Preoperative CT-Based Three-Dimensional Templating in Robot-Assisted Total Knee Arthroplasty More Accurately Predicts Implant Sizes than Two-Dimensional Templating.

Abstract

Patient dissatisfaction after total knee arthroplasty (TKA) is a concern. Surgical error is a common, avoidable cause of failed TKA. Correct femoral and tibial component sizing improves implant longevity, clinical outcomes, knee balance, and pain scores. We hypothesized that preoperative three-dimensional (3D) templating for robot-assisted TKA (RA-TKA) is more accurate than two-dimensional (2D) digital templating. Prospectively collected data from 31 RA-TKAs were assessed to determine accuracy pertaining to implant sizing and positioning. All cases undergoing RA-TKA undergo preoperative CT-scans as per protocol. Three blinded observers retrospectively templated these knees for TKA using standard radiographs. We compared whether 2D templating was as accurate as CT-guided templating. Postoperative radiographs were then evaluated for sizing and positioning. Intraclass correlation coefficients (ICCs) and the effect of learning curve were assessed. Preoperative femoral component 3D templating and retrospective blinded 2D templating accuracies were 96.6% and 52.9%, respectively (χ 2: 17.965; odds ratio [OR]: 24.957, 3.250-191.661; p < 0.001). Tibial component 3D and 2D templating accuracies were 93.1% and 28.7%, respectively (χ 2: 36.436; OR: 33.480, 7.400-151.481; p < 0.001). ICC for the three radiograph observers was 0.920 (95% confidence interval [CI]: 0.652-0.890; p < 0.001) for the femur and 0.833 (0.717-0.911; p < 0.001) for the tibia, showing excellent agreement. We conclude that preoperative CT-based templating for RA-TKA more accurately predicts the size of implants compared with traditional 2D digital templating. This may improve operating room efficiency and cost containment.