- Email: [email protected]
3D printed guides for controlled alignment in biomechanics tests
Journal of Biomechanics 49 (2016) 484–487
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Short communication
3D printed guides for controlled alignment in biomechanics tests Matthias A. Verstraete n, Laurent Willemot, Stefaan Van Onsem, Cyriëlle Stevens, Nele Arnout, Jan Victor Ghent University, Department of Physical Medicine and Orthopaedic Surgery, De Pintelaan 185, 9000 Gent, Belgium
art ic l e i nf o
a b s t r a c t
Article history: Accepted 18 December 2015
The bone-machine interface is a vital first step for biomechanical testing. It remains challenging to restore the original alignment of the specimen with respect to the test setup. To overcome this issue, we developed a methodology based on virtual planning and 3D printing. In this paper, the methodology is outlined and a proof of concept is presented based on a series of cadaveric tests performed on our knee simulator. The tests described in this paper reached an accuracy within 3–4° and 3–4 mm with respect to the virtual planning. It is however the authors' belief that the method has the potential to achieve an accuracy within one degree and one millimeter. Therefore, this approach can aid in reducing the imprecisions in biomechanical tests (e.g. knee simulator tests for evaluating knee kinematics) and improve the consistency of the bone-machine interface. & 2016 Elsevier Ltd. All rights reserved.
Keywords: Bone-machine interface 3D printing Virtual planning Knee simulator
1. Introduction
2. Methods
In the setting of biomechanical research, using cadaveric specimen it is customary to connect the tissue to the testing apparatus via a process called potting (Robinson et al., 2005; Victor et al., 2009a, 2009b; Wünschel et al., 2010). The specimen, usually a bony structure, is thereby encased in a container that can securely be mounted on the testing device of choice. The aim is to optimize the tissue-machine interface in order to create a safe and reproducible experimental environment. This includes minimizing the risk of loosening and breakage under applied loads, but should also allow for accurate specimen positioning with respect to the testing device. Specimen positioning is often rudimentary or simply done under visual control (Hull et al., 1995; Victor et al., 2009a, 2009b). Incorrect alignment of the tissue in the potting container is difficult to correct during testing and can adversely affect the interpretation of the results. In this paper we first present a theoretical framework that makes use of our new custom made 3D printed-guides in the potting of a distal femur and proximal tibia for testing in a knee simulator. In a second phase, the accuracy of the presented method is assessed by comparing the target values with the actual situations for a set of cadaveric experiments.
2.1. Boundary conditions
n
Corresponding author. Tel.: þ 32 9 332 11 85; fax: þ32 9 332 49 75. E-mail address: [email protected] (M.A. Verstraete).
http://dx.doi.org/10.1016/j.jbiomech.2015.12.036 0021-9290/& 2016 Elsevier Ltd. All rights reserved.
The focus of this paper is on the position of the distal femur and the proximal tibia relative to a knee simulator. The targeted knee simulator provides a single, rotational degree of freedom at the hip and four degrees of freedom at the ankle (Fig. 1) (Verstraete and Victor, 2015). Focusing on the hip first, the rotational freedom is created through a hinge joint between the simulator and the container holding the femur. To restore the specimen's specific kinematics, the femoral hip center should therefore coincide with this hinge axis (Fig. 2a). This represents a first boundary condition for the femur. Second, to control the direction of the applied loads in the knee simulator (quadriceps and hamstring muscle load), the femoral mechanical axis should be perpendicular to the hinge axis. Third, an excessive displacement of the femur along the hinge axis of the simulator should be minimized throughout flexion-extension. Therefore, the rotation along the femoral mechanical axis is controlled by prescribing parallelism between the hinge axis of the knee simulator and the posterior points of the medial and lateral condyle (also referred to as posterior condylar line) in the axial plane. Second, the four degrees of freedom of the tibia provide full rotational freedom and unrestrained translations in the mediolateral direction. This requires that the center of the ankle joint coincides with the rotation point of the mechanical joint. The latter is located on the mediolateral translational axis. The former implies two distinct requirements. On one hand, the distance to the center of the ankle should be mimicked in the simulator, resulting in the first boundary condition. On the other hand, the mechanical axis of the tibia ideally coincides with the rotational axis of the mechanical ankle joint (Fig. 2b). This is quantitatively assessed by means of two measures. The first measure quantifies the degree of parallelism between the mechanical axis of the tibia and the simulator axis and thereby provides the second angular boundary condition. The second measure describes the distance between both axes at given position along the mechanical axis and provides a linear measure that represents the third boundary condition. For a detailed description of the design of the 3D printed guides resulting from these boundary conditions, the reader is referred to Appendix A.
