Augmented reality for computer assisted orthopaedic surgery

Abstract

In recent years, computer-assistance and robotics have established their presence in operating theatres and found success in orthopaedic procedures. Benefits of computer assisted orthopaedic surgery (CAOS) have been thoroughly explored in research, finding improvements in clinical outcomes, through increased control and precision over surgical actions. However, human-computer interaction in CAOS remains an evolving field, through emerging display technologies including augmented reality (AR) – a fused view of the real environment with virtual, computer-generated holograms. Interactions between clinicians and patient-specific data generated during CAOS are limited to basic 2D interactions on touchscreen monitors, potentially creating clutter and cognitive challenges in surgery.

Work described in this thesis sought to explore the benefits of AR in CAOS through: an integration between commercially available AR and CAOS systems, creating a novel AR-centric surgical workflow to support various tasks of computer-assisted knee arthroplasty, and three pre–clinical studies exploring the impact of the new AR workflow on both existing and newly proposed quantitative and qualitative performance metrics.

Early research focused on cloning the (2D) user-interface of an existing CAOS system onto a virtual AR screen and investigating any resulting impacts on usability and performance. An infrared-based registration system is also presented, describing a protocol for calibrating commercial AR headsets with optical trackers, calculating a spatial transformation between surgical and holographic coordinate frames. The main contribution of this thesis is a novel AR workflow designed to support computer-assisted patellofemoral arthroplasty. The reported workflow provided 3D in-situ holographic guidance for CAOS tasks including patient registration, pre-operative planning, and assisted-cutting. Pre-clinical experimental validation on a commercial system (NAVIO®, Smith & Nephew) for these contributions demonstrates encouraging early-stage results showing successful deployment of AR to CAOS systems, and promising indications that AR can enhance the clinician’s interactions in the future. The thesis concludes with a summary of achievements, corresponding limitations and future research opportunities.