Establishment of a Finite Element Model and Biomechanical Analysis of Different Fixation Methods for Total Talar Prosthesis Replacement

Background: As a new technology, three-dimensional (3D)-printed personalized talar prostheses are associated with different xation methods for implanted prostheses, including xing the subtalar joint and talus-navicular joint with screws and xing the subtalar joint with screws only and xation without screws. No biomechanical study has been conducted yet. we aimed to build a 3D nite element model to compare the biomechanical effects of different xation methods. Methods: With 3D CT and MRI data of a volunteer's foot, Mimics research 19.0 and Geomagic wrap 2017 software were used to complete the geometric reconstruction of bone and cartilage, which was then input to NX12.0 software to build nite element models with different xation methods. Finally, the models were imported into Abaqus 6.14 software for meshing and assigning material properties and the different biomechanical effects in three gait phases (heel-strike, midstance and push-off) were simulated. The pressure changes in the articular surface around the talus or the prosthesis, the micromotion of the talus and the prosthesis and ankle motion were measured. Results: The 3D nite element model created in this study has been veried to be consistent with previous studies.Finite element analysis of the biomechanical mechanism showed that screw xation of the prosthesis in different gait phases mainly increased the pressure on the tibia-talus articular surface and decreased the pressure on the fused articular surface and joint micromotion, which hindered ankle motion. The indicator values were nearly the same in the models of xation without screws and the healthy state. Conclusions: The biomechanical mechanism varies by xation method according to the nite element analysis. Fixation of the prosthesis without screws yields values most similar to healthy values.


Background
The talus is the key bony structure connecting the lower limb and foot. Its anatomical structure is complex, and it is the mechanical point of rotation between the lower limb and foot. Stress is concentrated in this area, and the mechanical properties are particularly important [1][2][3]. Collapsible talus necrosis severely affects individuals' ability to stand and walk, with a disability rate of 100%. Ankle surgeons worldwide mainly perform total talus removal and partial joint fusion at the expense of talar function. Postoperative complications such as adjacent joint degeneration, joint stiffness, and the loss of foot exibility often occur, and the long-term e cacy of this method is very poor [4,5].
Due to advancements in modern three-dimensional (3D) printing and prosthesis casting technology [6], 3D-printed personalized talar prostheses have been used for clinical treatment [7][8][9][10], considerably improving the treatment for talus collapse necrosis worldwide. The surgical indications for prostheses are very similar to those used for conventional methods, but the method of prosthesis xation is different. In previous studies by Kadakia et al. [11] and Tracey et al. [12], the peritalar soft tissue was directly removed without xation of the talar prosthesis with screws. In studies conducted in China, the prosthesis was xed to the calcaneus using screws [13], or the prosthesis was screwed to both the calcaneus and navicula. However, there are no biomechanical studies on the above xation methods for total talar prostheses.
Finite element analysis (FEA) is also a new technology, and a mathematical model with high similarity can be established to re ect regional mechanical characteristics. The greatest advantage of FEA is that it can obtain research results that are di cult to obtain in objective human or animal experiments without causing damage [14,15]. Considering the ankle joint motion mechanics and gait cycle, the heel-strike phase, midstance phase and push-off phase represent the support and swing phases of the gait cycle and are the three periods that best re ect a healthy gait and pathological function. Therefore, these three phases were simulated to measure the stress and micromotion of the joints around the whole talar prosthesis as well as the ankle joint motion [16,17].
In this study, we intended to use the nite element method to explore the characteristics of the whole talar prosthesis under different xation methods to compare the biomechanical effects of different xation methods.

Data Collection
A male volunteer aged 30 years with a height of 177 cm, a body weight of 75 kg and a foot length of 250 mm participated in planning left arti cial talus replacement in this study. An X-ray examination on the right ankle[Ed31] was performed rst, and other diseases, such as foot tumors and deformities, were excluded. A 64-slice spiral CT scan (spatial resolution is 30 Lp/cm) of the right ankle was performed, with a slice thickness of 0.60 mm, and MRI was also performed to verify the cartilage boundaries. The data were output and saved in the DICOM format. In addition, we also collected MRI data of the right foot of volunteers to help determine the cartilage boundaries.

