Hemodynamic loads are highly relevant to the initiation and progression of a number of cardio/cerebrovascular diseases. Quantification of hydrodynamic forces (e.g., wall shear stress, pressure gradients) induced by blood flow is significant in the clinical diagnosis of various vascular pathologies. Computational modeling based on physical laws and physiological knowledge becomes an important tool for accurately quantifying these biomechanical features and understanding their connection to disease progression. However, most of the applications are too ill-defined to rely solely on the deterministic computational model due to the high physiological complexity and inter-patient variations.
On the other hand, medical measurement techniques such as MRI, CT, and TCD have been significantly advanced in recent years. Thanks to the continuous development of medical imaging techniques and accumulation of clinical measurements, databases of medical information have been constantly growing. The increased availability of data has revolutionized cardio/cerebrovascular research. Although medical data becomes essential to assist disease diagnosis and management, it alone is unable to provide predictive information. Moreover, the low signal-to-noise ratio of the medical imaging limits its capability of quantitative examination.
Combining computational modeling with medical data is promising to address the fundamental challenges and mitigate shortcomings of each side. In this project, the objective is to integrate the medical data into the computational model to improve predictive capability. Advanced data assimilation techniques, machine learning algorithms, and statistical and Bayesian inference methods are required to achieve this goal. Moreover, uncertainty quantification and reduction for the predictions will be another focus.
Faculty involved in this research area include Jian-Xun Wang