Patellofemoral arthroplasty (PFA) has higher revision rates than total knee arthroplasty (TKA) [Van der List, 2015; Dy, 2011]. Some indications for revision include mechanical failure, patellar mal-tracking, implant malalignment, disease progression and persistent pain or stiffness [Dy, 2011; Turktas, 2015]. Implant mal-positioning can lead to decreased patient satisfaction and increased revision rates [Turktas, 2015]. Morphological variability may increase the likelihood of implant mal-positioning. This study quantifies the morphological variability of the anterior-posterior (AP) and medial-lateral (ML) aspects of the patellofemoral compartment using a database of computed tomography (CT) scans. The analysis presented here used the custom CT based program SOMA (SOMA V.4.3.3, Stryker, Mahwah, NJ). SOMA contains a large database of 3D models created from CT scans. Anatomic analysis and implant fitting tools are also integrated into SOMA to perform morphometric analyses. A coordinate system is established from the femoral head center, the intercondylar notch, and a morphological flexion axis (MFA). The MFA is created by iteratively fitting circles to the posterior condyles and creating and axis through the circles' centers. The sagittal plane is created normal to this axis and through the notch. A coronal plane is created from the femoral head center and the flexion axis. The AP measurement is taken normal to the coronal plane from the anterior cortex sulcus to the intercondylar notch (Figure 1). A 5°-flexed anterior resection is created to run-out at the anterior cortex sulcus. The ML measurement is taken normal to the sagittal plane from the most medial to the most lateral points of the anterior resection (Figure 1). The ML measurements are broken down into medial and lateral components divided by a sagittal plane through the trochlea. Means and standard deviations of the AP and ML measurements are calculated. The mean and standard deviation for the AP measurement are 24.9mm and 2.8mm, respectively. The data predicts that 99.7% of the population will have an AP measurement between 16.5mm and 33.3mm. The mean and standard deviation for the ML measurement are 54.6 mm and 5.5mm, respectively. The data predicts that 99.7% of the population will have an ML measurement between 38.1mm and 71.1mm A Pearson Correlation value of 0.134 was calculated for AP/ML indicating a very weak positive correlation between the measures. The correlation value and the large measurement ranges indicate that there is high variability between the AP and ML measurements. A scatterplot was created to graphically represent the high variability between the AP and ML width measurements (Figure 2). A Pearson Correlation value of −0.649 was calculated for the medial and lateral components of ML (Figure 3). The results of this study suggest that patellofemoral morphology is highly variable with respect to the AP and ML dimensions. This variability may impact implant fit and positioning and should be taken into consideration in the design and use of prostheses for PFA.
The condylopatellar notch (CPN) represents the border between the patellofemoral articulation and the tibiofemoral articulation [Pao, 2001]. This could be a valuable landmark for establishing the boundaries of unicompartmental knee replacements. Its location on the distal femur has been described radiographically, but it has not, to our knowledge, been quantified with respect to anatomic landmarks [Hoffelner, 2015]. This study seeks to leverage a large database of computed tomography (CT) scans to quantify the location of the CPN with respect to well established anatomic landmarks of the knee. The analysis presented here used the custom CT based program SOMA (SOMA V.4.3.3, Stryker, Mahwah, NJ). SOMA contains a large database of 3D models created from CT scans. Anatomic analysis and implant fitting tools were also integrated into SOMA to perform morphometric analyses. 986 healthy distal femurs were analyzed. A coordinate system was established from the femoral head center, the intercondylar notch, and a morphological flexion axis (MFA). The MFA was created by iteratively fitting circles to the posterior condyles and creating and axis through the circles' centers. The sagittal plane was created normal to this axis and through the notch. A plane was created from the femoral head center and the flexion axis. A coronal plane was created from this plane and a point on the anterior cortex sulcus. Points were placed on a template bone model in the medial and lateral extents of the surface depressions of both the medial and lateral aspect of the CPN, where the depression of the CPN is most distinct. These points were then mapped to each of the 986 femoral specimens via a shape correspondence model. A line is created between the pairs of points representing the medial and lateral CPN's. The coordinates of the points are measured with respect to sagittal and coronal planes (Figure 1). Means and standard deviations of the anterior-posterior (AP) and medial-lateral (ML) coordinates of the CPN points are calculated. The mean coordinates for the lateral CPN line are (4.8±1.6, −33.6±6.8) and (29.1±5.4, −18.7±4.8). The mean coordinates for the lateral CPN are (−20.7±3.8, −2.2±4.4) and (−6.5±1.6, −29.7±3.2). The means with error bars representing two standard deviations are plotted on a scatter plot (Figure 2). Boxes representing the location of the CPN line for 95% of the population are included on the plots. Until now, the location of this anatomic feature of the knee has not been quantified with respect to known anatomical landmarks. The location of the CPN could serve as a valuable landmark for determining the border between the tibiofemoral and patellofemoral articulations. This data can be used to locate the CPN and inform the planning and design of compartmental knee replacements.
