Laura Ellwein

• 
  Image-based 3D quantification and reconstruction of coronary artery morphology in the context of stenting as treatment of cardiovascular disease
  Laura Ellwein

  One in six adults in the US have some form of coronary artery disease, characterized in particular by accumulation of atherosclerotic plaque. Though stenting is the most common treatment technique, it often leads to restenosis and thrombus formation. Computational modeling of human arteries from patient-specific image-based data offers a noninvasive way to investigate geometry, hemodynamics, and vascular disease corresponding with effects of stenting.

  
  

  Improved strategies for stent-based patient-specific treatment of atherosclerotic lesions at coronary bifurcations require a greater understanding of normal coronary vessel morphology. We developed a method to quantify morphology in the left coronary artery for eventual use in bifurcating stent design. Computational models of the left main coronary were created from computed tomography (CT) images of 54 patients using ITK-Snap. Metrics assessed using Visualization Toolkit-based software and MATLAB included cross?sectional area, length, eccentricity, taper, curvature, branching law parameters, and bifurcation angles. Traditional statistical analysis using parametric tests for comparing and correlating means revealed significant differences both within and between bifurcations for most metrics.

  
  

  Image-based computational models for quantifying hemodynamic indices in stented coronary arteries often employ biplane angiography and intravascular ultrasound for 3D reconstruction, but recent advances in optical coherence tomography (OCT) suggest more precise coronary artery reconstruction may be possible. We developed a patient-specific coronary artery reconstruction method that combines OCT, an intravascular imaging modality, with techniques for imaging wire pathway reconstruction adopted from graph theory. The pathway of the imaging wire was determined with a shortest path algorithm assuming minimum bending energy, and OCT images were registered orthogonal to the pathway with appropriate rotational orientation. Segments from both OCT in the stented region and CT upstream and downstream were imported into computational fluid dynamics software to quantify indices of wall shear stress (WSS). WSS results are presented using the method applied to imaging data of a left circumflex coronary artery acquired immediately post-stenting and after a 6-month follow-up period.

  
  

  Findings from computational modeling studies using patient-specific imaging data may ultimately enhance our knowledge of both healthy coronary arteries and of harmful hemodynamic indices induced by stenting and could be leveraged for future stent design.