Step-specific learning curve of the TMINI robotic system for total knee arthroplasty: a single-surgeon CUSUM analysis

Robotic total knee arthroplasty (R-TKA) adoption continues to expand, yet most learning curve analyses focus on global operative time rather than individual procedural steps. The TMINI system is a handheld, imageless robotic platform with limited published data. We evaluated step-specific learning curves in a single surgeon's first 75 consecutive cases with complete intraoperative timing data. We retrospectively reviewed 75 primary TKAs performed with the TMINI system (January-August 2025). Four intraoperative timing metrics were recorded: bone registration time, intraoperative planning time, bone resection time, and time from incision to implants. Learning curves were assessed using cumulative sum (CUSUM) analysis; turning points were identified at peak CUSUM values. Early-versus-later comparisons used Mann-Whitney U tests. Sensitivity analysis excluded extreme outliers (> mean + 3 SD). CUSUM turning points varied considerably by procedural step. Time to implants stabilized earliest (case 8; 12.0% median reduction; P = 0.006). Intraoperative planning time turned at case 12 (32.1% reduction; P < 0.001). Bone resection time trended toward improvement but did not reach significance in the primary cohort (P = 0.131); significance was achieved after outlier exclusion (35.0% median reduction; P < 0.001). Bone registration was most gradual, turning at case 56 (10.1% reduction; P = 0.002). Sensitivity analyses were consistent. The TMINI system demonstrated measurable but nonuniform step-specific learning curves across 75 cases. Cognitive and software-driven steps improved early, while platform-specific tasks such as bone registration required more cases to stabilize. Step-level CUSUM analysis provides a more actionable picture of TKA adoption than global operative time and may guide training and scheduling.