A methodology for assessing local bifurcated blood vessel shape variations

Abstract

The analysis of the progression of cardiovascular diseases is an active area of ongoing research. This paper develops an image registration-based methodology to quantify the patient-specific local blood vessel shape variations that occur in the radial direction (i.e. expansion or shrinkage) over an imaging follow-up period, and an example is presented as proof of principle. The methodology can be used for complex vessels with bifurcations, and it is able to identify and address vessel deformations if changes in tortuosity or longitudinal direction are small. The methodology consists of (a) overlapping the baseline and follow-up vessel surfaces by matching the lumen centerline, (b) dividing the region of interest into slices perpendicular to the centerline and centering each slice, and (c) dividing each centered slice into sectors. The local approach consists of analyzing a representative point in each sector of each slice (i.e. each patch). In this paper the algorithm is applied to a patient-specific abdominal aortic aneurysm (AAA) as a proof of principle of the method. Six patient-specific image reconstructions from a single subject followed for 28 months are analyzed in pairs, yielding five time spans to which the algorithm was applied. The algorithm was able to quantify the AAA radial growth. The average AAA radial growths for the five case studies are-2.13 mm, 3.43 mm,-0.25 mm, 1.41 mm, and 0.84 mm, whereas the maximum local growths are 4.76±0.15 mm, 9.30±1.13 mm, 2.08±0.05 mm, 4.10±0.14 mm, and 4.16±0.45 mm. The tolerance of the geometric local measurements is related to the matching processes (i.e. overlapping the geometries and centering each slice) because of the vessel deformation that took place over time. Thus, this methodology has been used to quantify the average AAA growth and the maximum local AAA growth (± the tolerance) as metrics of the vessel's radial growth.