A comparison between 3D printed models and standard 2D planning in the use of metal block augments in revision knee arthroplasty.

Objectives: The study focused on the ability to predict the need and size of femoral and tibial augmentation using standard two-dimensional (2D) templates and models created with three-dimensional (3D) printing in surgical planning.

Patients and methods: This observational cohort study included 28 consecutive patients (22 females, 6 males; mean age: 71±7.3 years; range, 54 to 82 years) with periprosthetic joint infection recruited between March 2021 and September 2023 undergoing revision total knee arthroplasty revision (TKA). Standard planning was made using calibrated X-ray images. The 3D planning started with computed tomography scans to generate a 3D template of the distal femur and proximal tibia. The model was exported to a 3D printer to produce a patient-specific phantom. The surgery was then simulated on the 3D phantom using revision knee arthroplasty instrumentation to evaluate the appropriate augmentation to use until a correct alignment was obtained.

Results: Three-dimensional planning predicted the need for femoral and tibial augments in 22 (78.6%) cases at both the tibial and femoral components, while 2D planning correctly predicted the need for augmentation in 17 (60.7%) for the tibial side and 18 (64.3%) for the femoral side. The Cohen's kappa demonstrated a significant agreement between the 3D planning for the femoral metal block and the intraoperative requirement (kappa=0.553), whereas 2D planning showed only nonsignificant poor agreement (kappa=0.083). In contrast, the agreement between 2D or 3D preoperative planning for tibial augment and the intraoperative requirement was nonsignificant (kappa=0.130 and kappa=0.158, respectively). On the femoral side, 2D planning showed only a fair nonsignificant correlation (r=0.35, p=0.069), whereas 3D planning exhibited substantial agreement with the actual thickness of the implanted augment (r=0.65, p<0.001). On the tibial side, 3D and 2D planning showed substantial agreement with the actual size of implanted augments (3D planning, r=0.73, p<0.001; 2D planning, r=0.69, p<0.001).

Conclusion: Prediction based on 3D computed tomography segmentation showed significant agreement with the intraoperative need for augmentations in revision TKA. The results suggest that planning with 3D printed models represents a stronger aid in this kind of surgery rather than standard 2D planning, providing greater accuracy in the prediction of the required augmentation in revision TKA.