Portable electronic devices for mechanotherapy and operative treatment of osteoarthritis

Abstract

Knee osteoarthritis (KOA) is fundamentally driven by abnormal mechanical loading and the subsequent loss of joint homeostasis. Effective therapeutic strategies, whether conservative or surgical, depend on the precise restoration of physiological kinematics and load distribution. This review synthesizes recent advances in sensor technologies designed to quantify the mechanobiological environment of the knee. In non-surgical management, wearable systems utilizing inertial measurement units (IMU) and flexible pressure sensors enable the continuous monitoring of gait cycles, joint angles, and muscle activation, providing objective data for neuromuscular rehabilitation. In surgical contexts, we analyze the evolution of intraoperative sensing from rigid force-sensing spacers to emerging soft electronics in total knee arthroplasty (TKA). A critical challenge remains in developing sensors with mechanical compliance similar to biological tissues and minimal thickness to fit the constrained joint space during joint-preserving procedures. We highlight the potential of novel transduction mechanisms—including piezoresistive, capacitive, piezoelectric, and triboelectric systems—to overcome these limitations. The integration of these flexible, self-powered technologies with data-driven analytics offers a pathway toward an integrated data-driven treatment framework, which could facilitate optimal biomechanical alignment and functional recovery.

Graphical abstract

1. Introduction

Osteoarthritis (OA) represents one of the most prevalent degenerative joint disorders worldwide, characterized by progressive cartilage degradation, altered joint mechanics and functional impairment.¹ As the understanding of OA pathophysiology evolves, clinical management has expanded from purely symptomatic control toward precision-oriented, biomechanics-informed therapeutic interventions. In this context, sensing technologies are emerging as a critical bridge that facilitates integration of joint function, mechanical environment and therapeutic decision-making.² Conventional clinical assessment modalities—including intermittent imaging examinations,³ manual goniometric measurements,⁴ and patient-reported outcome instruments⁵— offer only episodic and static evaluations. In contrast, emerging sensing technologies provide continuous, objective, and context-specific quantification of joint loading and locomotor function.

Current OA treatment strategies can be broadly categorized into non-surgical conservative therapy and surgical intervention, each presenting distinct monitoring demands. In conservative care, where exercise therapy, gait retraining and neuromuscular rehabilitation play essential roles, wearable sensing systems have revolutionized functional assessment.⁶ Commercially available platforms include integrated inertial measurement units (IMU)-enabled prosthetic extensions (e.g., Persona IQ) and skin-mounted multi-sensor patches (e.g., TracPatch). These devices now provide real-time tracking of joint angles, load symmetry, and muscle activity. They allow clinicians to monitor adherence, tailor rehabilitation protocols and identify aberrant motor patterns outside clinical environments. Furthermore, emerging flexible sensors based on piezoresistive, capacitive, piezoelectric and triboelectric mechanisms further enhance motion capture and soft-tissue biomechanical sensing, advancing conservative OA care toward personalized treatment plans based on sensor data.

In contrast, surgical management, including total knee arthroplasty (TKA) and high tibial osteotomy (HTO), demands precise control of intraoperative biomechanics.⁷ Restoring joint line alignment, ligament balance and compartmental load distribution is crucial for long-term implant survival and postoperative function.⁸ Traditional surgeries rely heavily on the surgeon's experience and visual estimation, which often makes mechanical axis correction and joint loading difficult to quantify. To address these limitations, sensor-assisted navigation systems, instrumented tibial inserts that measure joint forces, and robotic guidance platforms have been developed to provide real-time mechanical feedback during surgery. For TKA, instrumented polyethylene inserts, strain-based tibial prostheses, and piezoelectric -embedded platforms allow direct quantification of tibiofemoral forces.⁹ For HTO, fluoroscopy,¹⁰ mechanical alignment tools¹¹ and computer assisted navigation¹² assist in achieving accurate osteotomy angles and limb realignment. However, intra-articular load sensing remains limited. Recently, flexible and soft implantable sensors have shown promise in measuring multi-axis stresses within the joint space. This advancement paves the way for future biologically compatible, minimally invasive sensing strategies.

