201. Predicting Plaque Progression Using Patient-Specific Fluid-Structure-Interaction Models Based on IVUS and OCT Images with Follow-Up.
- Author
-
Xiaoya Guo, Dalin Tang, Molony, David, Chun Yang, Samady, Habib, Jie Zheng, Mintz, Gary S., Akiko Maehara, Jian Zhu, Genshan Ma, Mitsuaki Matsumura, and Giddens, Don P.
- Subjects
ATHEROSCLEROTIC plaque ,INTRAVASCULAR ultrasonography ,OPTICAL coherence tomography ,FLUID-structure interaction ,DIAGNOSTIC imaging - Abstract
Atherosclerotic plaque progression is generally considered to be closely associated with morphological and mechanical factors. Plaque morphological information on intravascular ultrasound (IVUS) and optical coherence tomography (OCT) images could complement each other and provide for more accurate plaque morphology. Fluid-structure interaction (FSI) models combining IVUS and OCT were constructed to obtain accurate plaque stress/strain and flow shear stress data for analysis. Accuracy and completeness of imaging and advanced modeling lead to accurate plaque progression predictions. In vivo IVUS and OCT coronary plaque data at baseline and follow-up were acquired from left circumflex coronary and right coronary artery of one patient with patient's consent obtained. Co-registration and segmentation of baseline and follow-up IVUS and OCT images were performed by experts. Baseline and follow-up 3D FSI models with cyclic bending based on merged IVUS and OCT data were constructed to obtain plaque stress/strain and flow shear stress data for plaque progression prediction. Nine factors (6 morphological factors and 3 mechanical factors) including average cap thickness, lipid area, calcification area, lumen area, plaque area, plaque burden, wall shear stress (WSS), plaque wall stress (PWS) and plaque wall strain (PWSn) were selected for each slice. Plaque area increase (PAI) and plaque burden increase (PBI) were chosen to measure plaque progression and serve as the target variables for prediction. All possible combinations of nine factors were fed to a generalized linear mixed model for PAI and PBI prediction and quantification of their prediction accuracies. In this paper, prediction accuracy was defined as the sum of sensitivity and specificity. The optimized predictor combining 9 factors gave the best prediction for PAI with accuracy=1.7087 (sensitivity: 0.8679; specificity: 0.8408). PWSn was the best single-factor predictor for PAI with accuracy=1.5918 (sensitivity: 0.7143; specificity 0.8776). A combination of average cap thickness, calcification area, plaque area, PWS and PWSn gave the best prediction for PBI with accuracy=1.8698 (sensitivity: 0.8892; specificity: 0.9806). PWSn was also the best single-factor predictor for PBI with accuracy=1.8461 (sensitivity: 0.8784; specificity 0.9677). Although WSS was commonly accepted as an important factor for plaque progression, it showed relatively poor ability for prediction of plaque progression in any measure (accuracy, sensitivity, specificity of PAI: 1.0607, 0.0893, 0.9714; PBI: 1.5431/0.6811/0.9032). Combining morphological and mechanical risk factors may lead to more accurate progression prediction, compared to the predictions using single factor. PWSn is better than WSS for plaque progression using single factor. IVUS+OCT formed basis for accurate data for morphological and mechanical factors. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF