Imageless and image‐based robotic‐assisted total knee arthroplasty achieve equivalent radiographic accuracy: A randomised controlled trial

PURPOSE: Robotic-assisted total knee arthroplasty systems offer both image-based and imageless workflows, but their comparative accuracy remains unclear. The robotic surgical assistant (ROSA) system uniquely provides both approaches within a single platform. This study compared radiographic accuracy between image-based and imageless ROSA-assisted total knee arthroplasty (ROSA-TKA). METHODS: This double-blind, randomised controlled trial included patients with primary knee osteoarthritis randomised to imageless (n = 47) versus image-based (n = 48) ROSA-TKA. Primary outcomes assessed the differences between final intraoperative planned measurements (FIPM), robotic-validated measurements (RVM) after bone cuts and postoperative scannogram measurements (PSM) across five radiographic alignment parameters measured at 12 weeks: hip-knee-ankle angle, femoral coronal and sagittal alignment, tibial coronal alignment and tibial slope. Secondary outcomes included functional scores and complications. RESULTS: Both techniques demonstrated equivalent radiographic accuracy with no significant differences between groups. Mean differences between FIPM and RVM were ≤1 degree for all parameters, achieving 93.6%-100% accuracy within 2-degree thresholds. No significant differences were observed between imageless and image-based approaches in comparisons of FIPM and RVM with postoperative scanograms. However, these comparisons showed larger differences (1.2-2.5 degrees) and lower accuracy rates (48.9%-87.5% within 2-degree thresholds), with sagittal alignment showing the greatest variation. Functional outcomes improved equally in both groups at 6 and 12 weeks, with comparable registration times and no complications. CONCLUSION: Both imageless and image-based ROSA-TKA achieved equivalent radiographic accuracy and clinical outcomes. The choice between techniques should be based on practical workflow considerations rather than accuracy concerns, as both approaches deliver comparable precision and excellent safety profiles. CLINICAL TRIAL REGISTRATION: This study was registered in the Thai Clinical Trials Registry (TCTR20240614001). LEVEL OF EVIDENCE: Level I.