Personalizing computational models to construct medical digital twins.

Area of Science:

• Biomedical Engineering
• Computational Biology
• Digital Health

Background:

• Digital twin technology, originating in engineering, is increasingly applied to biomedicine.
• Personalizing computational models with patient data is crucial for advancing personalized medicine.
• Complex biomedical models, including agent-based models, lack standardized personalization methods.

Purpose of the Study:

• To develop a novel algorithm for dynamically calibrating complex biomedical models to individual patients.
• To bridge the gap between clinically measurable macrostates and detailed microstate data for model personalization.
• To enhance the accuracy of model-based predictions in personalized medicine.

Main Methods:

• Application of the ensemble Kalman filter, a data-assimilation technique, at the macrostate level.
• Linking macrostate updates from the Kalman filter to corresponding microstate updates.
• Ensuring microstate consistency with desired macrostates and model dynamics.

Main Results:

• A new algorithm for personalizing complex biomedical models has been proposed.
• The method effectively bridges macrostate and microstate data for improved model calibration.
• Enhanced personalization of agent-based models and other complex biomedical simulations.

Conclusions:

• The proposed algorithm offers a standardized approach to personalizing complex biomedical models.
• This method improves the accuracy of patient-specific forecasts, advancing personalized medicine.
• Facilitates the integration of digital twin technology in clinical practice.