Novel Force Measurement System for Soft Tissue Balance in Total Knee Arthroplasty Based on Flexible Pressure Sensor Arrays

Total knee arthroplasty (TKA) is a common treatment for terminal knee arthropathy. Soft tissue balance is one of the key factors affecting the success rate of TKA. However, current measurement systems still face great challenges in terms of accuracy and timeliness. Herein, a novel force measurement system based on a smart spacer is proposed to measure soft tissue balance in a timely and accurate manner. The smart spacer is designed according to the simulated results, where six flexible pressure sensors with a high sensitivity of 99.56 N−1 and a wide detection range of 50 N are attached to the spacer base to directly detect the soft tissue balance force. Moreover, a soft tissue balance measurement system is constructed, which can measure the force at different positions, such as the extension position (0°), the middle position (45°), and the flexion position (90°), and display them on the computer in a timely manner. This novel real‐time knee soft tissue balance measurement system provides a way for surgeons or knee joint replacement surgery robots to accurately evaluate soft tissue balance and paves the way for improving the success rate of TKA and the patient's postoperative rehabilitation.

DOI: 10.1002/aisy.202100156 Total knee arthroplasty (TKA) is a common treatment for terminal knee arthropathy. Soft tissue balance is one of the key factors affecting the success rate of TKA. However, current measurement systems still face great challenges in terms of accuracy and timeliness. Herein, a novel force measurement system based on a smart spacer is proposed to measure soft tissue balance in a timely and accurate manner. The smart spacer is designed according to the simulated results, where six flexible pressure sensors with a high sensitivity of 99.56 N À1 and a wide detection range of 50 N are attached to the spacer base to directly detect the soft tissue balance force. Moreover, a soft tissue balance measurement system is constructed, which can measure the force at different positions, such as the extension position (0 ), the middle position (45 ), and the flexion position (90 ), and display them on the computer in a timely manner. This novel real-time knee soft tissue balance measurement system provides a way for surgeons or knee joint replacement surgery robots to accurately evaluate soft tissue balance and paves the way for improving the success rate of TKA and the patient's postoperative rehabilitation.
did not assess soft tissue in the dynamic state, which made the surgical component inaccurate and unreliable. Bardou-Jacquet et al. proposed a novel method for ligament balance in total TKA by combining a load sensor with robotic technologies for ligament balance. [38] This technique enabled quantifiable alignment and control of ligament tension. It was noted, however, that the loading sensors were not integrated with the spacer surface, making it difficult to measure soft tissue balance directly. Jiang et al. reported a visualized sensing system based on embedded sensors inside the spacer with the kinetic prosthesis pose reconstruction in TKA, which was used during the surgery as a trial component to provide pose reconstruction and contact force distribution. [39] It is worth noting that the aforementioned sensing systems were based on a sensor integrated into the internal cavity of the spacer (which is part of the knee prosthesis), owing to the limitations of integral design and pressure sensor arrays. In previous studies, [37][38][39] soft tissue tension acted on the hard surface of the spacer, followed by the hard curved surface exerting stress on the internal sensor chip. The initial tension was then transferred to the sensor chip through a secondary transformation. Thus, the original soft tissue tension could not be accurately and directly measured. [26,[40][41][42] Herein, a novel soft tissue balance measurement system is proposed using an integral design strategy based on flexible pressure sensor arrays to directly measure soft tissue balance. The flexible pressure sensor arrays were constructed using a conductive fabric, printed electrodes, and a biogel packaging layer, and showed a high sensitivity of 99.56 N À1 and a wide detection range of 0-50 N to meet the needs of the surgery. Importantly, the smart spacer based on the flexible sensor arrays could directly sense soft tissue tension because the sensors were distributed on the upper curved surface of the smart spacer to directly contact the femur component. In addition, the soft tissue balance measurement system model was integrated using a smart spacer, a multichannel signal acquisition platform, and a 3D-printed knee joint prosthesis. The resulting novel system can accurately measure the force of both sides of the knee joint at different positions, such as the extension position (0 ), the middle position (45 ), and the flexion position (90 ), in a timely manner. This promising, effective, and economical measurement system can greatly improve the success rate of TKA and postoperative rehabilitation.

