DFIS is derived from fluoroscopic imaging technology, which is widely used in the medical field for its penetration and noninvasive characteristics. However, fluoroscopic imaging mainly captures static bone images and thus cannot easily quantify the dynamic movement of humans. Therefore, researchers developed DFIS for the capture of bones and joints in vivo by combining fluoroscopic imaging, medical imaging and 3D-2D model registration technologies. This instrument consists of an X-ray fluoroscopic and a data analysis system. The X-ray fluoroscopic imaging system comprises two radiographic source and detector pairs, two mobile manipulators with fluorescence receivers and intensifiers and two matching digital cameras (Figure 2). By relying on computed tomography (CT), X-ray imaging and magnetic resonance imaging (MRI), the motion fluoroscopic image system can accurately quantify the movements of in vivo bony structures. The data analysis system is composed of 3D modelling, 3D–2D model registration and a motion analysis system (Figure 3). It is mainly responsible for registering and reconstructing 2D images and 3D models in 3D space to provide the exact position of each bone or joint. It is the basis for the final quantification of joint relative displacement, relative angle, cartilage contact area and cartilage contact stress/strain.

Numerous methods have been utilised to quantify human movement (Table 1). Some common methods include motion capture systems, which are extensively used to quantify joint kinematic by tracking reflective markers that have been adhered to the skin at bony landmarks (Roach et al., 2017). However, the measuring accuracy of this method is limited by artificial factors, such as marker location, and nonartificial factors, such as skin and soft tissue artefacts (Bauman and Chang, 2010). Furthermore, this system cannot analyse the movement of a single bone nor obtain the 6DOF motion of joints. Although computer simulation methods, for example, the finite element model (Gültekin et al., 2021), can simulate the 6DOF movement of an independent joint with computer programming, it cannot measure in vivo joint kinematics, let alone obtain the real movement of the joints in functional activities, such as walking, jumping and running. Human cadaver studies involving the implantation of steel beads or pressure sensors into the human body can also quantify the movement in a single bone but are invasive and raise ethical concerns. Traditional radiological technologies, such as X-ray plain film, MRI and CT, are also used to track bone positions (Kroupa et al., 2020). However, the subjects in the studies based on these technologies are usually in the supine position, which precludes the evaluation of human movement during activities. DFIS can compensate for the shortcomings of other biomechanical methods by noninvasively capturing the motion of multiple joints independently of other joints in dynamic human activities. DFIS is a new approach that is safe and reliable. Cross et al. (2017) estimated that the effective dose for a foot and ankle CT, plus one biplane fluoroscopic static trial and 10 dynamic trials in the current system, is180 μSv, which is far below the annual occupational limit of the whole-body effective dose of 500,000 μSv established by the US Nuclear Regulatory Commission (Brenner et al., 2003). In term of precision, DFIS demonstrated a bias range of −0.16–0.13 mm and −0.05–0.13°, a precision range of 0.05–0.86 mm and 0.06–0.69° and an overall dynamic root-mean-squared average error of 0.59 mm and 0.71° in static and dynamic trials (Cross et al., 2017).

The quantitative analysis of the kinematics of the foot and ankle joints provides a comprehensive understanding of the weight-bearing mechanisms and laws of ankle movement, ankle function and potential injury mechanisms. In the absence of a suitable location for placing a skin marker around the talus, infrared motion capture systems cannot measure the independent motion of the tibiotalar joint from the subtalar joint (Roach et al., 2016). Instantaneous helical axes (IHAs) are commonly used to investigate the maximum range of motion (RoM) of the tibiotalar joint and subtalar joint in three rotational directions (Akinnola et al., 2020). Arndt et al. (2004) found that during walking, the tibiotalar and subtalar joints have a RoM of inversion/eversion (IN/EV) of 6.3 and 8.3°, respectively, and the dorsiflexion/plantarflexion (DF/PF) of 18.7 and 3.7°, respectively. However, a study using DFIS found that the tibiotalar and subtalar joints have the RoM of 3.8 and 11.3° for IN/EV, respectively, and 16.3 and 8.6° for DF/PF, respectively; these results are different from those obtained from IHAs (Roach et al., 2016). These differences are related to the inability of IHAs to determine the precise position of the bone. Meanwhile, reliable kinematics data for the different stages of dynamic activity are insufficient when the results for IHAs are only for the joint angle at a certain time.

DFIS can noninvasively measure the movement characteristics of a single bone. It has been used to quantify foot and ankle kinematics in healthy populations under barefoot conditions (Table 2). Phan et al. (2019) used DFIS to observe the movement of the transverse tarsal joint during barefoot walking and found that this joint is positioned from the maximum pronation in the early stance phase to the maximum supination in the late stance phase. This action is combined with muscle activities to effectively transmit the push-off force in the late stance phase. This finding provides a new perspective for understanding ankle movement during walking. Additionally, the translation and rotation of the tibiotalar and subtalar joints vary at different stages of the support phase. During the initial landing, the tibiotalar and subtalar joints perform remarkable 6DOF movements, whereas only the subtalar joint performs a remarkable IN/EV and internal/external rotation (IR/ER) at the late support phase (Phan et al., 2018).

