Learning curves of two surgical robot systems for assisted total knee arthroplasty and their impact on early patient clinical outcomes: a retrospective study.

As robotic-assisted total knee arthroplasty (TKA) continues to evolve, a key clinical question is whether the learning curve and clinical outcomes differ between CT-dependent and image-independent robotic systems. This retrospective study compared the learning curves and early clinical outcomes of 101 patients undergoing TKA with either a CT-dependent robotic system (Beijing HURWA, Group A) or an image-independent system (Smith & Nephew CORI, Group B). A statistically significant intergroup disparity (P < 0.05) was noted in the rates of transition to conventional treatment, with Group A exhibiting a higher frequency during the learning phase. However, this phase did not witness a significant difference between the groups for other outcomes (p > 0.05). Cumulative sum (CUSUM) analysis was employed to chart the learning curve., peaked at the 15th case for Group A and the 14th for Group B, indicating a similar number of cases to achieve proficiency. During both learning and proficiency phases, the HURWA system demonstrated a shorter bone resection time but longer reference array mounting and registration times. Radiographically, the CT-dependent HURWA system showed superior performance in achieving optimal frontal and lateral femoral component (FFC, LFC) angles. In the early learning phase, Group A reported higher pain scores (VAS) at postoperative day 7, but this difference resolved by day 180. A comparison of the groups demonstrated comparable final knee function (KSS, ROM) and complication rates at all assessment points. The findings indicate that while the two robotic systems exhibit distinct operational time profiles and early radiographic advantages for the CT-dependent system in femoral positioning, both facilitate comparable and satisfactory early clinical outcomes after the initial learning period.