Accuracy and Interobserver Reliability of Preoperative Templating of the Tibial Component in Medial Unicompartmental Knee Replacement.

Background Preoperative templating is an important component of surgical planning in knee arthroplasty. In medial unicompartmental knee replacement (UKR), accurate tibial component sizing can be challenging due to limited surgical exposure and the risk of posterior cortical injury. This study aimed to evaluate the accuracy and interobserver reliability of preoperative templating using templating software in predicting tibial component size. Methods A retrospective observational study was conducted including 69 patients who had undergone medial UKR between January 2023 and December 2024. Two orthopaedic specialty doctors independently templated preoperative lateral knee radiographs using TraumaCad software (Brainlab, Munich, Germany), blinded to the actual implanted component size and to each other's assessments. Predicted sizes were compared between observers and against the implanted components. Interobserver reliability was assessed using the intraclass correlation coefficient (ICC) with a two-way random effects model for absolute agreement and weighted Cohen's kappa with quadratic weights. Results Exact agreement between observers was observed in 39 of 69 cases (56.5%), increasing to 68 of 69 (98.6%) within ±1 size. Observer 1 predicted the exact implanted size in 35 cases (50.7%), improving to 60 cases (87.0%) within ±1 size. Observer 2 demonstrated higher accuracy with exact prediction in 42 cases (60.9%) and 66 cases (95.7%) within ±1 size. Interobserver reliability was almost perfect with an ICC of 0.82 (95% CI: 0.72-0.88) and a weighted kappa value of 0.82. Bland-Altman analysis revealed minimal systematic bias between observers (mean difference 0.33 sizes). Observer 1 showed systematic oversizing compared to actual implants (mean bias +0.45 sizes, p<0.001), while Observer 2 demonstrated no significant bias (mean bias +0.12 sizes, p=0.19). Conclusion Preoperative templating using TraumaCad demonstrated almost perfect interobserver reliability and good-to-excellent accuracy within a clinically acceptable range. It represents a useful adjunct in surgical planning for medial UKR. Future studies comparing templating performance across different software platforms would be valuable.