Advances in both non-surgical wearable monitoring and intraoperative implantable sensing represent a paradigm shift from episodic evaluation to continuous, mechanics-aware and personalized OA management. By integrating soft-material electronics, self-powered transduction mechanisms and machine-learning analytics, next-generation sensor technologies hold the potential to facilitate the integration of preoperative evaluation, intraoperative decision-making, and postoperative rehabilitation, moving toward a more interconnected treatment framework. To provide a rigorous perspective on this interdisciplinary evolution, we surveyed relevant literature from databases including Web of Science, PubMed, IEEE Xplore, and Google Scholar. The search focused on studies published between 2010 and 2025, utilizing keywords bridging clinical orthopedics (e.g., TKA, HTO, biomechanics) and engineering innovations (e.g., piezoresistive, flexible electronics). Accordingly, this review summarizes recent progress in sensor-enabled OA management across the domains of non-surgical conservative care and surgical intervention, highlights emerging materials and device architectures, and discusses future directions toward fully intelligent orthopedic care ( Fig. 1).

Fig. 1.

Application of portable electronic devices in mechanotherapy and surgical management of osteoarthritis.

2. Knee anatomy and biomechanics

The integration of sensing technology into knee osteoarthritis (OA) management is not merely an engineering challenge but a biomechanically driven one. The fidelity of acquired data depends less on the sensor's intrinsic sensitivity and more on its strategic placement within the joint's load transmission pathway. Consequently, establishing a robust understanding of the knee's functional anatomy is a prerequisite for identifying optimal sensor locations and interpreting meaningful physiological signals. To provide a theoretical roadmap for device deployment, this section dissects the knee's mechanobiological environment into two functional domains: Section 2.1 analyzes the intra-articular structural determinants that govern static load distribution, while Section 2.2 examines the extra-articular dynamic stabilizers that control joint kinematics. This framework clarifies the distinct sensing requirements for monitoring internal contact stresses versus external dynamic stability.

2.1. Intra-articular structures and contact pressure

The knee joint is the largest hinge joint in the human body. It is formed by the femur, tibia, and patella, primarily facilitating flexion and extension movements. The contact surfaces of the femoral condyles at the distal end of the femur and the corresponding surfaces on the tibial plateau form two crescent-shaped articular surfaces. The patella is located anterior to the distal femur, embedded within the femoral trochlea. Articular cartilage covers the articular bone surfaces, creating a low-friction sliding surface that effectively protects the bones from direct impact and wear, thereby ensuring smooth and durable joint movement.¹³

As illustrated in Fig. 2, structural integrity is paramount for maintaining a physiological mechanical environment. During normal activities like walking and running, the reaction forces and stress distribution within the knee follow specific patterns, where coordinated muscle action distributes pressure evenly across articular surfaces, preventing excessive local stress concentration.¹⁴ However, this delicate balance is highly vulnerable in pathological states. In the state of OA (right side of Fig. 2), degeneration of the knee cartilage and meniscal damage alter the morphology of the articular surfaces, leading to abnormal stress distribution and noticeable local stress concentration. Cartilage degeneration roughens the articular surfaces, worsens tribological properties, increases mechanical wear during movement, and accelerates the joint degeneration process. Meniscal damage reduces the cushioning and dispersing effect on joint load, further exacerbating load imbalance. In reconstructive surgeries such as total knee arthroplasty (TKA) or high tibial osteotomy (HTO), the primary surgical goal is to reverse this pathology and restore ideal load distribution. Unfortunately, traditional static imaging (X-ray/CT) or subjective intraoperative “manual feel” cannot quantify contact forces during dynamic manipulation.¹⁵ This necessitates the integration of contact pressure sensors to provide objective intraoperative feedback ensuring uniform load distribution across the reconstructed joint surfaces.

