In Vivo Intravascular Ultrasound-based 3D Thin-walled Model And Fluid-structure Interaction Model For Human Coronary Plaque Progression Study

Cardiovascular disease(CVD)is the leading cause of death in the world.Considerable research has been done linking various risk factors to plaque progression and rupture which often lead to drastic clinical events such as heart attack and stroke.There has been evidence indicating that mechanical stress and strain may play an important role in plaque progression.Researchers have proposed various image-based computational biomechanical models to understand the mechanisms governing plaque progression and rupture and try to provide more useful information for clinical screen strategy.Due to the complexity of plaque structure and composition,3D fluid-structure interaction(FSI)plaque model construction is extremely time-consuming and making 3D models is near impossible for clinical implementations.2D structure-only model is easy to make,but its stress/strain predictions are not good enough to serve as approximation to 3D solutions.In this study,an in vivo IVUS based 3D Thin-Wall(TW)model was developed to approximate 3D FSI model for plaque progression investigation and potential clinical implementations.Eight 3D FSI models and 385 3D thin-walled models based on IVUS image and X ray data at baseline(T1)and follow-up(T2)were constructed in this thesis using existing data to extract plaque wall stress(PWS),strain(PWSn),and wall thickness(WT)values for analysis.Results on matched lumen nodes from TW and FSI models were compared to investigate their differences.For progression analysis,slices at baseline(T1)and follow up(T2)were paired.Plaque progression was measured by wall thickness increase(WTI)defined as WT at T2 – WT at T1.Correlation results using FSI models and TW models were compared.Moreover,impact of circumference shrinkage,pre-stretching and bending on results from FSI and TW models was investigated.Results from the 385 TW models and 8 FSI models for the 4 patients indicated that the differences of mean stress and strain values were less than 10%(stress error 9.1%;strain error 8.4%).Plaque progression correlation results from TW models were consistent with that from the FSI models(positive/negative correlation agreement rate 87.5%).In terms of time cost,TW models shortened the 10-15 days of FSI model construction and solution time per patient to several hours under the premise of ensuring the precision,which would make it possible to implement biomechanical models for realistic clinic applications.Large scale studies are needed to further investigate the feasibility of using TW models as approximation for FSI models.