Tool-tissue force estimation in laparoscopic surgery using geometric features.

Area of Science:

• Biomedical Engineering
• Computational Mechanics
• Soft Tissue Mechanics

Background:

• Accurate estimation of forces and stress in soft tissues is crucial for medical applications like haptic feedback and injury assessment.
• Current methods often rely on direct force measurements, which can be complex or invasive.
• Understanding the relationship between tissue deformation and mechanical forces is an ongoing research area.

Purpose of the Study:

• To introduce and validate three geometric features derived from soft tissue deformation for force and stress estimation.
• To develop and train neural networks using these features for predicting interaction forces.
• To assess the potential of these features for applications in haptic devices and tissue damage assessment.

Main Methods:

• Finite Element Method (FEM) simulations were used to create 2D and 3D models of porcine liver.
• Geometric features including maximum deformation angle, depth, and width of displacement constraint were calculated from simulated deformed shapes.
• Two neural networks were trained using the extracted geometric features and simulated interaction forces.

Main Results:

• The three geometric features demonstrated a strong correlation with probing force and maximum local stress.
• The trained neural networks showed high potential for accurate force estimation.
• The method was validated for large deformations up to 40%.

Conclusions:

• Geometric features of soft tissue deformation can reliably predict interaction forces and stress.
• This approach offers a promising, non-invasive method for force estimation in haptic devices.
• The findings contribute to better assessment of soft tissue damage during large deformations.