[home] [Personal Program] [Help]

---

Session: Vascular III
Session starts: Friday 27 January, 14:00
Presentation starts: 14:30
Room: Room 559

---

Esther Maas (Eindhoven University of Technology)
Arjet Nievergeld (Eindhoven University of Technology)
Judith Fonken (Eindhoven University of Technology)
Mirunalini Thirugnanasambandam (Eindhoven University of Technology)
Marc van Sambeek (Catharina Hospital Eindhoven)
Richard Lopata (Eindhoven University of Technology)

Abstract:
An abdominal aortic aneurysm (AAA), a local dilatation of the aorta, is a progressive disease. Between patients, there is a variety in growth rates and diameter at which the AAA ruptures [1-2]. In this study, these differences between patients were studied in a longitudinal study by analyzing 3D geometrical and mechanical parameters of AAAs using time-resolved 3D ultrasound (3D+t US). The mutual correlations between parameters, and their relationships with diameter and growth were studied. 3D+t US AAA images were acquired for 178 patients, who all had >3 follow-up diameter measurements with 2D US (d2D) from which the growth of their AAA was obtained. An automated analysis tool was developed to create a segmentation of the AAA wall for all time frames. From this, the maximal diameter (d3D) perpendicular to the centerline, volume (V), compliance (C) and distensibility (D) were determined. First, the geometry parameters were verified with computed tomography (CT) for a subset of patients (N=17), showing good correspondence with a median similarity index of 0.88. Correspondence between d3D and CT-based diameters is high (R2 = 0.90), although the d3D values are smaller (difference IQR [0.23 - 0.66] cm), a range similar to the differences for clinically used d2D values. Next, relationships between parameters were explored for all patients with Spearman correlations, and the parameters’ predictive value for growth were determined. A decrease in D with increasing blood pressure (p<0.0001) and diameter (p<0.05) was observed. Growth is related to d2D, d3D, V and C (p <0.005). Stepwise linear regression showed that C is a better predictor for growth than diameter (RMSE 1.70 and 1.78, respectively), despite diameter being a more established growth predictor [1,3]. In conclusion, an automated tool to determine geometrical and mechanical parameters of AAAs from 3D+t US has been developed, validated, and applied to a large group of patients. The AAA growth was found to be correlated to d2D, d3D, V and C, with C having the best predictive value for growth in a linear model. These findings contribute to the understanding of differences in AAA progression between patients, which is a step towards more patient-specific treatment.