Cross-platform adaptation in robotic total knee arthroplasty: transitioning from an image-based semiactive robot to autonomous and imageless systems

Despite the widespread adoption of robotic systems for knee arthroplasty, the impact of prior robotic surgical experience on the transition to different robotic systems remains limited. Adaptation patterns across robotic platforms during the transition may differ and have clinical relevance. We retrospectively studied patients undergoing total knee arthroplasty with newly introduced autonomous image-based (Cuvis) and imageless semiactive robotic systems (Velys), compared with an established semiactive, image-based (Mako) robotic program. Because the surgeon had performed over 500 Mako-assisted TKAs prior to the study period, the Mako program was considered mature and used as a benchmark for studying the pattern of adaptation to new systems. The procedures were performed by a single surgeon. Alignment accuracy was assessed using absolute alignment error, defined as the deviation between intraoperative robotically achieved hip-knee-ankle (HKA) alignment and postoperative HKA alignment. Secondary outcomes included alignment outliers (> 3°), tourniquet time, specific complications, conversion to conventional techniques, and cumulative sum (CUSUM) analysis of learning pattern. CUSUM analysis was used to sequentially assess deviations in operative performance from predefined acceptable targets and to identify changes in performance over the sequence of cases. Mean absolute alignment error differed significantly among platforms (Mako: 1.30 ± 1.59°, Velys: 1.58 ± 1.63°, Cuvis: 2.48 ± 1.93°; p = 0.007). Outlier rates (> 3°) were 6.0%, 16.0%, and 28.6%, respectively ( p = 0.018). Cuvis had a significantly higher outlier rate than Mako (adjusted p = 0.033), whereas Mako and Velys did not differ (adjusted p = 0.615). Mean tourniquet times were 60.0 ± 13.9, 63.5 ± 12.3, and 75.0 ± 13.7 min, respectively ( p < 0.001). Two vascular injuries and two conversions to conventional technique occurred with Cuvis. CUSUM analysis demonstrated stabilization of alignment performance after approximately 31 cases with Velys, whereas no clear proficiency point was observed within 35 Cuvis cases. The adoption pattern of the Velys system was comparable to that of the established Mako program. Adoption of the Cuvis system was associated with greater alignment error and higher outlier rates, longer tourniquet times, vascular injuries, and conversions to manual technique. Prior robotic experience may facilitate adoption of some robotic platforms more effectively than others. Because a mature Mako program was compared with newly adopted systems, these findings should be interpreted as platform-related patterns of early adoption rather than definitive differences in long-term performance, safety, or platform superiority.