The Accuracy of CT-Based Three-Dimensional Templating in Predicting Implant Sizes in Patients Undergoing Robot-Assisted Total Knee Arthroplasty

Abstract

Background

Computed tomography (CT) based three-dimensional templating is increasingly being used to predict implant sizes in total knee arthroplasty (TKA). However, the existing data is heterogeneous, and the majority of studies lack adequate statistical power. This study investigated whether preoperative CT-based planning in robot-assisted TKA (RA-TKA) helps in predicting the accurate size of implant used.

Methods

This is a single-center retrospective study of 632 consecutive RA-TKA surgeries. All surgeries were performed using a fully automatic Cuvis RA-TKA system. Cohen’s Kappa (κ) coefficient was used to measure the level of agreement between the predicted and the final implant sizes.

Results

A total of 632 knees were operated on 384 patients. A total of 136 unilateral cases whereas 248 patients had both knees operated on. For the tibial component, in 21.7% cases a bigger implant size was used while in 11.8% cases a smaller size was used. For the femoral component, in 5.1% cases a bigger implant size was used while in 4.9% cases a smaller size was used. The agreement between the predicted and actual implant sizes was moderate for the tibial component [κ = 0.56 (95% CI: 0.51 to 0.61); p < 0.001] and almost perfect for the femoral component [κ = 0.87 (95% CI: 0.84 to 0.90); p < 0.001].

Conclusion

This study suggests that planning of RA-TKA using a CT-based model can be valuable to surgeons in accurately predicting the component size for femur and to a lesser degree for tibia. Future studies should investigate the potential predictors of discordance between the predicted and actual tibial implant sizes.