Fig. 2.

Functional Anatomy and Biomechanical Loading Patterns of the Knee: Normal vs Osteoarthritis (OA).

Bridging the femoral and tibial articular surfaces, the meniscus (or its prosthetic equivalent) serves as the optimal “anchor point” for intra-articular sensing. The menisci are unique fibrocartilaginous structures of the knee joint, consisting of the medial and lateral menisci, located between the tibial plateau and the femoral condyles. The menisci not only act as cushioning structures in distributing joint load but also play a crucial role in maintaining joint stability. Moreover, their morphology and thickness vary across different regions, adapting to the specific patterns of load distribution.¹⁶ Studies indicate that the menisci transmit 50–70% of the compressive load in extension and up to 85% in flexion.¹⁷ Consequently, smart inserts or pressure sensors are invariably designed to mimic the meniscal geometry, situated within the joint gap to capture the authentic tibiofemoral contact pressure.

2.2. Extra-articular muscles and movement control

The muscular system surrounding the knee plays a pivotal role in both movement and stability of the joint. The major muscle groups involved are the quadriceps, hamstrings, sartorius, gracilis, gastrocnemius, popliteus, and the iliotibial band. Beyond the static intra-articular structures, the periarticular dynamic musculature represents another critical domain for sensor monitoring. Unlike the passive support provided by bones and ligaments, the muscular system acts as a “dynamic stabilizer,” actively modulating joint stiffness and absorbing ground reaction forces through coordinated contractions.

The quadriceps femoris, located anteriorly, is the primary extensor of the knee and a crucial stabilizer. It is essential for activities like walking, running, and jumping. The function of the quadriceps is particularly critical for knee stability and load-bearing capacity. Weakness or dysfunction of the quadriceps often leads to limited knee motion, pain, and functional impairment.¹⁸ Conversely, the hamstrings are primarily responsible for knee flexion and hip extension.^19,20 Their synergistic action contributes to dynamic stability, particularly during weight-bearing and postural transitions, by preventing hyperextension or excessive rotation.²¹

However, this precise muscular control is often compromised in OA patients. Studies have found decreased muscular coordination around the knee in OA patients, leading to altered movement patterns that further exacerbate abnormal loading.^22,23 Insufficient muscle strength or poor muscular coordination causes the knee to bear abnormal mechanical loads during movement, stimulating pathological changes in the articular cartilage and menisci. Furthermore, unstable movement postures, such as abnormal torsion of the knee or excessive flexion/extension angles, can cause localized instantaneous high pressure, leading to acute injuries like meniscal tears and ligament ruptures. Both clinical and biomechanical research suggest that maintaining stability during knee movement and coordinated muscle force generation are crucial for preserving mechanical balance in the joint and preventing injury.^24,25

A classic clinical phenomenon is “arthrogenic muscle inhibition (ami)”, where joint effusion or pain reflexively inhibits quadriceps activation.²⁶ This dynamic functional deficit acts as a “silent killer” accelerating disease progression, yet static imaging (e.g., MRI/CT) fails to diagnose these anomalies in muscle activation timing. By strategically deploying electromyography (EMG) sensors or inertial measurement units (IMU) over key muscle²⁷ bellies like the rectus femoris and biceps femoris, we can capture muscle co-contraction patterns. This data serves not only to assess joint stability but also as a biofeedback signal, guiding patients through precise rehabilitation to break the vicious cycle of muscle atrophy and joint degeneration.

3. Mechanobiology in knee osteoarthritis: From pathological progression to stage-specific treatment

Consistent with the aforementioned biomechanical framework, knee osteoarthritis is essentially a disease of biomechanical imbalance. While it manifests structurally as a whole-joint organ disease involving cartilage, synovium, and subchondral bone, its progression is relentlessly driven by the abnormal mechanical inputs—such as stress concentration and joint instability—described in Section 2. These mechanical stimuli are not benign; through mechanotransduction, they trigger inflammatory cascades and matrix degradation, creating a stage-dependent pathology that demands tailored therapeutic interventions.

