Systematic review of machine learning applications using nonoptical motion tracking in surgery.

Area of Science:

• Surgical Technology
• Biomedical Engineering
• Artificial Intelligence in Medicine

Background:

• Surgical motion analysis is crucial for skill assessment and improving surgical outcomes.
• Non-optical motion tracking systems (NOMTS) offer a practical alternative to optical methods for capturing surgical movements.
• Integrating machine learning (ML) with NOMTS presents opportunities to enhance the analysis of surgical performance.

Purpose of the Study:

• To systematically review machine learning applications in surgical motion analysis using non-optical motion tracking systems (NOMTS).
• To investigate the objectives, experimental designs, model effectiveness, and future research directions in this field.

Main Methods:

• Systematic literature review of 3632 records, including 84 studies.
• Analysis of machine learning models (e.g., Artificial Neural Networks, Support Vector Machines), objectives (e.g., skill assessment), NOMTS types, surgical settings, and procedure types.
• Evaluation of reported accuracy rates.

Main Results:

• Artificial Neural Networks (38%) and Support Vector Machines (11%) were the most common ML models.
• Skill assessment was the primary objective (38%).
• High accuracy (over 90%) was achieved in 36% of the reviewed studies, demonstrating the potential of NOMTS and ML.

Conclusions:

• Non-optical motion tracking systems combined with machine learning can significantly improve surgical precision, assessment, and training.
• Future research should focus on advancing ML in surgical environments, ensuring model interpretability and reproducibility.
• Larger datasets are needed for more accurate evaluation of ML models in surgical motion analysis.