M.A. Verstraete et al. / Journal of Biomechanics 49 (2016) 484–487
485
Fig. 1. Schematic representation of knee simulator (a) and picture taken during cadaveric experiment (b) (Verstraete and Victor, 2015).
3. Results For the femur, following the reconstruction of the mechanical axis and posterior condylar line, the effectiveness of the femoral guides is quantitatively assessed. A more in depth description of the analysis approach is described in Appendix B. To that extent, the following conditions are evaluated by fitting planes onto the bounding surfaces of the PU resin block in which the femur is embedded: 1. Mechanical axis perpendicular to proximal and distal plane 2. Posterior condylar line parallel to anterior and posterior plane 3. Distance from proximal plane to hip center with respect to target distance (85 mm) The averages and extremes measured for these variables are presented in Table 1. With respect to the first condition, an angular accuracy of 0.95° is obtained. However, if the difference between theoretically parallel planes (i.e. distal and proximal plane) is evaluated, it becomes clear that this difference is of the same order of magnitude (i.c. 0.90°). As a result, the obtained accuracy is expected to be within the accuracy range of the considered measurement platform. The second condition has a greater deviation, an average difference of 3.78° is observed between the actual and target value, ranging from 3.20° to 4.35°. From the evaluation of the third condition, it becomes clear that, on average, the target distance is realized with an accuracy of72.74 mm. Analogous to the femur, the effectiveness of the guides in restoring tibial alignment is evaluated following the reconstruction of the mechanical axis and quantifying the following three conditions considered during embedding:
Fig. 2. Schematic of boundary conditions for femoral (a) and tibial (b) fixation. 2.2. Embedding of specimen Three fresh frozen human cadaveric knees from donors with no evidence of prior injury and a mean age of 72 year (2 male/1 female) were obtained from a tissue bank after ethical approval was given by the local ethics committee (EC/ 2014/0847). Before testing, the specimens were thawed for 24 h. All soft tissues were removed from the proximal end of the femur and distal end of the tibia to allow fitting the alignment guides around the corpus femoris and tibiae. The distance measured between both guides as well as the distance between the hip and ankle center thereby served as a guide. After finding the most suitable position of the guides, the bone was resected to its final length, thereby assuring a correct position of the hip and ankle center in the proximal/distal direction, when placed inside the container. At the same time, the fibular head is fused to the tibia using a screw and the distal part of the fibular diaphysis is resected to avoid interference with the alignment guides.
1. Parallelism between mechanical axis of the tibia and cylinder axis of the container (rotational alignment). 2. Centricity of the mechanical axis with respect to the container axis (linear distance at the level of the proximal and distal plane). 3. Target cutting distance from distal plane to center of the ankle (90 mm). The measured values are presented in Table 2. With respect to the first condition, the angular accuracy obtained following this procedure equals 0.72° on average. The second condition is expressed in millimeter distance between the two intersection points of mechanical and cylinder axis, on average this distance equals 0.75 mm. Regarding the third condition, the target distance is realized with an accuracy of approximately 5 mm. However, a large variation is observed, with actual distances ranging between 84.47 and 95.13 mm.