Working Principle and Composition of the TKA Force-Sensing System
In TKA, the knee joint prosthesis is mainly composed of the femoral component, spacer, and tibial component. Surgeons use bone cement to attach the femoral and tibial components after an osteotomy with surgical instruments. The medial and lateral collateral ligaments of the knee joint are often unbalanced due to pathology and relaxation, which causes differences in stress distribution on the surface of the spacer ( Figure S1, Supporting Information), further affecting the lifetime of the knee prosthesis and the mobility of the patient's knee joint after surgery.
Herein, a novel TKA force-sensing system is proposed to measure the balance function of the knee joint, which will help the physician to loosen the soft tissue and improve the patient's degree of recovery. The measuring system mainly consists of a smart spacer, femoral component, tibial component, and display surface ( Figure 1). In surgery, unsuitable pathology and relaxation can cause an imbalance of the medial and lateral collateral ligaments of the knee joint, thereby resulting in different forces at different sides. Thus, attaching the proposed smart spacer to the surface first enables it to directly measure the forces at different angles and positions of the joint to evaluate soft tissue balance. The measurement process is as follows: the electrical signals of the pressure sensor arrays, which are located at different positions and angles, are first collected and then transmitted and displayed in real time on the LabVIEW software interface using the multichannel signal acquisition platform ( Figure S2, Supporting Information). According to the feedback information, the surgeon can adjust the balance of the soft tissue through arthrolysis, which greatly improves the patient's knee joint movement function.

Structure Design and Fabrication of the Smart Spacer
The diagram of the smart spacer structure is shown in Figure 2, and mainly includes a spacer base, printed electrodes, flexible pressure sensor arrays, and a biocompatible packaging layer. The spacer base was obtained using 3D laser scanning and reverse modeling according to the spacer in the actual knee joint prosthesis. A grooved structure was designed on the upper surface of the spacer to arrange the fabricated flexible sensor arrays, whose structure is similar to that reported in the literature. Therefore, the smart spacer can more realistically and accurately reflect the stress distribution of the contact surface between the femoral and tibial components because it is similar to the working environment of the actual prosthesis.
The flexible pressure sensor arrays were constructed based on the conductive fabric, printed electrodes, biogel packaging layer, and flexible polyethylene terephthalate (PET) substrate. The sensors can be distributed on the surface of the spacer to directly sense soft tissue tension and transmit the signals to the acquisition circuit located at the back end, which is different from the general embedded structure inside the spacer. [38] The size and  Figure S3, Supporting Information). The knee joint prosthesis was placed at three different bending angles (0 , 45 , and 90 ), and the corresponding stress distribution results were exhibited on the spacer. When the knee joint was in a straight state (0 ), the stress distribution on the spacer was at the top ( Figure S3a, Supporting Information). When the knee joint was bent at 45 , the contact position of the femoral component and spacer was in the middle position ( Figure S3b, Supporting Information). When the knee joint was bent at 90 , the contact stress was at the bottom position ( Figure S3c, Supporting Information). According to the abovementioned knee joint simulation results at different angles, the six pressure arrays attached to the peak contact stress position at the three bending angles could accurately detect the contact stress of the knee joint. In brief, the fabrication process mainly includes the following six steps: 1) preparation of a patterned mold; 2) fabrication of a patterned polydimethylsiloxane (PDMS); 3) plasma treatment of the packaging layer; 4) printing of the electrodes; 5) assembling the sensor array; and 6) combining them into a complete smart spacer. A detailed preparation process is presented in the experimental sections. It is worthy to note that the preparation method is simple, reliable, low cost, and convenient for large-area preparation and mass production.

Performances of the Core Detection Sensor
A schematic of the sensor structure is shown in Figure 3a. The sensor mainly includes the biogel packaging layer, a conductive fabric sensing layer, printed electrodes, and the PET substrate. The optimized multistructured film of poly(3,4-ethylenedioxythiophene) (PEDOT):poly(styrenesulfonate) (PSS)/PEDOT nanowires (NWs)/cellulose nanofiber (CNF) (PEDOT:PSS/PEDOT NWs/CNF fabric [PPCF]) is used as a sensing layer owing to the unique advantages of multiple structures and high conductivity, which confers high sensitivity, a wide detection range, and excellent flexibility. Screen-printed silver films were used as electrodes and a biocompatible gel was used as the packaging layer. The size of a single sensor was 3.3 mm Â 3.4 mm. The detailed fabrication process of the sensor is presented in the Experimental Section.
