Synthetic, Population-Based Virtual Patient Database Using a Digital Twin of the Cardiovascular System.

Area of Science:

• Cardiovascular physiology
• Computational modeling
• Medical informatics

Background:

• Developing realistic cardiovascular models is crucial for advancing medical diagnostics and decision-making.
• Virtual patient databases (VPDs) offer a powerful tool for simulating complex physiological scenarios.
• Integrating physiological and demographic data enhances the utility of VPDs for research and training.

Purpose of the Study:

• To develop a cardiovascular virtual patient database (VPD) integrating physiological and demographic data.
• To establish a foundation for applications in medical diagnostics, decision-making, and uncertainty analysis.
• To enable integration with 3D hemodynamic models and facilitate neural network training.

Main Methods:

• Generated an initial VPD by treating cardiovascular model parameters as stochastic variables.
• Ensured physiological plausibility and accuracy using literature data, sensitivity analysis, and resampling.
• Included key physiological quantities (e.g., aortic pressure, cardiac output, pulse wave velocities) and demographic factors (sex, age).
• Utilized the open-source hemodynamic solver, first_blood, for accuracy and efficiency.

Main Results:

• The final VPD comprises 34,347 virtual patients.
• VPD-derived aortic pressures closely matched literature values, with minimal differences.
• The VPD accurately reproduced sex-based differences in diastolic aortic pressure.
• Age-dependent increases in pulse wave velocity were accurately represented in the VPD.

Conclusions:

• The methodology effectively generates physiologically realistic virtual patient data.
• The first_blood solver facilitates the creation of robust cardiovascular simulations.
• The developed VPD serves as a valuable resource for 3D simulations, machine learning, and clinical decision support.