3.1. Stage-specific treatment strategies

OA is commonly stratified into early, mid, and late stages. Stage-specific alterations in the mechanical environment critically influence therapeutic choices (Fig. 3).

Fig. 3.

Stage-specific mechanobiological management strategies for knee osteoarthritis.

Early KOA: This stage is characterized by mild cartilage degeneration and altered functional joint loading. Management is anchored in non-pharmacological care, including: (1) an individualized, multicomponent management plan; (2) information, education, and self-management; (3) exercise with appropriate dose tailoring and progression; (4) an appropriate mode of exercise delivery; (5) maintenance of healthy weight and weight loss; (6) footwear, walking aids, and assistive devices; (7) work-related advice; and (8) behavior-change techniques to support healthier lifestyles.²⁸ If activity-related pain continues to impede exercise and weight management, short-term topical NSAIDs may be used, with oral NSAIDs considered when necessary.²⁹

Mid-stage KOA: With progression, cartilage degeneration accelerates and load distribution becomes more uneven, driving subchondral stress concentration and osteophyte formation. On a foundation of conservative care, intra-articular interventions can be used as an adjunct for short-term pain relief.^30,31 For patients with marked malalignment or excessive medial compartment loading, the strongest evidence for load reduction and outcome improvement supports “precision exercise prescription plus behavioral support”; adjuncts such as unloader bracing, contralateral cane use, and stable supportive footwear/custom biomechanical insoles can reduce pain and improve adherence.²⁸ When indicated, around-the-knee osteotomies—high tibial osteotomy (HTO) and/or distal femoral osteotomy (DFO)—can correct limb alignment and improve the local biomechanical milieu.^32,33

Late-stage KOA: Severe cartilage degeneration with aberrant load distribution results in deformity and major functional impairment. Management centers on knee arthroplasty, with goals of restoring alignment, increasing effective contact area, optimizing contact stresses, and re-establishing biomechanical stability of the knee.³⁴

3.2. Mechanotherapy

Non-pharmacological treatments aim to relieve OA symptoms and improve physical function, which are regarded as core management strategies. Exercise is the most effective non-pharmacological intervention, with high safety and notable cost-effectiveness. The latest meta-analyses show that aerobic exercise not only improves short-term quality of life (standard mean difference (SMD) = 1.53, 95% confidence interval (CI) 0.47 to 2.59) and reduces pain (SMD = −1.10, 95% CI −1.68 to −0.52), but also enhances mid-term physical function (SMD = 1.78, 95% CI 1.05 to 2.51) and gait performance (SMD = 0.85, 95% CI 0.55 to 1.14), yielding the most favorable overall efficacy. Mind–body exercises such as tai chi and yoga help improve short-term physical function; neuromotor exercise helps improve short-term gait performance; strengthening exercise and mixed exercise help improve mid-term physical function; and flexibility exercise helps alleviate long-term pain 35. Multimodal interventions (e.g., education + exercise + dietary management) outperform single interventions.³⁶

Accordingly, self-management plays an important role in non-pharmacological care for OA. Key components include information provision and education, digital support, self-monitoring, behavior-change techniques, goal setting, and action planning. By integrating multiple approaches (e.g., education and structured exercise programs) with ongoing follow-up, patients can better control symptoms, improve function, and enhance quality of life. Wearable devices can monitor physical activity, gait, joint range of motion, heart rate, and other parameters, providing real-time data to help clinicians and patients manage OA more precisely, facilitate self-management, track activity status, and support remote health management.^37,38 Although current evidence is still limited, with technological advances, wearables hold substantial promises for future OA management.