The performance of the pressure sensor was investigated in detail, as shown in Figure 3. The sensor was sensitive to a wide range of forces, and the current increased as the loading force increased in the range of 0-50 N (Figure 3b). Sensitivity is an important parameter for pressure sensors. Herein, the sensitivity of the obtained pressure sensor was first investigated. In general, the sensitivity (S) of the pressure sensor can be defined as [43] S ¼ δðΔI=I 0 Þ=δF where I 0 is the initial current without loading force and I is the current with loading force. ΔI is the relative change in current (IÀI 0 ), and F is the external force applied to the sensor surface. Similar to the literature, [44,45] the sensitivity of the sensor showed different changeable trends at different pressures. The optimized sensor unit achieved a high sensitivity of 99.56 N À1 in the low-pressure region (less than 10 N) and gradually stabilized to approximately 17.26 N À1 in the high-pressure region (10-50 N) (Figure 3c). Even when the loading force was 50 N, the prepared sensor maintained its pressure-sensing capability, demonstrating its potential application in soft tissue equilibrium measurements.
Moreover, the pressure sensor also exhibited a fast response to external forces, in addition to its high sensitivity and wide detection range. As shown in Figure 3d, the relative current change showed a fast response to the force from 0.8 to 50 N, indicating that the sensor had a good and fast sensing ability in a wide range of forces. The response and releasing times were 0.44 and 0.33 s www.advancedsciencenews.com www.advintellsyst.com ( Figure S5a,b, Supporting Information), respectively. Here, the response time is defined as the time required for the relative current change to increase from 10% to 90% of the peak value, while the release time is the opposite. In addition, the pressure sensor also showed a low hysteresis in the range of 0-50 N ( Figure S5c, Supporting Information), which may be attributed to the recoverability of the fabric. The relative current changes of the pressure sensor with repeated loading/unloading cycles under 8 N for 1000 cycles are shown in Figure 3e. The pressure sensor still showed a fast and stable response at 8 N after 1000 cycles. The relative current change was almost unchanged before and after 1000 loading cycles, demonstrating its remarkable electrical stability. Its high sensitivity, wide detection range, good stability, and fast response to pressure make it a good candidate for use in soft tissue measurements.

Performance of Sensor Arrays inside Smart Spacer
To further clarify the potential application of pressure sensors in the TKA force-sensing system, the performance of the pressure sensor arrays was investigated. First, the force distributions in different sensor arrays were investigated. As shown in Figure 4a, forces of 1, 1.6, 2.3, 3, 5, and 10 N were loaded on six sensor arrays attached to the smart surface. The six units  www.advancedsciencenews.com www.advintellsyst.com of the sensor arrays displayed the value and distribution of the loading forces in a timely manner. As shown in Figure 4b, the mapping color appeared at the corresponding areas of the sensor arrays, demonstrating their effectiveness and practicality. Moreover, the real-time responsiveness and sensitivity of the sensor arrays attached to the smart spacers are also illustrated. Forces were simultaneously loaded at different positions on the smart spacer. A schematic diagram of the sensor arrays is shown in Figure 5a (Figure 5b). As shown in Figure 5c, when one, two, and three points of the smart spacer experienced load forces simultaneously, the different voltage signals provided outputs correspondingly. The results demonstrate that the sensor arrays have stable electrical response characteristics and signal antiinterference capabilities.