Finite Element Model Establishment
The CT scan data were imported into the 3D reconstruction software Mimics 19.0, and the bone tissue and soft tissue were separated (thresholding, split masking, region grawing) to establish a geometric model of the whole foot, which was output as an STL le. Then, this le was imported into the reverse engineering software Geomagic Wrap 2017, removing the noise of the model and smoothing the model.
According to the geometric shape of each joint surface, the cartilage boundary was divided on each bone surface, the surface was tted, and the resulting le was output in the Stp format. Then, it was imported into the nite element preprocessing software NX12.0 to build different models. Next, the models were imported into ABAQUS6.14 software. Based on the ligament data, a model with ligaments was established by connecting the ligament attachment points with the 3D arrangement of the ber bundles. Finally, the solid model was subjected to mesh generation, material attribute selection, and other processing steps ( Figure 1).

Material Parameters
The bony structures and cartilage were modeled as isotropic linear elastic materials, and the ligaments had a nonlinear single-axis connection unit to simulate the characteristics of tension only without compression. The material properties of the bones, cartilage, titanium alloy and ligaments were determined according to previous studies and are listed in Tables 1 and 2 [18,19]. Finally, the nite element model and nodes were built and are shown in Table 3. Table 1 Properties of the bone and cartilage materials.

Boundary Conditions and Loads
There are many ways to divide a gait cycle, but it is usually divided into three phases: the heel-strike phase, the midstance phase and the push-off phase. The axis of exion-extension and center of rotation were determined in this study according to the research methods reported by previous scholars [18,19]; the centers of the arcs of the tibialis and peroneal circumferences of the talus pulley were determined, and they were related to the two centers of the circle, the rotation axis, and the midpoint of the two centers of the circle, the rotation center. In research on stress during gait, the load is approximated to be relatively static. Stress during the gait cycle is simulated by applying loads of different sizes and directions. A contact pair is established between the articular surfaces with a coe cient of friction (no friction between the joint surface and titanium alloy prosthesis interface) of 0.01 [18,19]. Table 4 and Figure 2 show the reference data for stress in different phases ( Figure 2). The healthy model was consistent with previous studies. Table 4 Finite element analysis parameters.

Results
3D nite element models of talar prosthesis xation with different methods were constructed and analyzed. The speci cally constructed nite element model is shown in Figure 3, and the distribution of pressure nephograms for the three different time phases of the healthy model is presented as an example in Figure 4. The results regarding the pressures on the joints adjacent to the talus were as follows: 1. For the tibia-talus joint, the contact forces in the three phases were uniform under healthy conditions. After the talar prosthesis was implanted, the pressure signi cantly increased in the three phases, but the changes in the no xation model were not signi cant. 2. For the subtalar joint, under healthy conditions, the pressure gradually increased in the heel-strike, midstance, and push-off phases in sequence. The trend was consistent after talar prosthesis replacement, and the xation method without screws yielded results similar to those under the healthy state. When the subtalar joint was xed, the pressure on the subtalar joint decreased signi cantly. Therefore, once the subtalar joint was xed again, the range of motion of the hindfoot was further limited, and the subtalar joint pressure decreased again. 3. For the talar joint, the pressure values increased across the heel strike, midstance and push-off phases. The trend was consistent after talar prosthesis replacement. When screws were not used to x the prosthesis, the pressure values in the midstance and push-off phases tended to increase. When screws were used to x the subtalar joint, the pressure in the three phases of the talar joint decreased. When the talar joint was xed again, the pressure value decreased further (Tables 5-7). Regarding slight movement of the talar adjacent joint, when a load was applied to simulate the three phases of the gait cycle, we compared the displacement of the talus relative to the adjacent joints, i.e., the tibia, calcaneus and scaphoid, and the results were as follows: The micromovement between joints was reduced by screw xation, and the situation was particularly obvious when the talus and subtalar joints were xed. The displacement of the xation without screw xation of the talar prosthesis relative to the tibia was reduced, probably because the corresponding ligament-supporting connection was lost, but there was no signi cant increase in relative motion at either the subtalar joint or the talus-navicular joint (Tables 8-10). For the ankle joint range of motion, the pressure change between the peri-articular surfaces of the talus and the micromotion of the talus were measured, with the plane axis extending from the central axis of the second metatarsal bone to the heel and the central axis of the tibia used as a reference; the ankle joint was in the neutral position in the midstance phase, in dorsi exion during the heel-strike phase, and in plantar exion during the push-off phase. The ankle joint range of motion in the three phases was measured. The results showed that after total talar prosthesis replacement, the ankle joint range of motion changed. Screw xation greatly limited the range of motion (in line with the characteristics of fusion surgery). There was also limited range of motion in the model with xation without screws, but this situation was the closest to the healthy situation (Table 11).

