Mimics innovation suite

Preoperatively, MRI scans were obtained for all patients according to the standard Signature scanning protocols and sent to the manufacturing company (Materialise NV, Leuven, Belgium) for further processing. Standard procedures were followed according to the Signature MRI total knee system (Biomet, Inc., Warsaw, Indiana, United States). MRI images were converted to virtual 3D models (Mimics, Materialise NV, Leuven, Belgium), anatomical landmarks were selected for definition of the coordinate system and a virtual planning was developed and proposed to the surgeon (Biomet Signature Planner, Materialize NV, Leuven, Belgium). The Signature software allows modifications of implant size, resection level and component orientation in the coronal, sagittal and axial planes. A default planning was proposed to the surgeon who made changes where needed and gave his final approval.
Additionally, each patient had a pre-operative full leg CT-scan. CT-images were segmented (RMS (root mean square) error of 0.55 mm [22] (link)) using the Mimics Innovation Suite (Materialise NV, Leuven, Belgium). The CT-scans were not used for surgical planning, but were required additional measurements for the post-processing of the data, to be able to link the post-operative CT-based model more reliable to the pre-operative MRI-based model for the accuracy assessment of component positioning.

Postoperatively (average 46 days, range 31-81 days), a full leg CT scan was obtained for all patients using the same scanning protocol as preoperatively to calculate the differences between the preoperative plan and postoperative measurements. Three dimensional virtual bone models were derived from the Signature planning and postoperative CT images using the Mimics Innovation Suite (Materialise NV, Leuven, Belgium). First, CAD (computer aided design) models of the implanted components were superposed on their segmented postoperative counterparts. Second, the postoperative CT bone models were registered on the preoperative CT bone models using an iterative closest point (ICP) algorithm. Similarly, the pre-operative CT bone models were registered on the Signature planning models using the same method and manually transformed to the same coordinate system as used preoperatively to perform the planning. The same transformation was applied to the CAD models of the components. Thirdly, we measured the orientation of the CAD models in the post-operative 3D model and calculate the difference with the planned orientation in the coronal, sagittal and axial planes (Figure 2). Femoral varus angle deviation was defined as the angle between the intersection lines of the pre-and postoperative distal cut planes with the femoral coronal plane.