486
M.A. Verstraete et al. / Journal of Biomechanics 49 (2016) 484–487
4. Discussion and conclusion This study presents a proof of concept, indicating the ability to restore the anatomical alignment in a knee simulator with a reasonable accuracy. The method presented here avoids the need for experience and/or arbitrary feeling in the process of positioning and potting a bone in a container for mechanical testing. In this case, the absence of direct visual control over the bones, due to the soft tissue envelope, would have further complicated an intuitive, non-guided positioning of the bones in their respective containers. It is the authors' belief that the performance of the proposed methodology has therefore been evaluated under realistic conditions. To position the guides on the bones, all soft tissue was removed around the location of the guides. Some soft tissue remained however, complicating the positioning in some cases. The inaccuracy in rotational alignment, in particular with respect to the posterior condylar line of the femur, is attributed to the above. More specifically, the femoral guides were slightly undersized to assure fitting in the container and accommodate some Table 1 Evaluation of accuracy for presented series of specimens. Condition 1
2
3
Angle between mechanical axis Angle between mechanical axis and normal to distal plane 0.91° and normal to proximal plane (0.46–1.33°) 0.98° (0.56–1.61°) Angle between posterior conAngle between posterior condylar line and medial plane 3.89° dylar line and lateral plane 3.66° (3.20–4.35°) (3.63–4.32°) Error on distance hip center to proximal plane 2.74 mm (1.85 mm– 3.34 mm)
Table 2 Evaluation of accuracy for presented case study (tibia). Condition 1
2
3
Angle between mechanical axis and cylinder axis 0.72° (0.61– 0.87°) Distance between intersection point mechanical axis and cylinder axis proximal plane 0.91 mm (0.39–1.37 mm) Error on distance distal plane to 5.53 mm)
Distance between intersection point mechanical axis and cylinder axis distal plane 0.60 mm (0.41–0.77 mm) center ankle 4.68 mm (3.35–
remaining soft tissue. The undersizing thereby created limited rotational freedom around the longitudinal axis of the femur. It is hypothesized that further control of the soft tissue removal in combination with an improved matching between the guides and the container has the potential to reduce this inaccuracy. The location of the cutting plane along the mechanical axis also demonstrated an inferior accuracy, with large scatter around the target values. This is attributed to the relatively constant cross section of the corpus femoris and tibiae around the location of the alignment guides. This constant cross section complicates the unique positioning of the alignment guides along the bones' longitudinal axis, compromising the accuracy in determining the cutting plane. To overcome this issue, the authors’ propose the use of specimen-specific 3D-printed cutting guides. In the particular case of a femur, for example, this cutting guide could fit the trochanter minor and corpus femoris to assure the positioning of the cut. The mechanical axis of the femur defines the normal to the cutting plane. The distance along the mechanical axis is calculated to match the specific knee simulator that is targeted in our test setup. Note additionally that the thickness of the sawing blade (i.e. 1.3 mm) was taken into account. The use of such a cutting block is schematically illustrated in Fig. 3. Although the use of a sawbone specimen introduces significant simplifications (i.e. no soft tissue remains), the cutting guide clearly improved the accuracy (actual value measured on CT of 84.19 mm, target of 85.00 mm). Notwithstanding this result represents a single case, the use of such guides appears therefore advisable to determine the location of the cutting plane. However, this approach has some limitations. Although the accuracy might benefit from a tighter design, it should be noted that an overly tight design risks being unmountable. The latter would compromise the use of these guides as it is impossible to adjust or replace these guides during testing (e.g. too long printing time, approx. 3 h for set of femur/tibia guides). This clearly points out a more general limitation, inherent to the use of 3D printed guides in combination with a pre-operative planning In conclusion, no standardized method was currently available for the specific application of connecting a cadaveric leg to a knee simulator. We believe that through this proposed potting process, the variability in the measured kinematic patterns can be reduced, as the position of the specimen and the direction of the applied loads is more reliably controlled between different specimens, notwithstanding the variability that is still observed in the measured accuracies.
Fig. 3. Illustration of the use of cutting guides on sawbone specimen with resection guide attached to femur and sawing blade positioned on top of sawing guide in left image and axial view on resected femoral head with resection guide attached in right image.
M.A. Verstraete et al. / Journal of Biomechanics 49 (2016) 484–487
Conflict of interest statement None.
Acknowledgments The authors would like to acknowledge the financial support of BIOMET who supported this research through an unrestricted research grant (Grant number: FWUGent_GE_022).
Appendix A. Supplementary information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jbiomech.2015.12.036.
487
References Hull, M.L., Berns, G.S., et al., 1995. Strain in the medial collateral ligament of the human knee under single and combined loads. J. Biomech. 29 (2), 199–206. Robinson, J.R., Bull, A.M.J., et al., 2005. Structural properties of the medial collateral ligament complex of the human knee. J. Biomech. 38 (5), 1067–1074. Verstraete, M.A., Victor, J., 2015. Possibilities and limitations of novel in-vitro knee simulator. J. Boimech. 48 (12), 3377–3382. http://dx.doi.org/10.1016/j. jbiomech.2015.06.007. Victor, J., Labey, L., et al., 2009a. The influence of muscle action on tibiofemoral knee kinematics. J. Orthop. Res. 28 (4), 419–428. Victor, J., Van Doninck, D., et al., 2009b. How precise can bony landmarks be determined on a CT scan of the knee? Knee 16 (5), 358–365. Wünschel, M., Muller, O., et al., 2010. The anterior cruciate ligament provides resistance to externally applied anterior tibial force but not to internatl rotational torque during simulated weight-bearing flexion. Arthroscopy 26 (11), 1520–1527.
