A pipeline for deriving computational fluid models of the heart-to-brain pathway from standard-of-care clinical workup for stroke

Embolic Stroke of Undetermined Source accounts for a significant fraction of all ischemic strokes, often with limited ability to disambiguate embolic stroke etiology - a critical step in improving treatment efficacy and reducing incidence of recurring stroke. In silico modeling of embolus transport can elucidate embolus source-destination dynamics but accurate hemodynamics modeling is needed to contextually simulate a stroke scenario. A common difficulty encountered when developing such computational models is integrating patient-specific physiology within the simulated hemodynamics that reflect each patient’s realistic condition. We will present a methodology for integrating patient-specific data derived from standard-of-care clinical workups from multi-modal imaging into data-rich computational fluid models of the heart-to-brain pathway. Specifically, we have used angiography, ultrasound, and perfusion imaging along with medical records data to derive arterial anastomosis, cerebral brain perfusion, systole and diastole cardiac timing, to inform in silico models. These efforts introduce a pipeline of transforming raw patient imaging from clinical workups into a patient-specific hemodynamic model that can further be used for digital twins of full embolic stroke scenarios.