Previous DFIS studies have demonstrated that the tibiotalar joint primarily performs DF/PF movements and the subtalar joint primarily performs IN/EV and IR/ER movements in the support phase of walking (de Asla et al., 2006). Roach et al. (2016) discovered that in addition to DF/PF movements, the tibiotalar joint also undergoes slight IN/EV and IR/ER movements, and the subtalar joint performs DF/PF movements in walking and single-leg balanced heel-rise tasks; the PF of the subtalar joint even appears earlier than that of the tibiotalar joint (Koo et al., 2015). Yamaguchi et al. (2009) used a single-plane fluoroscopic imaging system and found that the tibiotalar and subtalar joints execute DF/PF and IN/EV in single-leg heel-rise tasks. The difference in kinematic results indicates that DFIS effectively compensates for the lack of a single-plane perspective for the observation of the 3D motion of the joint and comprehensively analyses a single bone in a dynamic/static status. Furthermore, the tracking data can effectively supplement the previous kinematics results, thereby avoiding the erroneous judgment of movement law as much as possible.

Shod conditions seriously affect the kinematics of the foot and ankle joints (Roberts et al., 2011). For a long time, researchers used various technical means, amongst which motion capture systems were widely used, to identify the relationship between shoes and foot and ankle biomechanics. However, given that the kinematics calculated via skin marker motion analysis are visualised by using a generically scaled model, the relationship between individual anatomical features and the ensuing joint motion cannot be discerned (Roach et al., 2016). Meanwhile, in this method, bone movement is mostly calculated by pasting reflective markers on the upper shoe or destroying the shoe structure to paste the markers (Cigoja et al., 2020; Yang et al., 2020). This approach complicates the accurate analysis of the influence of shod conditions on the foot and ankle caused by the unavoidable technical defects of the motion capture system. By contrast, DFIS, which is based on fluoroscopic imaging technology, can directly observe the actual movement of the bone. It thus compensates for the inability of motion capture systems to quantify the bone inside of the shoe and is now the only method that can accurately capture the in vivo kinematics of the foot and ankle, even the metatarsophalangeal joint, under shod conditions.

Studies used DFIS to observe the real movement of bones inside shoes and found that shoes limit the PF movement of the ankle joint during walking (Campbell et al., 2016). The remarkably earlier appearance of the DF/PF and EV of the ankle joint under shod conditions than under barefoot conditions indicates that shoes may affect the initiation of foot movements and muscle activation. Furthermore, the RoM of the DF/PF and IR/ER of the tibiotalar joint under barefoot conditions is greater than that when wearing minimalist or motion control shoes during running (Peltz et al., 2014). This finding is inconsistent with a previous result that showed that minimalist shoes have little influence on foot and ankle movements and can simulate barefoot running (Nordin and Dufek, 2020). This inconsistency can be attributed to differences in conditions, such as the type of minimalist shoes used, and may also be related to the inability of previous studies to observe the real movement of the inner bone under shod conditions. Shod conditions also influence the movement of the rest of the foot and ankle joints. For example, a study that utilised DFIS discovered that the navicular drop rate is lower when running in motion control shoes than when wearing minimalist shoes or under barefoot conditions (Hoffman et al., 2015). Thus, wearing motion control shoes is more likely to prevent foot and ankle injuries, such as iliotibial band syndrome, periostitis, bursitis and stress fractures (Benca et al., 2020).

Lateral ankle sprain (LAS) is one of the most common injuries in sports and recreation; it accounts for up to 23% of all athletic injuries (Fong et al., 2009). Patients with LAS have a high probability of developing joint degeneration and chronic symptoms (Bell et al., 2006). Studies based on DFIS have found that the increasing rates of the cartilage contact deformation and contact strains of the ankle joints significantly increase in the early stance phase (Li et al., 2008). This finding indicates that foot injuries, such as LAS, may occur during the early stance phase of walking and proves the hypothesis regarding the mechanism of LAS (Gribble et al., 2016; Delahunt et al., 2018). The anterior talofibular ligament (ATFL), one of the lateral ligaments of the ankle joint, is linked to LAS and osteoarthritis development (Caputo et al., 2009). However, the injury mechanism of ATFL and the relationship between ATFL injury and the kinematics of the foot and ankle joint remains unclear.

A previous study used DFIS to explore the mechanism of ATFL injury and discovered that the ATFL elongates from the neutral position to the maximum PF position and from the maximum ER to the maximum IR (de Asla et al., 2009). This result suggests that ATFL injury may occur during the movement of PF with IR due to the excessive stretching of the ATFL. Caputo et al. (2009) quantified the foot and ankle kinematics of patients with ATFL injury and found that the IN and anterior and superior translation of the talus increase during walking. This increment potentially increases the load of the medial talus cartilage and shear force. Additionally, ATFL injury increases the cartilage deformation of the tibiotalar joint (Bischof et al., 2010). The correspondence between the increase in deformation and the location of osteoarthritis supports the hypothesis that ATFL injury may be potentially associated with osteoarthritis and joint degeneration.