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
Short communication
3D printed guides for controlled alignment in biomechanics tests Matthias A. Verstraete n, Laurent Willemot, Stefaan Van Onsem, Cyriëlle Stevens, Nele Arnout, Jan Victor Ghent University, Department of Physical Medicine and Orthopaedic Surgery, De Pintelaan 185, 9000 Gent, Belgium
art ic l e i nf o
a b s t r a c t
Article history: Accepted 18 December 2015
The bone-machine interface is a vital first step for biomechanical testing. It remains challenging to restore the original alignment of the specimen with respect to the test setup. To overcome this issue, we developed a methodology based on virtual planning and 3D printing. In this paper, the methodology is outlined and a proof of concept is presented based on a series of cadaveric tests performed on our knee simulator. The tests described in this paper reached an accuracy within 3–4° and 3–4 mm with respect to the virtual planning. It is however the authors' belief that the method has the potential to achieve an accuracy within one degree and one millimeter. Therefore, this approach can aid in reducing the imprecisions in biomechanical tests (e.g. knee simulator tests for evaluating knee kinematics) and improve the consistency of the bone-machine interface. & 2016 Elsevier Ltd. All rights reserved.
Keywords: Bone-machine interface 3D printing Virtual planning Knee simulator
1. Introduction
2. Methods
In the setting of biomechanical research, using cadaveric specimen it is customary to connect the tissue to the testing apparatus via a process called potting (Robinson et al., 2005; Victor et al., 2009a, 2009b; Wünschel et al., 2010). The specimen, usually a bony structure, is thereby encased in a container that can securely be mounted on the testing device of choice. The aim is to optimize the tissue-machine interface in order to create a safe and reproducible experimental environment. This includes minimizing the risk of loosening and breakage under applied loads, but should also allow for accurate specimen positioning with respect to the testing device. Specimen positioning is often rudimentary or simply done under visual control (Hull et al., 1995; Victor et al., 2009a, 2009b). Incorrect alignment of the tissue in the potting container is difficult to correct during testing and can adversely affect the interpretation of the results. In this paper we first present a theoretical framework that makes use of our new custom made 3D printed-guides in the potting of a distal femur and proximal tibia for testing in a knee simulator. In a second phase, the accuracy of the presented method is assessed by comparing the target values with the actual situations for a set of cadaveric experiments.
2.1. Boundary conditions
n
Corresponding author. Tel.: þ 32 9 332 11 85; fax: þ32 9 332 49 75. E-mail address: [email protected] (M.A. Verstraete).
http://dx.doi.org/10.1016/j.jbiomech.2015.12.036 0021-9290/& 2016 Elsevier Ltd. All rights reserved.
The focus of this paper is on the position of the distal femur and the proximal tibia relative to a knee simulator. The targeted knee simulator provides a single, rotational degree of freedom at the hip and four degrees of freedom at the ankle (Fig. 1) (Verstraete and Victor, 2015). Focusing on the hip first, the rotational freedom is created through a hinge joint between the simulator and the container holding the femur. To restore the specimen's specific kinematics, the femoral hip center should therefore coincide with this hinge axis (Fig. 2a). This represents a first boundary condition for the femur. Second, to control the direction of the applied loads in the knee simulator (quadriceps and hamstring muscle load), the femoral mechanical axis should be perpendicular to the hinge axis. Third, an excessive displacement of the femur along the hinge axis of the simulator should be minimized throughout flexion-extension. Therefore, the rotation along the femoral mechanical axis is controlled by prescribing parallelism between the hinge axis of the knee simulator and the posterior points of the medial and lateral condyle (also referred to as posterior condylar line) in the axial plane. Second, the four degrees of freedom of the tibia provide full rotational freedom and unrestrained translations in the mediolateral direction. This requires that the center of the ankle joint coincides with the rotation point of the mechanical joint. The latter is located on the mediolateral translational axis. The former implies two distinct requirements. On one hand, the distance to the center of the ankle should be mimicked in the simulator, resulting in the first boundary condition. On the other hand, the mechanical axis of the tibia ideally coincides with the rotational axis of the mechanical ankle joint (Fig. 2b). This is quantitatively assessed by means of two measures. The first measure quantifies the degree of parallelism between the mechanical axis of the tibia and the simulator axis and thereby provides the second angular boundary condition. The second measure describes the distance between both axes at given position along the mechanical axis and provides a linear measure that represents the third boundary condition. For a detailed description of the design of the 3D printed guides resulting from these boundary conditions, the reader is referred to Appendix A.
