Prediction of Coronary Artery Disease Progression in Human from Numerically Determined Endothelial Shear Stress CL Feldman 1,2 , AU Coskun 3 , PH Stone 1,2 1 Brigham & Women's Hospital, Boston, MA, USA 2 Harvard Medical School, Boston, MA, USA 3 Northeastern University, Boston, MA, USA Abstract Using a technique in which intravascular ultrasound images are fused with two planes of angiography, blood flow is measured, and the Navier Stokes equations are solved in 3 dimensions, we have studied the remodeling characteristic of 55 human coronary arteries and the relationship of endothelial shear stress (ESS) to remodeling and plaque progression in 13 human arteries. Results indicate that the remodeling characteristics of 87% of coronaries with minimal luminal narrowing are constant along the length of the artery: 60% percent demonstrate compensatory remodeling; 19% exhibit under-remodeling (consistent with stable CAD) and 21% exhibit excessive remodeling (consistent with unstable syndromes). Serial studies show that plaque progresses almost exclusively in regions of low ESS (<12 dyne/cm 2 ). This suggests a new paradigm that focuses on segments with low ESS, treating those with excessive remodeling as being at risk for unstable syndromes and those with inadequate remodeling as being at risk for stable CAD. 1. Introduction Coronary artery disease (CAD) is a diffuse, inflammatory disease, resulting from incorporation of lipoproteins into the wall of the coronary artery. However, the important sequelae of CAD are focal, with evidence gathered over several decades suggesting that the sites of plaque accumulation are determined by patterns of local endothelial shear stress (ESS)[1,2]. Focal progression of CAD is related to both plaque progression and vascular remodeling (outward or inward remodeling of the external elastic membrane [EEM]). Glagov and colleagues[3] observed that, in the aggregate, the response of the EEM to increasing plaque burden is compensatory outward remodeling until plaque occupies approximately 40% of the dilated EEM. However, there is great individual variability in this “Glagov” remodeling. Outward remodeling has been associated with unstable plaque, acute coronary syndromes and myocardial infarction while inward remodeling has been associated with stable CAD[4-7]. Using techniques we have previously described and have subsequently refined[8], we have studied both the remodeling characteristic of human coronary arteries and the relationship of ESS to plaque progression in a subset of those arteries during a follow-up period of approximately 8 months. Our objective is to develop tools to determine which minor lesions will progress and, of those that do progress, which will evolve to stable CAD and which will result in unstable coronary syndromes. 2. Methods Our methods of intracoronary profiling have been previously described[8-10]. In brief, the 3-D anatomy of the artery was reconstructed from IVUS images and 2 planes of coronary angiography. IVUS (Boston Scientific, Natick, MA) was performed with controlled pullback at 0.5 mm/sec. The ECG was recorded on the IVUS images. The arterial lumen and outer vessel wall were reconstructed from digitized and segmented end- diastolic IVUS frames[9]. The physical 3-D path of the IVUS transducer during pullback was determined using the corresponding biplane angiographic projections. The 3-D reconstructed catheter core served as the stem on which to rebuild the 3-D geometry. The 3-D position of each ECG-gated IVUS frame was determined from the reconstructed trajectory of catheter pullback and pullback speed[11]. The rotation of the frame was determined using computational geometry[9]. Each frame was aligned perpendicular to the catheter core. The boundary points of each frame were connected by spline curves to rebuild the luminal geometry in 3-D space. The 3-D geometry of the outer vessel wall (area within the EEM) was recreated in a manner similar to that described for the lumen geometry. The 3-D geometry of the plaque (plaque plus media thickness) was taken as the difference between the outer vessel wall and the lumen[13]. Figure 1 illustrates the reconstructed artery. 0276-6547/05 $20.00 © 2005 IEEE 105 Computers in Cardiology 2005;32:105-108.