Variability in measurement of abdominal aortic aneurysms. Abdominal Aortic Aneurysm Detection and Management Veterans Administration Cooperative Study Group.

Abstract

The purpose of this study was to report interobserver and intraobserver variability of computed tomography (CT) measurements of abdominal aortic aneurysm (AAA) diameter and agreement between CT and ultrasonography observed in the course of a large, multicenter, randomized trial on the management of small AAAs. CT measurements of AAA diameter from participating centers were compared with measurements made from the same scan by a central laboratory. Blinded central remeasurement of a randomly selected subset of these CT scans was used to assess intraobserver variability. Agreement between AAA measurements by CT and ultrasonography done within 30 days of each other was also assessed. For interobserver pairs of local and central CT measurements of AAA diameter (n = 806), the difference was 0.2 cm or less in 65% of pairs, but 17% differed by at least 0.5 cm. For intraobserver pairs of central CT remeasurements (n = 70), 90% differed by 0.2 cm or less, 70% were within 0.1 cm, and only one differed by 0.5 cm. Of 258 ultrasound-measured and central CT pairs, the difference was 0.2 cm or less in 44% and at least 0.5 cm in 33%. Ultrasound measurements were smaller than central CT measurements by an average of 0.27 cm (p < 0.0001). Local CT and ultrasound measurements showed a marked preference for recording by half centimeter. A high degree of precision is possible in CT measurement of AAA diameter, but this precision may not be obtained in practice because of differences in measurement techniques. Differences between imaging modalities increase variability further. Variations in AAA measurement of 0.5 cm or more are not uncommon, and this should be taken into account in management decisions. Efforts to reduce variation in measurement are warranted and might include (1) seeking agreement between surgeons and radiologists on a precise definition of AAA diameter, (2) limiting the number of radiologists who measure AAAs, and (3) use of calipers and magnifying glass for CT measurements.