The objective of this study was to evaluate the sensitivity of CGM to marker placement. We simulated different combinations of markers on the left lower limb, moved in different directions by 10 mm. Overall, measurements in the transverse plane demonstrated the greatest sensitivity to marker displacement, while markers moved in the sagittal plane resulted in the highest RMSD angles relative to the original kinematics.
Pelvic kinematics showed very low sensitivity to marker displacement, with all of their calculated RMSD angles within an acceptable limit of 5° and the majority of their simulations resulting in RMSD angles below 2°. For the other calculated joint angles, the transverse plane is most affected by marker displacement, with approximately 47% of simulations returning an error beyond the 5° limit of acceptability. These results were consistent with previous literature indicating that the transverse plane was the least reliable in gait analysis.2.16.
An analysis of the ten worst-case marker displacement scenarios allowed us to better understand the effects of a combination of marker displacements on the lower extremity model. For example, the “worst” marker configuration for the pelvis was calculated when the anterior iliac spine markers were moved in opposite directions on the vertical axis and the ACSS was moved on the horizontal axis. With this simulated marker setup, the pelvis was both tilted and rotated relative to its original definition. Since the CGM is a hierarchical, anatomical, top-down model, one would expect it to affect the estimation of hip joint center, hip kinematics, and all distal angles of the joint . As shown in Fig. 3, the thigh and tibia flexion-extension axes of the CGM are defined as being orthogonal to the plane connecting the proximal and distal joint centers when the rod is placed along the segment. Thus, the medial-lateral axes of these segments are estimated to be orthogonal to the flexion-extension and proximal-distal axes. The simulated displacement of the femoral rod (LTHI) in the antero-posterior axis directly affects the coronal plane of the femur, thus modifying the flexion-extension axis and the medio-lateral axis. As a result, the kinematics of the hip and knee joints will be directly affected, as well as the center of the knee joint which is defined along the mid-lateral axis of the femur (in the absence of the femoral epicondyle marker medial). A similar impact was noted for the tibia. Finally, the medial displacement of the LTOE marker was responsible for a rotation of the angle of the foot with respect to the direction of walking and an impact on the angle of progression of the foot.
Regarding the displacements of the individual markers, the displacements of the thigh and tibia rods and the knee marker in the antero-posterior axis had the greatest calculated impact on the kinematics, all with an RMSD angle greater than 5 ° in the transverse plane (Fig. 4). These results confirmed previous findings demonstrating the high impact of the knee marker in the antero-posterior axis in the transverse plane, but its very low impact when moved in the proximal-distal axis.12.13. Even though some studies have reported improvements in calibration methods, such as the knee aligner, the reproducibility and reliability of marker placement remains the most significant limitation of CGM.17. The high sensitivity of the CGM to wand orientation is even more critical as the lack of anatomical landmarks makes its placement somewhat subjective. The current user manual specifications for stick placement are simply: “Adjust the position of the marker so that it is in the plane that contains the centers of the hip and knee joints and the axis of flexion/extension of the knee”.18.
CGM is characterized by a hierarchical, anatomical, top-down approach; therefore, a displaced marker affects the kinematics of each joint located distal to the anatomical segment containing that marker and the joint most proximal to it15. In addition, the slight impact we calculated on the angle of foot progression demonstrates that without the medial knee and ankle landmarks, the definition of the joint centers is affected by multiple landmark displacements. Thus, an error in placement of the knee joint center marker impacts the definition of the ankle joint center and consequently the angle of progression of the foot. Overall, the calculated impact of a displaced marker could be noted in both simulated displacements in opposite anatomical directions.
Gait scores, like GPS, are very effective in classifying a patient’s gait by comparing it to a reference database of an asymptomatic general population. As the calculation uses kinematic data, the variability noted due to marker displacement also introduces variability into the final gait classification and therefore can also have a significant impact on the interpretation of the gait data. We therefore investigated the impact of marker displacement on overall gait scores. Moving the marker in one leg caused GPS variations of up to 7°. This is comparatively well above the 1.6° considered the minimum variation of clinical significance.19. As the GPS is calculated from the kinematics of both lower limbs, the expected variation if our simulations were applied to both sides would be even higher.
The impact of marker placement variability on our simulated gait kinematics is shown in Fig. 5 by the calculated maximum RMSD angle corridors per frame in the walk cycle added around a subject’s original curve. We note that the error can be defined by a global offset added to the original data. This finding is consistent with previous results indicating that the impact of errors on axis definition was more like a shift in the kinematics than a change in their overall pattern.4. Such results can be useful for estimating the expected variability of kinematics when considering the expected variability of marker placement. To more accurately assess the impact of marker misplacement, our results could be used in combination with those of studies investigating the accuracy of marker placement, such as Della Croce et al.11. Thus, the magnitude of displacement of each marker would be defined based on the experimentally observed error.
Given the overall results provided in this study, different solutions can be proposed to arbitrate the displacement of the markers. First, the identification of anatomical landmarks must be followed carefully and with proper training of the responsible assessors. The guidelines used for marker placement in our data are recommended14. Second, the referred evaluator should pay extra attention to markers and directions that have a large impact on the kinematics, as shown in Fig. 4. In order to resolve the high sensitivity seen on the wand, lateral femoral epicondyle and lateral tibial malleolus at anterior-posterior misplacement, knee alignment device or medial femoral and tibial markers could be a solution, but specific studies are needed to validate the possible solution5.20. Third, in patients who underwent 3D imaging, a fusion between medical imaging and motion capture system could limit marker displacement but seems difficult to apply to all patients who performed clinical gait analysis.21.
This study had certain limitations. First, the lack of literature regarding the sensitivity of gait analysis to marker placement makes comparisons with our results difficult. Second, the marker displacement was performed virtually, so the effects of soft tissue artifacts could not be accounted for. Different distances and axes of marker movement could also induce different soft tissue artifacts22.23. Additionally, our benchmark marker placements cannot be considered “true” benchmarks as they were also subject to the uncertainty of marker placement. We only applied displacements of 10 mm in only four directions, although this distance was set based on Della Croce’s results and to serve as a potential standard reference for future comparisons.11. Finally, the enormous amount of simulations required to calculate each potential combination of marker displacement for the twenty subjects required enormous computing time. This imposed limitations on testing many distances and directions of travel, as discussed earlier for single-marker travel13.
To conclude, we performed a very thorough sensitivity analysis combining 390,625 simulated marker placements. We successfully identified the most sensitive angles contributing to an overall measure of marker displacement simulation and quantified the RMSD angles associated with displacements of different lower limb markers. We also identified and analyzed simulated worst-case marker displacement scenarios. In addition, we indicated which markers and which axes caused the greatest variability in the measured angles. Greater precision in the placement of thigh and tibia rods (or markers) and lateral femoral epicondylar markers in the anteroposterior axis is needed to improve the reliability of gait analysis using the MCG.