Automated software supported versus manual aorto-iliac diameter measurements in CT angiography of patients with abdominal aortic aneurysms: Assessment of inter- and intraobserver variation

Abstract

Background: Open surgical or endovascular abdominal aortic aneurysm (AAA) relies on precise preprocedual imaging. Purpose of this study was to assess inter- and intraobserver variation of software-supported automated and manual multi row detector CT angiography (MDCTA) in aortoiliac diameter measurements before AAA repair. Patients and methods: Thirty original MDCTA data sets (4 times 2mm collimation) of patients scheduled for endovascular AAA repair were studied on a dedicated software capable of creating two-dimensional reformatted planes orthogonal to the aortoiliac center-line. Measurements were performed twice with a four-week interval between readings. Data were analysed by two blinded readers at random order. Two different measurement methods were performed: reader-assisted freehand wall-to-wall measurement and semi-automatic measurement. Results: Aortoiliac diameters were significantly underestimated by the semi-automatic method as compared to reader-assisted measurements (p < 0.0031). Intraobserver variability of AAA diameter calculation was not significant (p > 0.15) for reader-assisted measurements except for the diameter of the left common iliac artery in reader 2 (p = 0.0045) and it was not significant (p > 0.14) using the semi-automatic method. Interobserver variability was not significant for AAA diameter measurements using the reader-assisted method and for proximal neck analysis with the semi-automatic method (p > 0.27). Relevant interobserver variation was observed for semi-automatic measurement of maximum AAA (p = 0.0007) and iliac artery diameters (p = 0.024). Conclusions: Dedicated MDCTA software provides a useful tool to minimize aortoiliac diameter measurement variation and to improve imaging precision before AAA repair. For reliable AAA diameter analysis the reader-assisted freehand measurement method is recommended to be applied to a set of reformatted CT data as provided by the software used in this study.