NOVEL IMAGING TOOLS AND ARTIFICIAL INTELLIGENCE FOR VULNERABLE PLAQUE CHARACTERIZATION

Cardiovascular diseases remain the leading cause of death globally, with coronary artery disease (CAD) being the primary contributor. The identification of high-risk plaques in coronary arteries, particularly non-flow-limiting lesions, is vital for improving risk stratification and reducing adverse cardiac events. This thesis investigates the application of novel imaging tools and artificial intelligence (AI) in evaluating vulnerable plaque characteristics, focusing on wall shear stress (WSS) as a key haemodynamic parameter. WSS plays a crucial role in atherosclerosis progression, plaque destabilization, and rupture. Through computational fluid dynamics (CFD) and three-dimensional quantitative coronary angiography (3D-QCA), this research explores the potential of WSS metrics in predicting lesion outcomes. The first study validated the CAAS Workstation WSS prototype against traditional methods, showing high reproducibility and accuracy in calculating WSS and identifying lesions with unfavourable haemodynamic environments. This prototype allows for WSS assessment in under five minutes, enabling its potential use in clinical settings. The second study evaluated 352 patients with borderline non-flow-limiting lesions and found that both area stenosis (AS) and WSS metrics improved risk stratification beyond baseline demographics, providing more precise prognostic insights. While WSS variables alone did not add predictive value beyond AS in multivariable analysis, patients with elevated WSS combined with high AS were identified as a high-risk group for adverse events in Kaplan-Meier curves. Additionally, the third study implemented AI and deep learning (DL) to predict WSS in stenosed coronary arteries. Using CFD-derived datasets and synthetic vessel models, the DL model demonstrated promising accuracy and speed in predicting WSS distribution, although it showed limitations in predicting shear stress in tortuous segments and in vessels with multiple stenoses. This advancement could significantly reduce the time required for CFD analysis, making WSS assessment more feasible for larger patient populations. In conclusion, this research underscores the importance of WSS in assessing plaque vulnerability and stratifying cardiovascular risk. The development of fast, reliable software, combined with AI advancements, holds potential for broader clinical application in identifying high-risk patients and improving outcomes in CAD management. Further studies are required to validate these tools and expand their use across diverse clinical populations.