Nuclear Magnetic Resonance

Eight healthy male subjects (mean age, 30 ± 7 years; range, 23–41 years) were evaluated using an IRB approved protocol. All volunteers had no previous history of lower extremity injury or surgery prior to completing the test protocol.
One knee of each subject was imaged using a 3T MR scanner (Trio Tim, Siemens Medical Solutions USA, Malvern, PA) at the Center for Advanced Magnetic Resonance Development at Duke University. Coronal, sagittal, and axial images were acquired from the subjects while lying in a supine position using a double-echo steady-state sequence (DESS) and an eight-channel receive-only knee coil with a field of view of 15 × 15 cm², a matrix of 512 × 512 pixels², and slice thickness of 1 mm (flip angle: 25°; repetition time: 17 ms; echo time: 6 ms). From the three views of the MRI scans, outlines of the femur and tibia were segmented using solid-modeling software (Rhinoceros 4.0, Robert McNeel and Associates, Seattle, WA), as described in previous studies.^{2 (link),56 (link)} Additionally, the attachment site of the ACL was outlined in the three planes of view. Knowing the voxel size, these outlines were then used to create 3D models of the distal femur and proximal tibia, as well as the footprints of the ACL on each. Orthogonal image sets were used to confirm the shape and position of the ACL. The ACL footprint was further divided into anteromedial (AM) and posterolateral (PL) bundles,^{28 (link)} as described previously in the literature (Fig. 1).^{33 (link),38 (link)} A previous validation study has shown that this methodology can locate the center of the ACL footprint to within 0.3 mm.^{2 (link)} Based on a previous parametric study,^{38 (link)} we expect this relatively small difference to have minimal effect on our results.
Following MRI, each subject’s knee was imaged while standing on a level platform from orthogonal directions using fluoroscopes (BV Pulsera, Philips, The Netherlands).^{33 (link)} Each fluoroscopic image had a resolution of 1024 × 1024 pixels². The protocol consisted of the following single-legged static knee positions (Fig. 2): full extension, 30° of flexion, and 30° of flexion with 10° of external rotation of the tibia and maximal internal rotation at the hip to simulate a valgus collapse position.^{36 (link),46 (link),52 (link),60 (link)} For each pose, subjects were guided on how to position their knees by one investigator using a goniometer.
To create the in vivo joint model (Fig. 2), the orthogonal images were imported into the solid-modeling software in order to recreate the biplanar fluoroscopic system used during testing.^{1 (link),12 (link)} Next, the 3D MR knee model was imported into the virtual fluoroscopic environment. Using custom-written edge detection software as a modeling aid to highlight the bone contours on the fluoroscopic images,^{1 (link),12 (link)} the bones were moved individually in six degrees of freedom until their projections matched the bony outlines in the two orthogonal planes when viewed from the x-ray sources. Previous validation studies have shown that this approach can reproduce joint motion to within 0.1 mm and 0.3°.^{12 (link),15 (link)}From these 3D models, knee joint kinematics and the length of the ACL and its functional bundles were measured. First, a coordinate system was drawn on each knee model.^{15 (link)} The long axis of the tibia was determined by fitting a cylinder to the tibial shaft. Next, a mediolateral axis was drawn perpendicular to the long axis of the tibia and tangent to the posterior extremes of the tibial plateau. Finally, the anteroposterior axis was drawn orthogonal to the long and mediolateral axes of the tibia. On the femur, the long axis was determined by fitting a cylinder to the femoral shaft. The femoral coordinate system consisted of this proximodistal axis and an axis through the transepicondylar line. The kinematic measures examined by this study included flexion, internal/external rotation, and varus/valgus angle.^{27 (link)} The transepicondylar line was used as a flexion/extension rotational axis. The internal/external rotation of the tibia was measured as the angle between the mediolateral axis of the tibia and the transepicondylar line projected on to the tibial plateau. Varus/valgus angle was measured as the change in angle between the long axis of tibia and transepicondylar line of the femur (Fig. 3). However, varus/valgus calculated this way is different from valgus measurements made by various videographic studies.^{11 (link),46 (link)} Therefore, we used the coronal plane angle to approximate these measurements of valgus when viewed from a broad perspective outside the knee. Coronal plane angle was defined as the angle between the long axis of the femur and the long axis of the tibia projected on the tibial coronal plane (Fig. 3). ACL and bundle lengths were calculated as the distance between the area centroids of the femoral and tibial ACL attachment sites.^{1 (link),56 (link)}Repeated measures ANOVA and Student–Newman–Keuls post hoc tests were used to detect statistically significant differences in flexion angle, as well as the lengths of the ACL and its functional bundles at each of the three knee positions. In addition, a two-way repeated measures ANOVA was used to detect differences between the coronal plane and varus/valgus angles in each knee position. Differences were considered statistically significant where p < 0.05.