Can human judgment match robotic precision? A concordance study in robotic-assisted knee arthroplasty

Background: Robotic-assisted total knee arthroplasty (TKA) enhances intraoperative precision. However, the degree to which experienced surgeons can independently predict implant sizes before the robotic plan is generated remains unclear. This study evaluated the accuracy and inter-surgeon concordance in predicting femoral and tibial component sizes using the CORI robotic system. Methodology: Three experienced surgeons predicted femoral and tibial sizes in 20 consecutive TKA cases after surface registration but before the robotic plan was displayed. Their predictions were compared with the final implant sizes suggested by the CORI robotic system. Inter-surgeon agreement and surgeon-robot concordance were analyzed using kappa statistics. Results: Femoral sizing prediction accuracy was 95% (Cohen’s κ = 0.95), showing near-perfect agreement between surgeons and the robotic system. Tibial sizing prediction accuracy was 70% ( κ = 0.65). Inter-surgeon agreement was almost perfect for femoral sizing ( κ = 0.91) and substantial for tibial sizing ( κ = 0.89). Conclusion: This study demonstrates that experienced surgeons achieve high concordance with robotic femoral sizing, validating their intraoperative judgment. However, the lower tibial concordance (70%) underscores robotics’ critical role in addressing anatomical complexity, supporting a complementary rather than competitive relationship.