Comparative Analysis of Sagittal-Plane Radiographic Landmarks Used to Identify the Femoral Attachments of Lateral Knee Structures

To compare previously described radiographic parameters for the localization of the lateral knee (LK) structures, including the popliteal tendon (Pop), anterolateral ligament (ALL), and lateral collateral ligament (LCL), to determine which method best estimates the femoral attachment of each LK structure. Twenty-nine human cadaveric knee specimens were carefully dissected to identify the LCL, ALL, and Pop. The femoral attachment for each structure was labeled with a radiopaque bead. LK radiographic images were obtained using fluoroscopy. Two radiographic approaches were used to identify each LK structure (Pop-A, Pop-B, LCL-A, LCL-B, ALL-A, and ALL-B) via previously published methods based on radiographic landmarks including the posterior femoral cortex and the Blumensaat line. The identification of radiographic landmarks was performed at 2 different time points by 2 different surgeons to determine the Pearson correlation between values, as well as interobserver and intraobserver reliability and reproducibility. The paired t test was conducted to compare the distance between the actual attachment site (as determined by the bead location) and the 2 radiographically identified estimations of attachment locations. For the LCL, the mean difference between the actual location and the estimated location via application of the LCL-B method (5.0 ± 2.4 mm) was significantly less than that estimated using the LCL-A method (8.2 ± 3.3 mm, P < .0001). Likewise, the Pop-B (5.7 ± 2.0 mm) and ALL-B (9.3 ± 4.5 mm) methods were shown to have smaller differences between the actual and estimated femoral attachment sites of the Pop insertion and ALL insertion, respectively (P < .0001). Methods for estimating the ALL femoral origin were the worst among the LK structures analyzed, with 90% of estimated values greater than 5 mm from the anatomic origin. Interobserver and intraobserver intraclass correlation coefficients were 0.785 or higher. Previously described radiographic methods for localization of the femoral attachment sites of the LK structures resulted in estimated locations that were significantly different from the locations of the radiographic beads placed at the anatomic femoral attachment sites of these structures. Therefore, radiographic methods used to localize the femoral attachments of the LK structures may not be reliable. This study shows the variability of the anatomy of the LK structures and the lack of reproducible radiographic criteria to identify these stru