LAS also affects walking kinematics. Compared with healthy controls, patients with LAS exhibit larger tibiotalar anterior/posterior translation during walking and excessive tibiotalar IN during stair descent (Cao et al., 2019a; Cao et al., 2019c). Chronic ankle instability (CAI) is a consequence of the interaction between mechanical and sensorimotor insufficiencies/impairments following acute LAS (Kobayashi and Gamada, 2014; Delahunt et al., 2018). Some researchers used DFIS to analyse the influence of CAI on ankle movement and found that CAI restricts the activities of the tibiotalar and subtalar joints during walking and single-leg balanced heel rise (Roach et al., 2017). This finding indicates that balancing tasks, such as single-leg heel rise, may be one of the best ways to assess the prognosis of patients with CAI because they can reveal additional differences between the foot and ankle kinematics of patients with CAI and the healthy population. The greater IN and anterior translation of the tibiotalar joint during stair descent (Cao et al., 2019a) and greater lateral/medial translation and IN/EV of the subtalar joints in the stance phase of walking shown by patients with CAI relative to those shown by patients with LAS and healthy subjects indicate that lateral ankle injury persists in patients with CAI (Cao et al., 2019c). Meanwhile, previous studies have discovered that patients with CAI experience excessive IN movement in the ankle and that the subtalar joint is mainly responsible for IN/EV in the global movement of the ankle joint (de Asla et al., 2006). Thus, the functional training of the subtalar joint may have considerable clinical importance in the treatment of CAI.

DFIS studies have also found that ankle braces can limit the IN/EV of the tibiotalar and subtalar joints during the stance phase of walking (Zhang et al., 2019). However, ankle braces neither limit anterior translation and the PF of the subtalar joints nor help the RoM of the subtalar joints return to a normal or near-normal level in patients with LAS and CAI (Cao et al., 2019b). This result differs from the findings of previous studies that utilised the traditional motion capture system and showed that ankle braces could limit excessive ankle movement and reduce re-sprain risk for patients with CAI (Dewar et al., 2019). This difference occurred because the traditional motion capture system can only analyse the overall motion of the ankle joint but not the individual motion characteristics of the tibiotalar and subtalar joints. Additional well-designed studies, especially studies using DFIS, are needed to verify the real effect of ankle braces in pathological populations.

Functional flat foot is a common orthopaedic problem that may result in disability; its prevalence can reach 20% amongst athletes (Michelson et al., 2002). Compared with people without functional flatfoot deformities, those with functional flatfoot deformities are more likely to develop tendinopathy of the tibialis posterior or Achilles, plantar fasciitis, patellofemoral pain syndrome or even lower back pain as a result of the change in foot structure (Lakstein et al., 2010; Beeson, 2014). Many techniques have been used to assess functional flat foot. The footprint method is one of the most popular and widely used techniques (Wozniacka et al., 2013). However, this method cannot directly measure the real arch movement of people with functional flat feet.

For the past few years, DFIS has been applied to quantify the medial longitudinal arch (MLA) angle and shown that people with functional flat foot have a large MLA angle, and this angle increased during dynamic activities (Balsdon et al., 2016). This result is predictable because the arch of the functional flat foot collapses and the height of the navicular bone is relatively low. In dynamic tasks, the skeletal and ligamentous structures that constitute the MLA play a major role in transferring and dampening forces through their deformation to protect the foot (Caravaggi et al., 2019). Foot orthotics can reduce the angle of MLA in people with functional flat foot and increase the arch height to adjust abnormal foot morphology to a certain extent (Balsdon et al., 2019).

Previous studies have typically used motion capture systems to assess functional flat foot (Su et al., 2017). For example, Jung et al. (2011) found that the MLA angle in healthy people when standing is 145.1 ± 5.5°, which is considerably higher than that measured by DFIS under the same condition (129.2 ± 7.6°) (Balsdon et al., 2016). This difference could be attributed to the subtler motion of the foot arch than that of the rest of the lower extremity kinetic chain. Meanwhile, the size, position and relative displacement to the skin of markers make the arch kinematics measured via traditional measurements very different from the real state (Gorton et al., 2009). Therefore, DFIS has a higher potential for clinical application than the traditional measurement method that uses inverse dynamics to identify joint damage mechanisms or disease evaluation by clinicians. Furthermore, DFIS can be exploited to observe in vivo bone movement directly for the analysis of potential pathogenic mechanisms and the provision of favourable conditions for the accurate diagnosis and treatment of ankle injury.

Current studies on foot and ankle pathologies have clarified the influence of injuries and abnormal arch on movements from the perspective of kinematics. The role of clinical rehabilitation devices and potential factors that cause arthritis and joint degeneration have also been discussed. However, only a few reports have focused on DFIS combined with kinetic measurements. Therefore, future works should combine kinetic measurements, such as force platform and foot pressure technology, with DFIS to quantify the biomechanical characteristics of foot and ankle movements in a multidimensional manner and understand the relationship between ankle movement and injury.