Augmented Reality-Guided Individualized Total Knee Arthroplasty: Initial Clinical Implementation and Results

Contemporary total knee arthroplasty (TKA) practice has evolved from uniform mechanical alignment to individualized alignment planning in response to growing evidence that a single alignment strategy does not optimize outcomes for all patients. However, precise execution of patient-specific alignment strategies remains challenging with manual instrumentation. Augmented reality (AR) technology addresses this limitation by providing real-time intraoperative visualization of anatomic landmarks and planned resection parameters. In this study, individualized alignment was applied within predefined safety envelopes, including femoral valgus relative to the mechanical axis of 0° to 3°, tibial varus of 0° to 3°, and hip-knee-ankle alignment within 0° to 3° of neutral. All procedures were performed using the Pixee Knee + AR system (Pixee Medical, Besançon, France). Despite growing interest in AR-assisted surgery, clinical data regarding the accuracy, precision, and early outcomes of AR-guided individualized TKA remain limited, providing the rationale for this study.

This retrospective consecutive case series evaluated 50 patients with end-stage knee osteoarthritis who underwent primary TKA using the Pixee Knee + AR system between April and August 2024. All procedures used individualized alignment strategies with patient-specific planning parameters. Radiographic accuracy was assessed using postoperative CT imaging, and validated patient-reported outcome measures were collected at minimum 6-month follow-up.

The cohort comprised 45 men and 5 women with mean age 71.4 ± 7.6 years, American Society of Anesthesiologists score 3.0 ± 0.2, and body mass index 31.0 ± 4.1 kg/m2. Alignment accuracy was excellent, with 96% of coronal measurements within 3° of planned values and 53.3% within 1°. Component-specific accuracy rates within 1° were as follows: tibial component 72%, femoral implant 62%, and overall hip-knee-ankle alignment 46%. Four percent of measurements exceeded 3° deviation from planned alignment. No revision surgeries, infections, or mechanical complications occurred during follow-up. Patient-reported outcomes demonstrated notable improvements from preoperative scores: modified single assessment numeric evaluation scores improved 4.7 ± 3.1 points (P < 0.001) and KOOS Jr. normalized scores improved 21.6 ± 23.2 points (P < 0.001).

AR-guided individualized TKA demonstrated high accuracy in executing patient-specific alignment plans with favorable early clinical outcomes and no complications at 6 months. These initial results suggest that AR technology may provide a viable alternative for precise implementation of individualized alignment strategies, although larger studies with longer follow-up are needed to establish long-term efficacy and safety.