The Amsterdam Foot Model: Advancing the clinical assessment of multi-segment foot kinematics during gait: Advancing the clinical assessment of multi-segment foot kinematics during gait

The human feet are essential for our unique way of bipedal locomotion. A dysfunctional foot limits our mobility and consequently our participation in daily life. Hence, it requires careful clinical assessment before any treatment is provided to restore or maintain foot function. Musculoskeletal foot and ankle disorders can be evaluated with marker-based gait analysis. Many marker models have been developed to measure foot and ankle kinematics and several of these are used in a variety of clinical populations. However, a critical evaluation on whether the outcomes of current models actually represent the movement of the bony structures of the foot is still lacking. To further advance the clinical assessment of foot and ankle kinematic measurements during gait, measurement errors that affect the anatomical validity (i.e. accuracy) and reliability (i.e. precision) need to be quantified, and improved if required. Simultaneously, these models’ definitions and output have to align with the clinical perception of foot and ankle motions for a successful clinical application. In this thesis we aim to establish the most accurate and precise skin-marker-based method to measure clinically relevant foot and ankle kinematics during gait. First we explore and compare existing foot models. We show that the ankle angle and triceps surae muscle-tendon complex length are likely overestimated in a mono-segment foot model compared to a multi-segment foot model. Additionally, in chapter 3, we show that the two most-frequently used multi-segment foot models (i.e. the Oxford and Rizzoli Foot Models) provide different angular output for the same gait trials of normal and voluntarily pathological gait. In the second part of this thesis, we evaluate potential sources of measurement error in the Oxford and Rizzoli Foot Models to gain insight in their accuracy and precision. A well-known source of error within marker-based gait analysis are soft tissue artifacts, which are the relative motions of the skin-mounted markers with respect to their underlying bones. To measure these artifacts, we developed a low-dose loaded computed tomography scanning protocol including 10 scans with a total radiation dose that is similar to 1.5 times the dose of a single clinical foot and ankle scan (chapter 4). This protocol is used in chapter 5 to quantify the soft tissue artifacts, which are shown to be substantial and causing clinically relevant errors in multi-segment foot kinematics. Another potential source of measurement error is inconsistent marker placement. In chapter 6, we quantify the marker placement sensitivity and show that for both models each segment is highly sensitive to misplacement of at least one marker. In chapter 5 and 6, markers are identified that are most affected by soft tissue artifacts or inconsistent marker placement. The role of these markers within the model definitions should be reconsidered to improve the accuracy of the measurements and the robustness of the models to inconsistent marker placement. After these evaluation studies, we concluded that the evaluated models are affected by clinically relevant measurement errors that can lead to misinformation. Therefore, in part three of this thesis (chapter 7), we define a novel clinically-informed multi-segment foot model: the Amsterdam Foot Model. A foot and ankle expert panel was consulted to assure that the model definitions and its output align with the clinical perception. The model is explicitly based on minimizing the aforementioned measurement errors. Both types of errors were shown to be smaller in the Amsterdam Foot Model compared to the Oxford and Rizzoli Foot Models. Within the limitations of marker-based motion analysis, the clinically inspired Amsterdam Foot Model accurately and precisely measures foot and ankle kinematics during gait, while striving for an optimal combination of anatomical accuracy, clinical needs and practical applicability.