Background: Ischemic stroke is a major cause of death worldwide. Atherosclerosis in the carotid arteries is an established predictor of these events. Ideally, patient-specific prevention plans can be developed that target advanced plaque development prior to events. Morphometry of advanced calcified plaque phenotypes can be assessed through Computed Tomography Angiography (CTA) but predicting progression on morphometry alone is insufficient as atherosclerosis is nonlinear and heterogeneous. Therefore, it was hypothesized that biomechanical triggers stimulate advanced plaque growth. The biomechanical response can be observed through structural stress and strain in the vessel wall as a response to plaque composition, blood pressure and tethering. So, the research aim of this study is to develop a framework that allows for local and global assessment of structural biomechanical stimuli and morphometrical changes in atherosclerotic carotid arteries

Methods Nine CTA scans paired at baseline and follow-up were selected from the Plaque At RISK study. The carotid bifurcation at the cervical spine was segmented by two independent observers using QAngioCT (Medis Medical Imaging, Leiden, The Netherlands). Contour data was reconstructed into a 3D geometry using a two-phase developed method in which 3D surface was computed first, and the second phase converted this to volumetric parts. Finite Element (FE) models were developed for baseline geometries where Neo-Hookean for calcified and Holzapfel-Gasser-Ogden for non-calcified materials were assigned.
For both timepoints, local morphometry was defined in wall thickness (WT), principal curvatures and calcium localization. Contour maps allowed the local association analysis between biomechanics and morphometry. Global plaque progression was computed using the morphometrical parameters and overall plaque burden (PB).


Results: This pipeline was successfully run for nine different carotid arteries, which proved overall robustness. The Dice similarity index computed an average segmentation observer similarity of 0.80 (St. Dev. 0.06) and 0.87 (St. Dev. 0.07) for surface reconstruction. Throughout reconstruction, the FE-modeling set the requirements for reconstruction outcome thus the focus was laid on connecting these phases. No patient data was made available during this study, so standardized systolic blood pressure resulted in an average maximum stress of 269.07 kPa and 0.14 strain.
Morphometrical analysis detected diseased WT in seven out of nine cases at baseline and all at follow-up. Local principal curvatures uncovered a relation with diseased thickening, indicating its success in detection of irregularities on a surface. Calcified tissue was found in all cases but one at baseline. The average morphometrical change increased 2.71 mm (St. Dev. 7.86 mm) in maximum WT, 0.69 % (St. Dev. 5.75 %) for maximum PB and 12.91 mm³ (St. Dev. 12.50 mm³) for calcium. These preliminary results highlighted the importance of multicomponent morphometry analysis. Moreover, correlations between calcium growth and stress (R² = 0.33) and WT increase (R² = 0.67) indicate that future studies should focus on comprehending atherosclerotic pathways involved in calcified plaque formation.

Conclusion: This developed method has laid the groundwork for future research and exposed important relations between analysis methods. The close dependency between reconstruction and FE-modeling, and anatomy and biomechanics emphasize that there is still a lot to discover.