Automated outcome scoring in a virtual reality simulator for endodontic surgery

Area of Science:

• Dental Surgery Simulation
• Virtual Reality (VR) Technology
• Medical Education Technology

Background:

• Automated outcome assessment is crucial for effective feedback in VR surgical simulators.
• Current methods lack objective scoring and natural language communication for dental students.
• This study addresses the need for comprehensive, automated feedback in endodontic surgery training.

Purpose of the Study:

• To develop and evaluate an automated system for outcome assessment in VR endodontic surgery simulators.
• To provide objective, consistent, and natural language feedback to dental students.
• To establish a foundation for automated generation of comprehensive feedback in surgical training.

Main Methods:

• Computed virtual reference templates based on tooth anatomy.
• Assessed outcomes using a 3D score cube with voxel-level weighted scores.
• Transformed detailed scores into standard dental school scoring language.
• Evaluated the system with fifteen outcome samples and five expert endodontists.

Main Results:

• The automated system demonstrated a high degree of agreement with expert endodontist scores.
• Analysis showed some disagreement among human expert scores, highlighting the subjectivity of manual assessment.
• The system's scores effectively discriminated between clinically acceptable and unacceptable outcomes (AUC values).

Conclusions:

• The proposed automated scoring system offers objective consistency and aligns well with expert evaluations.
• Translating scores into common dental terminology facilitates natural communication between students and instructors.
• The automated template creation and scoring approach is robust and applicable to various dental drilling procedures.