The consistency of the sensor arrays was also investigated because it could affect the accuracy of the measuring system. The voltage-force curves of the six sensor units at different positions in the smart spacer are shown in Figure S6, Supporting Information. It can be seen that the curves nearly coincide in the range of 0-50 N, indicating that the sensor arrays have excellent stability. The output voltages of the different sensor units at 10, 30, and 50 N are shown in Figure S6b, Supporting Information. There was a small difference in the output values of each unit of the sensor array. At the same time, the performance of the sensor unloading force at different temperatures was investigated, as shown in Figure S7, Supporting Information. The output voltage did not change significantly when the temperature increased from 30 to 80 C, indicating that the sensor possessed the capacity to resist thermal disturbances. In addition, measurement errors in the detection range were also recorded in the range of 0-50 N ( Figure S8, Supporting Information). The maximum error was 0.08-0.14 N, which was less than 2%. The above results demonstrate that the sensor arrays attached to the smart spacer have good stability, excellent consistency, and signal anti-interference capabilities, which can meet the requirements of the TKA force-sensing system.

Knee Joint Simulation Test Based on Designed TKA Force-Sensing System
To further demonstrate the application of the proposed TKA force-sensing system, a knee joint simulation test system was fabricated, as shown in Figure 6a. The knee joint simulation test system consists of a smart spacer, the signal platform, and 3D printed physical models of the femur, tibia, femoral component, spacer, and tibial component ( Figure S9, Supporting Information). The femoral component and femur were installed together, as were the tibial component and tibia. The two parts constituted a model of the human knee joint, which was convenient for simulation tests. The designed TKA force-sensing system is used to measure soft tissue balance during TKA, which assists the surgeon in regulating the soft tissue. This soft tissue balance measurement process is shown in Figure 6b. When the smart spacer containing the pressure sensor arrays is inserted directly between the femur and tibia, the pressure sensors measure the contact stress caused by soft tissue tension at the two sides between the femur and tibia. The pressure sensor arrays then output the corresponding electrical signals at both sides. The circuitry then finally completes signal acquisition and processing. Thus, different electrical signals can show the actual stress and its distribution within the knee joint.
To directly display the differences in the forces at both sides of the joint, the output voltage signals need to be converted into force information because the measuring system outputs the relationship between the voltage and the pressure. The actual measurement signals between the force and voltage were fitted. The obtained functional relationship between the force and voltage is shown in Figure S10, Supporting Information. The curves conform to the characteristics of the power function. By optimizing the function parameters, the expression of the fitting function was obtained ( Figure S10b, Supporting Information), which is convenient for converting the collected voltage signals into corresponding force values. Thus, the force information at both sides of the joint can be directly obtained from the measured surface according to the fitted results.
According to clinical needs, the soft tissue measurement system mainly measures three-activity positions, including the extension position (0 ), the middle position (45 ), and the flexion position (90 ). A physical view of the test setup is presented in Figure S11, Supporting Information. The smart spacer is placed in the gap of the knee joint to test the contact stress caused by soft tissue tension. When the knee test model is extended, the tension in the soft tissue causes a change in the output signals of the sensor arrays on the smart spacer. The display interface can directly provide various information, such as force values, 3D histograms, and intensity images. The knee joint stresses at the three different angles were measured to simulate actual surgery for testing (Figure 6c and S12, Supporting Information).
As shown in Figure 6c-I, the pressure signal is greatest at the top position of the smart spacer, indicating that the contact force on the spacer in extension is concentrated at the topmost position. The differences between the medial and lateral contact forces can be clearly seen in the contact force distribution histogram in Figure S13a, Supporting Information. Meanwhile, the www.advancedsciencenews.com www.advintellsyst.com relationship between the magnitude of the lateral and medial soft tissue tension can be easily compared. According to Equation (1) in the supporting information, the imbalance coefficient of the knee joint in this state can be calculated. Thus, the surgeon can unitize the imbalance coefficient to adjust the balance of soft issues during TKA. Similarly, when the knee test model was in flexion at 45 and 90 , pressure signals appeared at the middle and bottom positions of the smart spacer, respectively, as shown in Figure 6c-II and III. From the contact force distribution histograms in Figure S13b,c, Supporting Information, the relationship between the magnitude of the lateral and medial soft tissue tension could be compared, and the imbalance coefficient of the knee joint in this state could be obtained. The real-time measurement process of the proposed TKA force-sensing system in the three-activity positions is shown in the Videos of Supporting Information. Based on the results, it is apparent that the measurement system can display force differences on different sides of the joints in a timely manner. Thus, the doctor can adjust the balance according to the results of Equation (1). The above results confirm that the proposed TKA force-sensing system is a novel strategy for the timely and accurate measurement of soft tissue balance in TKA, providing valuable information for surgeons performing intraoperative balance correction. Moreover, this system also provides a strategy to detect soft tissue balance for the knee joint replacement surgery robot, thus increasing the surgical success rate.

