Automated surgical planning in spring-assisted sagittal craniosynostosis correction using finite element analysis and...

Area of Science:

• Biomedical Engineering
• Computational Biology
• Pediatric Neurosurgery

Background:

• Sagittal synostosis, caused by premature fusion of the sagittal suture, leads to skull deformities in infants.
• Current surgical planning for spring-assisted cranioplasty lacks precision, potentially resulting in suboptimal outcomes.
• Existing methods focus on skull anatomy or require extensive simulations, hindering clear parameter selection.

Purpose of the Study:

• To develop an automated tool architecture for predicting post-operative outcomes in spring-assisted cranioplasty for sagittal synostosis.
• To integrate machine learning and finite element analysis for enhanced surgical planning.
• To improve the accuracy and efficiency of surgical parameter selection.

Main Methods:

• Developed a finite element model simulating calvarium properties, osteotomy sizes, and spring characteristics.
• Tested six machine learning algorithms, including XGBoost, against the finite element model.
• Utilized a statistical shape model of a sagittal synostosis calvarium for algorithm assessment.

Main Results:

• The XGBoost algorithm demonstrated high accuracy in predicting the post-operative cephalic index.
• Finite element simulations validated the predictions made by the XGBoost algorithm.
• The developed architectural structure provides a robust framework for outcome prediction.

Conclusions:

• The proposed automated tool architecture can significantly improve surgical planning for spring-assisted cranioplasty.
• Accurate prediction of post-operative cephalic index aids in optimizing surgical parameters.
• This approach has the potential to enhance patient outcomes in sagittal craniosynostosis correction.