Reproducibility and Precision of CT Scans to Evaluate Tibial Component Rotation

Abstract Background Component rotation likely plays a greater role on the survivorship and outcomes of total knee arthroplasties (TKAs) than is currently known. Our goal was to evaluate the precision, inter-observer reliability, and intra-observer reproducibility of tibial component rotation as measured by CT scan, regardless of measurement technique. Methods Three fellowship-trained, academic arthroplasty surgeons independently measured tibial component rotation on CT scans of 62 TKAs using their methods of choice. Measurements were repeated at least 2 weeks after the initial measurement. The precision of the measurements was assessed using a formal eight-step protocol as the gold standard. Intra-class correlation coefficients (ICCs) were calculated to evaluate precision, inter-observer reliability, and intra-rater reproducibility (IRR). Comparison of surgeons’ measurement to a validated gold standard revealed only moderate precision for tibial component rotation (ICCs were 0.68, 0.57, and 0.64, respectively). Results The inter-observer agreement between the three surgeons for tibial component rotation was also moderate (ICC = 0.48). The intra-rater reliability of tibial rotation was excellent (IRR = 0.80). Conclusions Practicing surgeons measuring tibial rotation were internally consistent, but failed to demonstrate satisfactory precision and inter-observer agreement. We support the adoption of standardized criteria for the measurement of tibial component rotation on CT scans.