and lowest (n 4) with the enhanced power Doppler test." "One patient had nondiagnostic ultrasound examinations with all 4 test modalities due to bowel gas; the CT showed iliac limb dislocation."
Withdrawals explained? All tests	Yes	2 participants excluded because radiology staff failed to follow protocol during 5 imaging sessions.