Conclusions
In summary, a novel, accurate, and real-time force measurement system based on flexible sensor arrays is proposed to assist surgeons in balancing soft tissue during TKA. The high-performance pressure sensor arrays based on conductive fabrics possess both high sensitivity and a wide detection range; thus, they can meet the surgical requirements for soft tissue balance measurement. The smart spacer has been constructed using multipositiondistributed flexible sensor arrays, which can directly sense soft tissue tension, resulting in more accurate and direct measurements than the traditional embedded spacer devices. Moreover, the soft tissue balance measurement system model was integrated with a smart spacer, multichannel signal acquisition platform, and 3D-printed knee joint prosthesis. The obtained system can measure the force of both sides of the knee joint at different positions, such as the extension position (0 ), middle position (45 ), and flexion position (90 ), as well as provide surgeons with a variety of selectable display interfaces about stress information for conducting better surgeries. This novel soft tissue measurement system provides a convenient way for surgeons or the knee joint replacement surgery robot to accurately evaluate soft tissue balance, and paves the way for improving the TKA success rate and the patient's postoperative rehabilitation.
Fabrication of Pressure Sensor Arrays Based on the PPCF: A sandwich structure of the pressure sensor was constructed. A conductive fabric film based on the PEDOT:PPCF was used as the sensitive layer, which was fabricated by an in situ polymerization method, similar to that reported in the literature. [47] The size of the well-cut conductive PPCF was 3.3 mm Â 3.4 mm Â 0.6 mm. The silver array electrodes based on the PET were fabricated using the screen-printing method. A mixture of PDMS prepolymer and curing agent (10:1) was poured into the premade mold and heated at 80 C for 1 h to form a flexible packaging layer. The prepared printed electrode, conductive fabric, and packaging layer were fixed together using silicone glue to form pressure sensor arrays.
Fabrication of the Smart Spacer Based on the Flexible Sensor Arrays: The smart spacer for the TKA force-sensing system consisted of a pressure sensor array, a knee joint spacer, and a signal acquisition platform based on a personal computer (PC). A groove structure was designed for the knee joint spacer to fix the flexible pressure sensor arrays. The pressure sensor arrays were fixed on this grooved spacer structure to form a smart spacer for soft tissue measurement. The homemade signal acquisition platform mainly included a data acquisition card, a signal conversion module, and a display interface. The data acquisition card was used in conjunction with a signal conversion module to convert the resistance signals into measurable voltage signals. The LabVIEW software converted the measured voltage signal into an image on a programmed display interface, which was convenient for the surgeon to observe and operate.
Fabrication of the TKA Force-Sensing System: First, a scanning photograph (512 Â 512) of the joint was measured. The scanning distance was 1 mm; the X-ray voltage was 120 kV. And next, according to the rules of the TKA, a simulation model of the femur, tibia, and femoral component was constructed. Then, the physical models of the femur, tibia, and femoral component were fabricated using the 3D printing method. The femoral component and femur were installed together, as were the tibial component and tibia. These two parts constituted a model of the human knee joint, which was convenient for simulation tests. Finally, the smart spacer tibial component and signal platform were integrated to form the TKA force-sensing system.
Characterizations and Measurements: The electric properties of the pressure sensor units were measured using a Keithley 4200 source meter. The loading force instrument used was a Mark-10. The properties of the pressure sensor arrays were measured using a homemade signal acquisition platform consisting of a data acquisition card, a signal conversion module, and a display interface. A National Instrument data acquisition card (USB-6210) was used to collect pressure signals caused by soft tissue tension. A signal conversion module was used to convert the resistance signal into a detectable voltage signal. The LabVIEW software based on the acquisition program could display acquired signals in real time on the PC.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.