Estimating the error of CT-based measurements of aortic lumen volume used in endovascular planning

Abstract

IntroductionPlanning of endovascular sealing of abdominal aortic aneurysms requires measurement of the aortic lumen volume. The aim of this study was to investigate mathematically the effect of intra- and inter-observer variability error, as well as cardiac cycle-related variability, on these measurements. MethodsMean (±2SD) intra- and inter-observer error in lumen measurements and mean (+2SD) cardiac cycle-related variability were obtained from published literature and added to the measurement of the flow lumen volume of a 57 mm abdominal aortic aneurysm to calculate average and extreme error possibilities. ResultsThe aneurysm volume was measured at 165 ml. The calculated possible mean measurement error due to cardiac cycle variation, intra- and inter-observer variability was +11.0%, resulting in a potential measurement of 183.1 ml. The calculated extreme errors were +24.3% (if 2SD of all errors were added to the mean) and +3.5% (if 2SD of all errors, except cardiac cycle, were subtracted from the mean), resulting in potential measurements of 170.8 ml and 205.1 ml, respectively. When considering the errors combined, the proportion of patients who may have volume measurement errors of up to ±2.5 ml, ±2.6 to ±5.0 ml and ±5.1 to ±7.5 ml were 18%, 17% and 15%, respectively. ConclusionMeasurement of CT-based aortic lumen volumes in abdominal aortic aneurysms is imprecise. This has practical implications for the planning and the performance of complex endovascular therapies.