Treatment of abdominal aortic aneurysm (AAA) is indicated if risk for rupture exceeds the risk for aortic repair. Estimation of the individual risk for rupture in AAA is therefore essential. The diameter of AAA is known as an independent risk factor for rupture and therefore the base of indication for surgical or endovascular therapy. For more sensitive patient selection, other morphological or hemodynamic predictors such as volume or peak wall stress have to be evaluated.
The purpose of this study was to analyze the reproducibility of diameter measurement, volume estimation and peak wall stress calculation in AAA by finite element analysis.
Computed tomography angiography (CTA) scans of 10 patients with AAA and 4 volunteers with healthy infrarenal aortas were analyzed by three independent investigators. A semiautomatic reconstruction using two- and three-dimensional deformable (active) contour models was used to segment vascular bodies from CTA data. Centreline calculated maximal diameter and volume measurements, as extracted from the reconstructed abdominal aorta, as well as peak wall stress, as predicted by three-dimensional non-linear finite element models, were analyzed. Specifically, aortic wall and thrombus tissue were captured by isotropic, non-linear and finite strain constitutive models. Likewise, mean arterial pressure was applied at the luminal surface, the vessels were fixed at the renal arteries and the aortic bifurcation and no contact with surrounding organs was considered.
Inter- and intra-observer variabilities for diameter, volume and peak wall stress measurements were assessed by calculating the coefficient of variation
(CV=SD*100/mean in %) of the five fold determinations. The methodological variation was expressed as deviation of diameter (mm), volume (ml) and peak wall stress (kPa) amongst the three observers.
Reproducibility measurements in healthy vessels of aortic diameters between 16.1mm to 16.6mm varied from CV=2.5% to CV=4.9%. Abdominal aortic volumes of 14ml to 15ml were measured in the healthy cohort with a reproducibility of CV=5.8% to CV=11.5%. Peak wall stress varied between 53 kPa and 55 kPa, where CV ranged from 3-13%. Inter-observer variation was <10% for diameter, volume and peak stress in healthy volunteers.
Aortic diameter in three AAAs was measured to 58.9 mm; 54.6 mm; and 71.2 mm respectively. The coefficient of variation showed high agreement with values less than 5%. AAA volume varied between 130 ml and 300 ml (CV < 10%) and Peak wall stress was predicted between 172 kPa and 296 kPa (CV <10%).
Variability between the 3 observers in AAA measurements was 0.7 mm – 6.0 mm for diameter, 11 – 28 ml for volume and 4-27 kPa for peak wall stress, respectively.
Volume and diameter measurements based on geometrical models reconstructed from CTA scans showed quit good reproducibility for serial measurements in normal and degenerative arteries. Peak wall stress predictions exhibited high accordance between different observers, and in serial measurements within one observer. Volume and peak wall stress analysis could be an additionally module for assessment of individual rupture risk in AAA in the future, which however needs to be validated by additional studies.