M.A. Verstraete et al. / Journal of Biomechanics 49 (2016) 484–487
485
Fig. 1. Schematic representation of knee simulator (a) and picture taken during cadaveric experiment (b) (Verstraete and Victor, 2015).
3. Results For the femur, following the reconstruction of the mechanical axis and posterior condylar line, the effectiveness of the femoral guides is quantitatively assessed. A more in depth description of the analysis approach is described in Appendix B. To that extent, the following conditions are evaluated by fitting planes onto the bounding surfaces of the PU resin block in which the femur is embedded: 1. Mechanical axis perpendicular to proximal and distal plane 2. Posterior condylar line parallel to anterior and posterior plane 3. Distance from proximal plane to hip center with respect to target distance (85 mm) The averages and extremes measured for these variables are presented in Table 1. With respect to the first condition, an angular accuracy of 0.95° is obtained. However, if the difference between theoretically parallel planes (i.e. distal and proximal plane) is evaluated, it becomes clear that this difference is of the same order of magnitude (i.c. 0.90°). As a result, the obtained accuracy is expected to be within the accuracy range of the considered measurement platform. The second condition has a greater deviation, an average difference of 3.78° is observed between the actual and target value, ranging from 3.20° to 4.35°. From the evaluation of the third condition, it becomes clear that, on average, the target distance is realized with an accuracy of72.74 mm. Analogous to the femur, the effectiveness of the guides in restoring tibial alignment is evaluated following the reconstruction of the mechanical axis and quantifying the following three conditions considered during embedding:
Fig. 2. Schematic of boundary conditions for femoral (a) and tibial (b) fixation. 2.2. Embedding of specimen Three fresh frozen human cadaveric knees from donors with no evidence of prior injury and a mean age of 72 year (2 male/1 female) were obtained from a tissue bank after ethical approval was given by the local ethics committee (EC/ 2014/0847). Before testing, the specimens were thawed for 24 h. All soft tissues were removed from the proximal end of the femur and distal end of the tibia to allow fitting the alignment guides around the corpus femoris and tibiae. The distance measured between both guides as well as the distance between the hip and ankle center thereby served as a guide. After finding the most suitable position of the guides, the bone was resected to its final length, thereby assuring a correct position of the hip and ankle center in the proximal/distal direction, when placed inside the container. At the same time, the fibular head is fused to the tibia using a screw and the distal part of the fibular diaphysis is resected to avoid interference with the alignment guides.
1. Parallelism between mechanical axis of the tibia and cylinder axis of the container (rotational alignment). 2. Centricity of the mechanical axis with respect to the container axis (linear distance at the level of the proximal and distal plane). 3. Target cutting distance from distal plane to center of the ankle (90 mm). The measured values are presented in Table 2. With respect to the first condition, the angular accuracy obtained following this procedure equals 0.72° on average. The second condition is expressed in millimeter distance between the two intersection points of mechanical and cylinder axis, on average this distance equals 0.75 mm. Regarding the third condition, the target distance is realized with an accuracy of approximately 5 mm. However, a large variation is observed, with actual distances ranging between 84.47 and 95.13 mm.