3.3. Operative treatment

3.3.1. Around-the-knee osteotomies

Around-the-knee osteotomies re-establish mechanical balance by altering the joint line/limb alignment, particularly suitable for younger, more active patients. From a mechanobiological perspective, the procedure functions as a “load-shifting” intervention. By structurally adjusting the tibial mechanical axis, HTO redirects the weight-bearing line (WBL) from the degenerated medial compartment to the relatively healthy lateral compartment (typically targeting the Fujisawa point at 62% of the tibial width³⁹). This mechanical unloading reduces contact stress on medial cartilage, thereby creating a favorable mechanobiological environment for fibrocartilaginous regeneration and symptom relief.

3.3.2. Knee arthroplasty

Total or unicompartmental knee arthroplasty is the gold standard for late-stage OA. Unlike HTO, which preserves native anatomy, TKA involves resurfacing the articular interface with prosthetic components. Mechanically, the success of TKA depends not merely on geometric alignment but on soft tissue balancing. The surgical objective is to ensure that the extension and flexion gaps are equal and rectangular, ensuring uniform stress distribution across the polyethylene insert.^40,41 Any imbalance can lead to eccentric loading, accelerated wear, and aseptic loosening, highlighting the critical need for precise load management.⁴²

While HTO and TKA employ divergent strategies—realigning the native axis versus resurfacing the joint—their long-term success converges on a single determinant: the precision of mechanical reconstruction. Whether it is preventing stress concentration from over-correction in HTO or avoiding eccentric loading in TKA, the margin for error is narrow. However, traditional reliance on static imaging and subjective manual feel leaves a critical blind spot regarding the dynamic mechanical environment. Consequently, unquantified overload or imbalance remains the primary driver of postoperative failure (e.g., varus recurrence or aseptic loosening). This clinical reality underscores the imperative for the sensor technologies discussed in the following sections, which aim to transform these subjective procedures into data-driven precision surgeries.

4. Sensors currently applied in non-surgical and surgical treatments of osteoarthritis

4.1. Monitoring of exercise therapy

With the growing adoption of digital orthopedics and remote rehabilitation concepts, postoperative and exercise rehabilitation monitoring has gradually evolved from traditional outpatient follow-ups and imaging assessments to continuous, home-based, and wearable digital health models. Currently, the most established systems in clinical and commercial applications mainly include implantable motion monitoring devices and multi-sensor surface rehabilitation patche 43, 44, 45, 46. Representatively, the Persona IQ,⁴⁷ jointly developed by Zimmer Biomet and Canary Medical, integrates an accelerometer and an inertial measurement unit (IMU) within the tibial implant. This system enables long-term recording of parameters such as joint range of motion, step count, cadence, and gait symmetry, which are wirelessly transmitted to a remote monitoring platform. Another representative example is the TracPatch series of surface motion-tracking systems,^48,49 which utilize dual-patch IMUs to record knee flexion–extension angles, gait rhythm, and rehabilitation compliance. These systems provide real-time motion feedback for early postoperative functional training and home-based rehabilitation. Some versions are integrated with remote physiotherapy programs and cloud-based analytics platforms, supporting personalized rehabilitation planning and outcome tracking. In addition, several clinically validated wearable systems for lower-limb rehabilitation have emerged on the market. For instance, the Moticon smart insole⁵⁰ is used for gait load and balance training analysis; the Delsys 51, 52, 53 and Noraxon 54, 55, 56 electromyography systems are applied in monitoring quadriceps reconstruction and strength training after surgery; and smart knee braces integrating joint angle and gait monitoring functions have also been developed. Overall, existing rehabilitation devices primarily focus on quantifying movement behavior and muscle function, but limitations remain in soft tissue mechanosensing, energy management, and long-term unobtrusive monitoring. With advances in flexible electronics, skin-conformal sensors, and self-powered materials, monitoring during the rehabilitation phase is evolving from simple “motion tracking” toward an intelligent rehabilitation paradigm that integrates tissue mechanics, electrophysiology, and behavioral data. This shift enables more refined and dynamic sensing capabilities for postoperative management following joint replacement.^57,58

