Abstract
Introduction
Removal of primary components during revision TKA procedure can damage underlying bone, resulting in defects that may need filled for stability of the revision reconstruction. Special revision components including cones and/or augments are often used to compensate for the missing bones. Little work has been done to characterize metaphyseal geometry in the vicinity of the knee joint, however, in order to motivate proper size and shape of cones and augments. The objective of this study was to use statistical shape modelling to evaluate variation in endosteal anatomy for revision TKA.
Methods
Digital models of the femur and tibia were generated through segmentation of computed tomography scans, for the femur and the tibia (n∼500). Custom software was used to perform virtual surgery and statistical shape analysis of the metaphyseal geometry.
A representative and appropriately sized revision femoral component was placed on each bone, assuming anterior referencing with an external rotation of 3 degrees from the posterior condyle axis. The outer and inner boundaries of the cortical bone were determined at the resection level and at 5 mm increments proximally, up to 40 mm. Similar analyses were performed on the tibia, using a typical revision resection (0 degrees medial and posterior slope), with outer and inner boundaries of the cortical bone were determined in 5 mm increments up to 40mm distal to the resection.
Metaphyseal contours were exported relative to the central fixation feature of the implant, and average geometries were calculated based on size, and across the entire cohort. Principal Component Analysis (PCA) was used to quantify the variability in shape, specifically to evaluate the +/− 1 and 2 standard deviation geometries at each cross section level of Principal Component 1 (PC1).
Results
Representative results illustrating the effect of size for the femur at single depth and the effect of depth and PC1 for tibia are reported. The average inner metaphyseal geometry of the femur (30mm proximal to resection) varied from 25.1×47.7 mm (AP x ML) at the smallest size to 54.5×78.0 at the largest size. The overall average tibia geometry decreased from 51.5×69.5 mm at the base resection level to 33.5×31.3 mm at the most distal resection level (40mm) distal to the resection. At the 20 mm level, the average tibia contour of 45.0×47.8 mm changed to 32.2×33.4 at −2 standard deviations of PC1 and 57.9×62.4 mm at the +2 standard deviations of PC1
Discussion
The generated contours can be used as a design input to optimize the shape of cones and augments, in order to fit potential defects in the femur and tibia encountered during revision TKA while respecting the anatomical constraints of the bone. Statistical shape analysis shows that these constraints are not strictly uniform scaling, based on bone size or on location in the metaphysis, but rather reflect variations in shape that may be used to optimize fit and stability of the prostheses.