486
M.A. Verstraete et al. / Journal of Biomechanics 49 (2016) 484–487
4. Discussion and conclusion This study presents a proof of concept, indicating the ability to restore the anatomical alignment in a knee simulator with a reasonable accuracy. The method presented here avoids the need for experience and/or arbitrary feeling in the process of positioning and potting a bone in a container for mechanical testing. In this case, the absence of direct visual control over the bones, due to the soft tissue envelope, would have further complicated an intuitive, non-guided positioning of the bones in their respective containers. It is the authors' belief that the performance of the proposed methodology has therefore been evaluated under realistic conditions. To position the guides on the bones, all soft tissue was removed around the location of the guides. Some soft tissue remained however, complicating the positioning in some cases. The inaccuracy in rotational alignment, in particular with respect to the posterior condylar line of the femur, is attributed to the above. More specifically, the femoral guides were slightly undersized to assure fitting in the container and accommodate some Table 1 Evaluation of accuracy for presented series of specimens. Condition 1
2
3
Angle between mechanical axis Angle between mechanical axis and normal to distal plane 0.91° and normal to proximal plane (0.46–1.33°) 0.98° (0.56–1.61°) Angle between posterior conAngle between posterior condylar line and medial plane 3.89° dylar line and lateral plane 3.66° (3.20–4.35°) (3.63–4.32°) Error on distance hip center to proximal plane 2.74 mm (1.85 mm– 3.34 mm)
Table 2 Evaluation of accuracy for presented case study (tibia). Condition 1
2
3
Angle between mechanical axis and cylinder axis 0.72° (0.61– 0.87°) Distance between intersection point mechanical axis and cylinder axis proximal plane 0.91 mm (0.39–1.37 mm) Error on distance distal plane to 5.53 mm)
Distance between intersection point mechanical axis and cylinder axis distal plane 0.60 mm (0.41–0.77 mm) center ankle 4.68 mm (3.35–
remaining soft tissue. The undersizing thereby created limited rotational freedom around the longitudinal axis of the femur. It is hypothesized that further control of the soft tissue removal in combination with an improved matching between the guides and the container has the potential to reduce this inaccuracy. The location of the cutting plane along the mechanical axis also demonstrated an inferior accuracy, with large scatter around the target values. This is attributed to the relatively constant cross section of the corpus femoris and tibiae around the location of the alignment guides. This constant cross section complicates the unique positioning of the alignment guides along the bones' longitudinal axis, compromising the accuracy in determining the cutting plane. To overcome this issue, the authors’ propose the use of specimen-specific 3D-printed cutting guides. In the particular case of a femur, for example, this cutting guide could fit the trochanter minor and corpus femoris to assure the positioning of the cut. The mechanical axis of the femur defines the normal to the cutting plane. The distance along the mechanical axis is calculated to match the specific knee simulator that is targeted in our test setup. Note additionally that the thickness of the sawing blade (i.e. 1.3 mm) was taken into account. The use of such a cutting block is schematically illustrated in Fig. 3. Although the use of a sawbone specimen introduces significant simplifications (i.e. no soft tissue remains), the cutting guide clearly improved the accuracy (actual value measured on CT of 84.19 mm, target of 85.00 mm). Notwithstanding this result represents a single case, the use of such guides appears therefore advisable to determine the location of the cutting plane. However, this approach has some limitations. Although the accuracy might benefit from a tighter design, it should be noted that an overly tight design risks being unmountable. The latter would compromise the use of these guides as it is impossible to adjust or replace these guides during testing (e.g. too long printing time, approx. 3 h for set of femur/tibia guides). This clearly points out a more general limitation, inherent to the use of 3D printed guides in combination with a pre-operative planning In conclusion, no standardized method was currently available for the specific application of connecting a cadaveric leg to a knee simulator. We believe that through this proposed potting process, the variability in the measured kinematic patterns can be reduced, as the position of the specimen and the direction of the applied loads is more reliably controlled between different specimens, notwithstanding the variability that is still observed in the measured accuracies.
Fig. 3. Illustration of the use of cutting guides on sawbone specimen with resection guide attached to femur and sawing blade positioned on top of sawing guide in left image and axial view on resected femoral head with resection guide attached in right image.
M.A. Verstraete et al. / Journal of Biomechanics 49 (2016) 484–487
Conflict of interest statement None.
Acknowledgments The authors would like to acknowledge the financial support of BIOMET who supported this research through an unrestricted research grant (Grant number: FWUGent_GE_022).
Appendix A. Supplementary information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jbiomech.2015.12.036.
487
References Hull, M.L., Berns, G.S., et al., 1995. Strain in the medial collateral ligament of the human knee under single and combined loads. J. Biomech. 29 (2), 199–206. Robinson, J.R., Bull, A.M.J., et al., 2005. Structural properties of the medial collateral ligament complex of the human knee. J. Biomech. 38 (5), 1067–1074. Verstraete, M.A., Victor, J., 2015. Possibilities and limitations of novel in-vitro knee simulator. J. Boimech. 48 (12), 3377–3382. http://dx.doi.org/10.1016/j. jbiomech.2015.06.007. Victor, J., Labey, L., et al., 2009a. The influence of muscle action on tibiofemoral knee kinematics. J. Orthop. Res. 28 (4), 419–428. Victor, J., Van Doninck, D., et al., 2009b. How precise can bony landmarks be determined on a CT scan of the knee? Knee 16 (5), 358–365. Wünschel, M., Muller, O., et al., 2010. The anterior cruciate ligament provides resistance to externally applied anterior tibial force but not to internatl rotational torque during simulated weight-bearing flexion. Arthroscopy 26 (11), 1520–1527.