Peak cap stress calculations in coronary atherosclerotic plaques with an incomplete necrotic core geometry

Annette M. Kok, Lambert Speelman, Renu Virmani, Antonius F. W. van der Steen, Frank J. H. Gijsen, Jolanda J. Wentzel

Stress calculations in atherosclerotic coronary vulnerable plaques can aid in predicting coronary cap rupture. In vivo plaque geometry and composition of coronary arteries can merely be obtained via intravascular imaging. Only optical driven imaging techniques have sufficient resolution to visualize the fibrous cap, but due to limited penetration depth deeper components such as the backside of the necrotic core (NC) are generally not visible. The goal of this study was to investigate whether peak cap stresses can be approximated by reconstructing the backside of the NC. Manual segmentations of coronary histological cross-sections served as a geometrical ground truth and were obtained from seven patients resulting in 73 NCs. Next, the backside was removed and reconstructed according to an estimation of the relative necrotic core thickness (rNCt). The rNCt was estimated at three locations along the NC angle and based on either group averaged parameters or plaque specific parameters. Stress calculations were performed in both the ground truth geometry and the reconstructed geometries and compared. Good geometrical agreement was found between the ground truth NC and the reconstructed NCs, based on group averaged rNCt estimation and plaque specific rNCt estimation, measuring the NC area difference (25.1 % IQR 14.0-41.3 % and 17.9 % IQR 9.81-32.7 %) and similarity index (0.85 IQR 0.77-0.90 and 0.88 IQR 0.79-0.91). The peak cap stresses obtained with both reconstruction methods showed a high correlation with respect to the ground truth, r(2) = 0.91 and r(2) = 0.95, respectively. For high stress plaques, the peak cap stress difference with respect to the ground truth significantly improved for the NC reconstruction based plaque specific features (6 %) compared to the reconstruction group averaged based (16 %). In conclusion, good geometry and stress agreement was observed between the ground truth NC geometry and the reconstructed geometries. Although group averaged rNCt estimation seemed to be sufficient for the NC reconstruction and stress calculations, including plaque specific data further improved stress predictions, especially for higher stresses.