Intraoperative Assessment of Kinematics Using Robotic-Assisted Total Knee Arthroplasty Is Reliable: A Cadaver-Based Study

Computed tomography-based robotic-arm-assisted total knee arthroplasty (RATKA) enables three-dimensional surgical planning and intraoperative adjustment of implant positioning based on ligament laxity. Stability and kinematic assessments may offer enhanced insight into multiplanar knee laxity, but their reproducibility remains underexplored. This study evaluated the reliability of intraoperative knee kinematic (dynamic), sagittal and transverse stability assessments in a cadaver setting under different support conditions. Cruciate-retaining RATKA was performed on five fresh-frozen cadaver knees by three experienced surgeons. Medial and lateral anteroposterior translation (MAP, LAP) and internal-external rotation (IE) were measured at 10, 45, and 90 degrees of flexion before and after component implantation. Dynamic assessments across the full range of motion were used to calculate the average medial contact position (AMCP) and medial pivot ratio (MPR). Inter- and intra-rater reliability were determined using intraclass correlation coefficients (ICC: poor < 0.4, good 0.4 to 0.74, and excellent ≥ 0.75). Analyses compared a leg-holder-only condition with all surgeons, including manual support. Intra- and inter-rater reliability across all surgeons was generally good to excellent. For intact knees, reliability ranged from ICC 0.52 to 0.84 for MAP, 0.44 to 0.57 for LAP, and 0.48 to 0.62 for IE. With components, reliability remained good to excellent for MAP, LAP, and IE (ICC 0.47 to 0.80). Dynamic AMCP assessments demonstrated excellent inter-rater reliability (ICC 0.84 to 0.93), while MPR showed good reliability (ICC 0.57). The leg holder reduced variance for MAP/LAP and IE, maintaining error within two mm or 5 degrees, respectively. Intra-rater reliability was consistently excellent across nearly all measures (ICC 0.69 to 0.99). Intraoperative stability and kinematic assessments during RATKA are reproducible, particularly for AMCP. The leg holder generally improved consistently across observers and reduced variance. These findings support the reliability of robotic-assisted intraoperative stability and kinematic measures for evaluating knee function and guiding surgical planning.