An accurate scapula registration process in shoulder arthroplasty using mixed reality

Area of Science:

• Medical Engineering
• Surgical Technology
• Computer-Assisted Surgery

Background:

• Shoulder arthroplasty requires precise glenoid component positioning to prevent revision surgeries.
• Limited surgical exposure in arthroplasty presents a significant challenge.
• Mixed reality (MR) offers potential guidance solutions through bone-to-model registration.

Purpose of the Study:

• To develop and evaluate a mixed reality registration workflow for shoulder arthroplasty using HoloLens 2.
• To achieve high accuracy in glenoid guidewire placement and component orientation.
• To ensure the registration process is efficient, completed within 5 minutes.

Main Methods:

• A marker-based tracking system and an improved Iterative Closest Point algorithm were employed.
• The workflow was integrated with the HoloLens 2 Head Mounted Display for real-time guidance.
• Accuracy targets were set at 1.5 mm for entry point and 1.5 degrees for inclination/version.

Main Results:

• Average accuracy for glenoid guidewire entry point was 0.84 ± 0.58 mm (AP) and 0.49 ± 0.41 mm (SI).
• Inclination and version accuracy averaged 0.89 ± 0.6 degrees and 0.97 ± 0.8 degrees, respectively.
• The mean process time was 1 minute 24 seconds, well within the target.

Conclusions:

• A complete, embedded registration workflow for shoulder arthroplasty was successfully developed.
• The MR system demonstrated promising accuracy for glenoid guidewire placement.
• The workflow enhances surgeon concentration by keeping all guidance within their field of view.