Digital Twins of Heart Veins With Artificial Intelligence Comes More Reliable

The accuracy of medical simulations can save life for patients. New research enables digital twins of aort (ana artery) models using artificial intelligence **more reliable. The research team tested how well the machine learning can be transferred between different patients.

Basic findings of the study:

• The aoretic model of four different patients is subjected to realistic fluid-structure interaction simulation
• The machine learning algorithm called LightGBM is trained based on geometry data
• The research measures how successful information learned from a patient can be applied to other patients
• In the surface where interpolation (a value estimate) and transfer between completely new anatomy (transfer) was compared

The results are important for us to understand what artificial intelligence is really reliable in medical simulations. Digital twins serve virtual laboratory for surgeons; it allows you to test treatment plans before they interfere with the patient. This research answers critical questions about the accuracy and overallability of these virtual models.

The implementation of the techniques used can help heart surgeons and cardiovascular medical experts make better decisions. The research is shared on the arXiv platform and open to the review of the academic community.