3-D Morphology of the Distal Femur Viewed in Virtual Reality (AAOS 2001 annual meeting - Scientific Exhibit No. SE28) Donald G Eckhoff, MD, Denver, CO Thomas F Dwyer, MD, Denver, CO Joel M Bach, PhD, Aurora, CO Victor M Spitzer, PhD, Aurora, CO Karl D Reinig, PhD, Aurora, CO Introduction: The morphologic shape of the distal femur dictates the shape, orientation and kinematics of prosthetic replacement in total knee arthroplasty. Traditional prosthetic designs incorporate symmetric femoral condyles with a centered trochlear groove. Traditional surgical techniques center the femoral component to the distal femur and position it relative to variable bone landmarks. However, failure patterns documented in retrieval studies13,15, case series 9, and kinematic studies demonstrate how traditional design and surgical technique reflect a poor understanding of distal femoral morphology and knee kinematics. It has been shown that the flexion/extension (F/E) axis of the knee is fixed within the femur and that the articular surfaces of the condyles are circular in profile11,12. Ligament length patterns are significantly altered by abnormal axis alignment when using a hinged knee brace 14. It is expected that a malaligned femoral component would have the same effect in TKA. The purpose of this exhibit is to demonstrate with conventional images and with interactive animations in virtual reality the three-dimensional shape of the naturally asymmetric distal femur, illustrating the sulcus axis of the trochlear groove and flexion axis of the condyles relative to conventional axes (mechanical, anatomic, epicondylar & posterior condylar axes). Correlations between the morphologically determined rotation axes and experimentally determined kinematic axes are illustrated. Methods: Eighty-five mummified cadaveric knees (fig. 1 below left) were measured with a stereotactic micrometer (fig. 2 below center). The location and orientation of the sulcus were obtained by repeated horizontal passes of the stereotactic stylus over the distal femur beginning at the top of the articular surface and progressing down to the intercondylar notch (fig.3 below right). With each horizontal pass, the lowest depression of the trochlea (sulcus) was identified by the stereotactic stylus and the coordinates were recorded. After each horizontal pass, the stereotactic device was lowered by 2 mm to provide sequential horizontal tracings.           85 cadavers from the University of Colorado Department of Anthropology collection of skeletal remains were naturally mummified by burial in natron clay (fig. 1 above left). The stereotactic micrometer, originally designed to localize intra-cranial lesions in neurosurgery, was modified to hold cadaveric femora for topographical mapping of the condyles and trochlear groove (fig.2 above center). The stylus of the micrometer moves horizontally and vertically in millimeter increments to allow measures of depth in millimeters of the articular surface of the condyles and trochlea in the horizontal plane (fig.3 above right). Ten cadaveric knees were studied using a 6-degree-of-freedom motion analysis apparatus1 (fig. 4-7 below) based on a floating coordinate system10. Initial alignment of the flexion/extension (F/E) axes of the specimens with the corresponding axis of the apparatus was performed using the trans-epicondylar axis.           (click on images to enlarge)   A side view of the knee simulator with the femoral unit on the left and the tibial on the right. The knee is oriented with the patella pointed down and the tibia parallel to the horizontal. The femoral carriage swings through a vertical plane to allow flexion/extension (fig. 4 above left). An end view of the knee simulator looking at the tibial unit from distal to proximal. Movement of the tibial carriage to the left or right allows for abduction/adduction. The three large wheels in this image have been replaced by air bearings to further reduce friction. A downward motion represents anterior tibial displacement whereas upwards would be posterior (fig. 5 above, 2nd from left). Schematic representation of the knee simulator. The upper image is a top view oriented from posterior to anterior. The lower image is a side view with the tibial unit on the left and the femoral on the right (fig. 6 above 2nd from right). A side view of the knee simulator focusing on the details of the flexion/extension actuation. The hamstring muscle actuator can also be seen in the upper-right. The axial rotation actuator can be seen in the foreground to the lower left of the image (fig. 7 above right). The specimens were then passively flexed between -5� (hyperextension) and 120� under physiologic loads (fig.8 below) while six-degree-of-freedom (6-DOF) kinematics were measured.     (click on image to enlarge)   A table showing the design load capabilities of the knee simulator. The design external loads are capable of disrupting the structures of the healthy knee joint. The muscle loads approach physiologic levels (fig. 8 above). The femur position was interactively adjusted in the alignment fixture while observing the coupled abduction/adduction (A/A) and compression/distraction (C/D) motion during F/E (fig.9 below).     An image of the femoral unit of the knee simulator with the femoral alignment fixture installed. This alignment fixture allows 6-DOF adjustment of the femur with respect to the femoral unit (fig. 9 above). There is a similar alignment fixture for the tibia (not shown). Following this functional alignment 1,2, passive flexion of the specimens between -5� (hyperextension) and 120� was repeated while 6-DOF kinematics were measured. The kinematics of the two alignment conditions were compared. Seventy-four patient knees (fig.10 below left, 11 below right) and 34 volunteer knees (68 control knees) were evaluated with computed tomography (CT) 3-5,7-8.       40 patients (40 knees) with medial compartment osteoarthritis presenting for total knee arthroplasty were measured by computed tomography and compared to 40 age matched normal knees (fig. 10 above left) 34 patients (34 knees) with anterior knee pain were measured by computed tomography and compared to 34 age matched normal knees (fig. 11 above right) A CT scan of the femoral condyles, immediately proximal to the notch, and the tibial plateau, immediately proximal to the tubercle, was obtained in each knee in full extension (fig.12 below left, 13 below right). The extended knee was selected because this is the only position in which the tibia assumes a reproducible position in rotation relative to the femur. Additional CT cuts were obtained through the femoral head, femoral shaft at the lesser trochanter, and across the malleoli to determine limb rotation.         Computed tomographic scans were performed with the knee in extension by transversely "cutting" the femur proximal to the intercondylar notch and the tibia proximal to the tubercle (fig. 12 above left) The distal femora and proximal tibial computed tomography "cuts" were superimposed to measure relative translation and rotation. The proximal and distal femoral cuts were superimposed to measure femoral version (fig. 13 above right) The morphologic and biomechanical characteristics of the knee defined in these studies were measured in cyberspace and illustrated with a computer visualization program as an Interactive Anatomic Animation (IAA) in three additional cadaveric specimens. These three cadaveric knees were sectioned with CT into 0.1-1.0 mm slices, digitized, and reconstructed in virtual space (fig. 14-16 below).             The transverse image in figure 14a is one of the 1,877, 1 mm spaced, transverse images of the Visible Human Male. The collection of 1,877 images forms a volume of photographic data. This volume can be resliced at any angle to form other cross-sectional images like the coronal cross-section in figure 14b and the sagittal cross-section in figure 14c.) (fig. 14 a,b,c above)             The transverse image in figure 15a is a portion of one of the 5,189, 0.33 mm spaced, transverse images of the Visible Human Female. This transverse image is comparable in resolution to that of the male in figure 14. Like the male volume, this volume can also be "resliced" at any angle to form other cross-sectional images like the coronal cross-section in figure 15b and the sagittal cross-section in figure 15c. These "resliced" images are more than three times higher in resolution than the male counterpart because of the thinner slice thickness for the female. (fig. 15 a,b,c above)             The