Discussion
The talus plays an important role in the biomechanics of the ankle. Abnormal anatomical structures have large effects on the function of the foot. The ankle joint bears a heavy load in the human body. Any injury to its anatomical structure will affect its stability. Talus osteochondral injuries are common ankle joint injuries that considerably affect the ankle joint [20].
Ischemic collapse necrosis of the talus is challenging to treat [21,22]. To address this clinical challenge, there are currently three main therapeutic approaches: 1. core decompression, which can preserve joint motion and effectively relieve pain, but the disadvantage is that it is suitable for only patients with early talus necrosis and is not effective for end-stage necrosis [23]. 2. Ankle joint fusion surgery has been suggested to relieve pain and is suitable for patients in almost all stages of necrosis, but it will greatly limit the range of motion of the ankle and affect the quality of life of patients [24]. After the talus collapses, structural bone grafting is often performed during fusion to prevent the force lines of the lower limbs from being affected. If the blood supply around the talus is damaged extensively, tibialis calcaneal fusion or posterior ankle arthrodesis is needed. However, for cases of severe collapse and necrosis of the talus, fusion surgery is not suitable, and in earlier studies, fusion surgery has been proven to be inferior to ankle prosthesis replacement in terms of mobility, e cacy and prognosis [25]. 3. Regarding ankle joint replacement for collapse necrosis of the talus, the requirement of residual bone mass of the talus is very high to reduce the probability of revision or refusion.
Whole talar prosthesis implantation was rst performed and reported by Harnroongroj and Vanadurongwan [26] in 1997, but there were many postoperative complications due to limitations of the casting method. In recent years, with the development of modern computer processing technology, 3D printing technology has been widely used in the clinical practice of orthopedics and has yielded good curative effects.
3D printed, personalized all-talar prostheses can be used for collapse necrosis of the talus. The talus, the core of ankle-hind foot movement, has seven joint surfaces, so it is the rst choice for personalized treatment. However, there is always controversy about whether the whole talar prosthesis should be xed after implantation. Regardless of whether the subtalar joint or subtalar joint is fused, the degree of exibility and range of motion of the foot are affected [13]. If we choose not to x the talar prosthesis [11,12] and use the bony structure of the talus between the ankle points and the ankle-foot complex to obtain self-stability, damage to the adjacent articular cartilage and complications such as prosthesis dislocation may occur. The clinical e cacy of different xation methods has also been assessed in many studies. Due to the relatively short follow-up times and limited number of cases in this study, more scienti c and objective data cannot be provided. Therefore, biomechanical studies are urgently needed to verify the biomechanical differences between several different xation methods so that operators can select the best surgical method.
Traditional orthopedic biomechanical experiments are mainly based on animal or cadaver models (also gait analyses). Although the results of these experiments are more reliable than those of simulations, it is often very di cult to obtain ideal experimental data without changing the physiological state of the model due to limitations in experimental methods, the need to adhere to ethical standards, and the in uence of other factors. In recent years, with the development of medical imaging technology and computer processing technology, nite element analysis, a new biomechanical research method, has been widely used in orthopedic mechanics research. Simulation experiments performed using the nite element method have the advantages of a short experimental time, a low cost, the capability of simulating complex boundary conditions, the ability to provide comprehensive mechanical property testing, and good repeatability [14,15]. In this study, a nite element model was used to simplify and effectively simulate a healthy model and models of different methods of xing a talar prosthesis.
