[Study on Prediction Model of Soft Tissue Deformation during Needle Insertion].

Area of Science:

• Biomedical Engineering
• Materials Science
• Computational Modeling

Background:

• Simulating soft tissue is crucial for understanding medical device interactions.
• Flexible needle insertion requires precise control to avoid tissue damage.
• Existing methods for predicting tissue deformation have limitations.

Purpose of the Study:

• To develop a quantitative model for soft tissue deformation during flexible needle insertion.
• To evaluate the accuracy of a back propagation neural network in predicting tissue displacement.
• To assess the impact of the predictive model on improving needle insertion accuracy.

Main Methods:

• Polyvinyl alcohol (PVA) hydrogel was fabricated to mimic human soft tissue properties.
• Markers were embedded in the hydrogel, and their displacements were tracked using an image acquisition system.
• A back propagation (BP) neural network was trained to model the Y-direction displacement based on marker data.

Main Results:

• The BP neural network model achieved a fitting degree above 95% compared to experimental data.
• The maximum relative error for valid data was limited to 30%, with a maximum absolute error of 0.8mm.
• The model demonstrated quantitative prediction capabilities for soft tissue deformation.

Conclusions:

• The developed PVA hydrogel effectively simulates soft tissue for experimental analysis.
• The BP neural network provides a reliable method for quantitatively predicting soft tissue deformation.
• This predictive modeling approach can enhance the accuracy of flexible needle insertion procedures.