Analysis of Waveform Parameters in the Retinal Vasculature via Mathematical Modeling and Data Analytics Methods.

Abstract

Patient-specific mathematical modeling combined with data analytics methods presents a promising approach for analyzing retinal hemodynamics. In this study, we build upon previous developments in retinal blood flow modeling, integrating clinical measurements and physiological insights to reconstruct parameters such as the central retinal artery velocity (CRA) Doppler profile and pressure wave. By leveraging personalized input data, including CRA velocity profile, systemic blood pressure, heart rate, and intraocular pressure, we evaluate the performance of our in silico approach. Our investigation highlights the significance of methodological considerations in combining automatic image processing and mathematical modeling, particularly concerning the selection of appropriate strategies and the inclusion of personalized supplementary data. Through extensive validation and comparison with prior works, we demonstrate the impact of different assessment methods on clinically meaningful quantities, particularly biomarkers related to blood flow. Furthermore, our study introduces a novel metric based on the Wasserstein distance for monitoring temporal changes in retinal blood flow dynamics, providing valuable insights into the evolution of vascular function. Overall, our findings underscore the importance of patient-specific input data, automatic image processing, and personalized mathematical modeling to ensure robust and clinically relevant outcomes in retinal vasculature analysis.