• PURPOSE
    • This study sought to validate intraoperative robotic measurements of femoral and tibial component coronal alignment in total knee arthroplasty (TKA) by comparing to pre- and post-operative standing, double stance, long-leg radiographs (LLR).
  • METHODS
    • This retrospective cohort study included 59 unique patients undergoing primary TKA at a single institution. Pre- and post-operative femoral and tibial coronal alignment were measured on LLRs using a deep learning artificial intelligence model and compared to measurements obtained from the imageless robotic system to evaluate the robot's accuracy and reliability.
  • RESULTS
    • Robotic measurements were highly correlated with measurements from preoperative LLR (Pearson r2 = 0.68). There was no significant difference in preoperative constitutional alignment between the two methodologies (p = 0.28). Additionally, the intraoperative and post-operative alignment of femoral and tibial implants were not significantly different (p = 0.12 and p = 0.95, respectively) and were strongly correlated (Pearson r2 = 0.5 and Pearson r2 = 0.6 respectively). The mean difference in femoral alignment was 0.43° and the mean difference in tibial alignment was 0.01°.
  • CONCLUSIONS
    • The findings of this study suggest that there were no significant differences in the coronal alignment of TKA when assessed by a robotic system compared to LLR. This signifies the robotic system's high intraoperative accuracy and reliability in determining coronal alignment.
  • LEVEL OF EVIDENCE
    • Level III.