A real-time digital twin for human cardiovascular applications
Digital twin (DT) technology is transforming many industries and has great potential in healthcare as well. Such a DT has numerous potential cardiovascular applications, including veno-arterial extracorporeal membrane oxygenation (VA-ECMO), a complex life-support therapy used to provide cardiac support to critically ill patients. Deciding when a patient can safely be weaned from ECMO requires detailed mechanistic information about cardiac function. However, this information is difficult to obtain. Therefore, this research focuses on developing a DT that estimates left ventricular (LV) contractility from LV pressure measurements.
A human DT is a virtual representation of a part of the human body, such as an organ or tissue. It reflects the current physiological state and predicts how it changes over time. This makes it possible to monitor physiological processes that are difficult or even impossible to measure directly. In addition, a DT can simulate disease progression and predict responses to treatment, supporting clinical decision-making and personalization of treatment plans. Despite its great potential, developing a real-time digital twin remains challenging.
In this project, the DT was developed based on data assimilation (DA), a method that combines physiological measurements with a mathematical model to continuously update model predictions based on recent patient data. Specifically, the proposed approach uses a reduced-order unscented Kalman filter (ROUKF) with a lumped-parameter model of the systemic circulation. Proof of principle was evaluated using different types of data: synthetic, in vitro, and in vivo. In addition, several analyses were performed to assess the performance, reliability, and robustness of the proposed approach.
The ultimate goal is to enable real-time, patient-specific monitoring of cardiac parameters to support clinical decision-making. More broadly, the potential of a DT extends far beyond real-time monitoring, offering new opportunities for personalized medicine with the potential to improve patient outcomes.
Stan Snijders
Supervisors
Collaborators