It is traditionally believed that xation must be performed after the prosthesis is inserted, which is similar to fusion surgery. The subtalar joint [13] needs to be xed or the scaphoid joint needs to be xed at the same time to stabilize the prosthesis at the ankle. For prosthesis-bone interfaces needing screw xation, a special coating is often used to achieve the biomechanical effect of bone ingrowth. Professor Tang Kanglai conducted a series of studies on this topic and made clinical progress [27,28]. The operation is similar to fusion surgery. Hindfoot motion is limited to a certain extent, which was shown in the nite element model. When loads in different directions were applied, the range of motion of the talus relative to the screw-xed joint surface decreased; during a simulated gait cycle, the pressure on the tibia-talus joint increased, and the pressure on the prosthesis-bone interface decreased. Screw xation does limit the motion between joints and reduce the pressure between fusion joints; however, the nite element results showed that the reduced pressure is completely compensated by the tibialis joint. For the talus joint, there is not only a loss of range of motion and an increase in contact pressure but also an increased probability of osteoarthritis in the talus joint surface and an increased possibility of late prosthesis loosening in the long term. However, the effect on ankle range of motion is similar to that of fusion surgery. Clinical research on total talus replacement with a prosthesis indicates that talar prostheses have a better curative effect in the short term [29]. However, studies on bone ingrowth between the bone and prosthesis interface and long-term clinical follow-up studies are currently underway.
Other scholars have used methods other than screw xation. The rst report of using an un xed method was published by Assal and Stern [30] in 2004, and good curative effects were achieved within the ve- year follow-up period. In the heel-strike phase of the gait cycle, when the ankle is in dorsi exion, the talus was locked upward in the ankle, and there was a force component exerted vertically downward on the calcaneus; therefore, xation was not required. In the midstance phase, the upper surface of the talar prosthesis incurred downward stress from the tibia, which exerted a force against the calcaneus and scaphoid at an angle of 140°. The talus was relatively stable and did not need xation. During the pushoff phase, the ankle began to plantar ex, the hindfoot was locked, and the ankle was in the "unlocked state". The moment arm of the talus against the scaphoid increased, and xation was not required [31].
The nite element model of this study also con rmed the assumption that talar prosthesis xation without screws yields stable xation, with biomechanical and ankle range of motion values closest to normal values. To the best of our knowledge, no one has conducted relevant studies on self-stability of the talus in the past.
If screw xation is used, the probability of prosthesis dislocation is relatively low. For cases without screw xation, dislocation is possible in the following conditions. When the forefoot is off the ground, the Achilles tendon is pulled upward, the hindfoot is plantar exed, the talus is unlocked forwards, and the anterior ankle is loose. Then, the talus is displaced to a large extent forward and upwards. Due to the containing effect of the scaphoid (bony structure) and the limiting effect of the tibialis anterior and anterior joint capsules (soft tissue) on the prosthesis, the probability of prolapse is relatively low, and the speci c biomechanical mechanism needs to be studied further.

Conclusions
As has been veri ed by similar research, the nite element model established in this paper is a reliable model that can effectively re ect the biomechanical methods of different xation methods after talus prosthesis implantation. By comparison, the use of screw xation prostheses limits the ankle joint range of motion to a certain extent and changes its original biomechanical characteristics. Instead of using screw xation of talar prostheses, xation without screw is the closest to the normal xation method.

Limitations
Some shortcomings of this study were unavoidable. First, only the bony structures were simulated, and the soft tissues were simpli ed, which may affect the accuracy of this model to some extent with respect to real conditions. Second, the model was veri ed by repeating the experiments in previous studies, which does not yield the strongest form of evidence. Therefore, in the future, we plan to verify the results of this YQD, CL and ML completed the collection of the clinical data, while BXT and CW completed the statistics. TKL and TX served as reviewers and gatekeepers of the overall quality of the article.