transverse image in figure 16a is one of the 2,500, 0.10 mm spaced, transverse images of a right knee specimen. Like the whole body volumes, this knee volume can also be "resliced" at any angle to form other cross-sectional images like the coronal cross-section in figure 16b and the sagittal cross-section in figure 16c. These "resliced" images are more than 90 times higher in resolution than the male and nearly 30 times higher in resolution than the female counterparts because of thinner slice thickness and higher "in-plane" resolution. (fig. 16 a,b,c above) Computer generated cylinders were "grown" within the confines of the articular surface of the distal femur to confirm and illustrate the cylindrical geometry of the condyles as well as demonstrate the position of the cylindrical axis relative to conventional axes (mechanical, anatomic, epicondylar & posterior condylar axes). (fig. 17 below)     The femur of the Visible Human Male has been extracted from the volumetric computed tomography data represented in figure 14. A cylinder was then fit to the osseous surface of the distal femur. Note that the radius of the cylinder fit to the medial femoral condyle is slightly larger in radius compared to the cylinder fit to the lateral femoral condyle but the center of each cylinder lies on a single axis. (fig. 17 above) Results: The sulcus of the trochlear groove lies lateral to the mid-plane and is oriented between the mechanical and anatomic axes of the femur (fig. 18 below left,19 below right).     (click on images to enlarge)   The sulcus (lowest point) is a near-linear depression in the trochlear groove that lies lateral to the midplane defined as the plane perpendicular to the posterior condylar axis. (fig. 18 above left) The sulcus is oriented between the traditional mechanical axis (line joining center of femoral head and center of knee) and anatomic axis (center of femoral shaft). (fig. 19 above right) The cross-sectional center of the femur lies medial and anterior to the cross-sectional center of the tibia (fig.20 below).     (click on images to enlarge)   The cross-sectional centers of the distal femur and proximal tibia are not superimposed, but are translated 4+6 mm anteroposterior and 5+4 mm mediolateral in both normal and pathologic knees (ostheoarthritic and anterior knee pain). (fig. 20 above) The tibia is not rotated relative to the femur in the normal knee but is externally rotated in the osteoarthritic knee and the knee with anterior pain (fig. 21 below left). Femoral version is decreased in the osteoarthritic knee and increased in the knee with anterior pain when compared to the normal knee (fig. 22 below center, 23 below right).     (click on images to enlarge)   In the extended pathologic knee, there is a rotation of the tibia to the femur fixed in soft-tissue, a "rotation" contracture. The tibia is externally rotated to the femur in both the osteoarthritic knee (4+1 degrees) and the knee with anterior knee pain (7+1 degrees). There is no rotation of the tibia to the femur or rotation contracture in the normal knee. (fig. 21 above left) The distal femur is malrotated relative to the proximal femur in the osteoarthritic knee and the knee with anterior pain when compared to the normal knee. (fig. 22 above center) The distal femur of the knee with anterior pain viewed here by computed tomography is rotated when the proximal femur is normally oriented to the acetabulum. (fig. 23 above right) Kinematic analysis demonstrated that when the coupled A/A motion was eliminated, the functional F/E axis of the femur was parallel to the F/E axis of the apparatus. Subsequently, when the coupled C/D motion was eliminated, the functional F/E axis of the femur was coincident with the F/E axis of the apparatus. Variability was extremely small (fig. 24 below).     (click on image to enlarge)   Experimental data for 8 cadaveric knees. This data shows the coupled axial rotation of the tibial during flexion/extension of the femur following functional alignment. Note the extremely small variability attributable to the alignment procedures. (fig. 24 above) The medial and lateral femoral condyles geometrically approximate cylinders. Although the medial condyle has a larger radius than the lateral condyle, both cylinders are oriented around a common axis. This common "cylindrical" axis represents a single, fixed, flexion/extension axis in the human knee, which is distinct from other conventional axes (mechanical, anatomic, epicondylar & posterior condylar axes). (fig. 25 below)     One view of the Interactive Animation of the distal femur demonstrated in this exhibit illustrates the 3-D relationship between the epicondylar axis (green line) and the "cylindrical axis" (red line) defined by the center of the cylinders which most closely reproduce the geometry of the condyles. (fig. 25 above) Discussion and Conclusion: This study carefully documents the asymmetric morphologic features of the distal femur and correlates these features with the kinematics of the knee. The location and orientation of the femoral sulcus is lateral to the midplane between the femoral condyles and oriented between the anatomic and mechanical axes of the femur. The center of the femur in cross-section is off-set, medial and anterior, to the center of the tibia. These asymmetric, off-set morphologic features of distal femur should be incorporated in the design (fig. 26 below) and positioning of prosthetic replacements in the knee.     A femoral component designed with asymmetrical features (patents #5,681,354 & #5,728,162) addresses the natural asymmetry of human knee. (fig. 26 above) This study further documents the asymmetric cylindrical shape of the condyles and establishes the cylindrical axis of rotation of the condyles about which the tibia rotates. The measured kinematic data supports the morphologic findings for a fixed, cylindrical flexion/extension axis of rotation. This study provides kinematic and morphologic validation for a single cylindrical flexion/extension axis of the knee. This scientific exhibit is the first to illustrate these observations in 3-D stereo using interactive animations and virtual reality. (This Project has been funded in whole or in part with Federal funds from the National Library of Medicine under Contract No. N01-LM-0-3507.) References 1) Bach JM, Hull ML. A New Load Application System for In Vitro Study of Ligamentous Injuries to the Human Knee Joint. J Biomechanical Eng. 1995;117:373-382. 2) Berns GS, Hull ML, Patterson HA. Implementation of a Five Degree of Freedom Automated System to Determine Knee Flexibility In Vitro. J Biomechanical Eng. 1990;112:392-400. 3) Eckhoff DG, Aukerman R. Femorotibial Offset. Clin Orthop. 1999;368:162-165. 4) Eckhoff DG, Brown AW, Kilcoyne RF, Stamm ER. Knee Version Associated with Anterior Knee Pain. Clin Orthop. 1997;339:152-155. 5) Eckhoff DG, Montgomery WK, Stamm ER, Kilcoyne RF. Location of the Femoral Sulcus in the Osteoarthritic Knee. J Arthrop. 1996;11:163165. 6) Eckhoff DG, Burke BJ, Dwyer TF, Pring ME, Spitzer VM, VanGerwen DP. Sulcus Morphology of the Distal Femur. Clin Orthop. 1996;331:23-28. 7) Eckhoff DG, Montgomery WK, Kilcoyne RF, Stamm ER. Femoral Morphometry and Anterior Knee Pain. Clin Orthop. 1994;302:64-68. 8) Eckhoff DG, Johnson RJ, Stamm ER, Kilcoyne RF, Wiedel JD. Version of the Osteoarthritic Knee. J Arthrop. 1994;9:73-79. 9) Figgie HE, Goldberg VM, Figgie MP, Inglis AE, Kellly M, Sobe M. The Effect of Alignment of the Implant on Fractures of the Patella after Condylar Total Knee Arthroplasty. J Bone Joint Surg Am. 1989;71:1031-1039. 10) Grood ES, Suntay WJ. A Joint Coordinate System for the Clinical Description of Three-Dimensional Motions: Application to the Knee. J Biomed Eng. 1983;105:136-144. 11) Hollerbach K, Hollister AM. Mathematical Analysis of Errors Inherent in Using Euler Angles to Represent Joint Motion. Proceedings Int Soc Biomech. 1995. 12) Hollister AM, Jatana S, Singh AK, Sullivan WW, Lupichuk AG. The Axes of Rotation of the Knee. Clin Orthop. 1994;290:259-68. 13) Lewis P, Rorabeck DH, Bourne RB, Devane P. Posteromedial Tibial Polyethylene Failure in Total Knee Replacements. Clin Orthop. 1994;299:11-17. 14) Walker PS, Rovick JS, Robertson DD. The Effects Of Knee Brace Hinge Design And Placement On Joint Mechanics. J Biomech. 1988;21:965-74. 15) Wasielewski RC, Galante JO, Leighty RM, Natarajan RN, Rosenberg AG. Wear Patterns on Retrieved Polyethylene Tibial Inserts and their Relationship to Technical Considerations during Total Knee Arthroplasty. Clin Orthop. 1994;299:31-43.