Incremental kernel ridge regression for the prediction of soft tissue deformations

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Computational Mechanics

Background:

• Maxillofacial surgery often results in soft tissue deformation.
• Predicting this deformation is crucial for surgical planning and patient outcomes.
• Current methods may lack precision in forecasting post-operative facial changes.

Purpose of the Study:

• To develop a nonlinear regression model for predicting soft tissue deformation post-maxillofacial surgery.
• To integrate finite element model (FEM) derived features with patient-specific biomechanical properties.
• To establish a generalizable and accurate method for forecasting facial changes.

Main Methods:

• Utilized a nonlinear regression model to analyze the relationship between extracted features and facial deformation.
• Employed finite element model (FEM) for feature extraction from 3D data.
• Calculated facial deformation using pre-operative and post-operative 3D scans.
• Combined a general predictive relationship with individual patient biomechanical properties for new predictions.

Main Results:

• Established a statistically and biomechanically relevant model for predicting soft tissue deformation.
• Demonstrated the model's effectiveness and efficiency through validation on eleven patients.
• Achieved accurate predictions by integrating generalizable relationships with patient-specific data.

Conclusions:

• The proposed nonlinear regression model offers an effective and efficient approach for predicting post-maxillofacial surgery soft tissue deformation.
• The model's ability to incorporate biomechanical relevance and statistical accuracy enhances its clinical applicability.
• This method provides a valuable tool for surgical planning and improving patient outcomes in maxillofacial procedures.