4.2. Monitoring of intraoperative procedures

In the surgical treatment of knee osteoarthritis, whether in total knee arthroplasty (TKA) or high tibial osteotomy (HTO) aimed at restoring limb alignment and delaying joint degeneration, precise intraoperative control of the knee's mechanical environment is of critical importance.⁵⁹ In a normal knee joint, balanced medial–lateral loading, consistent joint spacing, and stable alignment are maintained during walking, flexion–extension, and weight-bearing. However, traditional surgical procedures rely heavily on the surgeon's experience and tactile feedback, making it difficult to accurately assess joint contact pressure, soft tissue tension, and imbalances in three-dimensional kinematics. As a result, postoperative complications such as pain, prosthetic loosening, uneven wear, or overcorrection leading to persistent medial compartment stress abnormalities may occur. Therefore, intraoperative monitoring technologies are essential.^60,61 Representative systems such as Mako (Stryker)^62,63 and ROSA (Zimmer Biomet)^64,65 are optical/inertial navigation platforms that use preoperative image registration and reflective marker tracking to provide real-time information on bone segment orientation, limb alignment, and knee motion trajectories. Meanwhile, intelligent spacer systems such as VERASENSE,⁶⁶ which incorporate piezoresistive sensor arrays, can directly measure joint contact pressure, enabling quantitative adjustment of soft tissue balance and load symmetry.

However, most of the aforementioned systems are designed for TKA and are not suitable for HTO procedures. Compared with joint replacement, high tibial osteotomy (HTO) focuses less on prosthesis interface loading and soft tissue balance; its primary goal is to adjust the proximal tibial mechanical axis to redistribute knee joint loading, thereby reducing medial compartment stress, delaying cartilage degeneration, and improving functional outcomes.⁶⁷ Therefore, accurately obtaining intraoperative lower-limb mechanical axis and osteotomy angle information is a key factor influencing the success of HTO.⁶⁸ Current clinical monitoring systems primarily rely on imaging and navigation assistance, including intraoperative C-arm fluoroscopy and mechanical alignment rod verification, as well as computer-assisted navigation systems based on preoperative CT/MRI planning and intraoperative bone marker tracking, such as the Brainlab^69,70 navigation platform and certain navigation modules from Stryker.⁷¹ These systems can provide real-time visualization of changes in the mechanical axis, osteotomy opening or closing angles, and the weight-bearing line (WBL), thereby improving osteotomy precision and consistency in individualized alignment correction. Additionally, patient-specific instruments (PSI) and 3D-printed osteotomy guides can assist in reducing reliance on surgeon experience through preoperative personalized planning and precise intraoperative positioning. However, current HTO monitoring approaches remain primarily focused on geometric and alignment accuracy and cannot directly quantify changes in peri-knee soft tissue tension or load redistribution after osteotomy. They also lack real-time biomechanical feedback mechanisms similar to those provided by intelligent spacers or instrumented implants. As a result, it remains challenging to dynamically assess the degree of correction and postoperative biomechanical evolution during HTO, leading to risks of undercorrection, overcorrection, or long-term regression of the mechanical axis in some patients 72, 73, 74. Overall, current HTO monitoring systems have progressed from traditional “image-based positioning” toward “personalized and precise alignment correction,” but they have yet to achieve mechanically feedback-driven precision adjustment. The integration of flexible, implantable, and dynamic sensing technologies holds promise for addressing this critical gap.

5. Wearable technologies in osteoarthritis treatment and rehabilitation

5.1. Overview of wearable sensor technologies (piezoresistive, capacitive, triboelectric, and piezoelectric)

With the rapid advancement of flexible electronics and smart materials, wearable sensors have become a crucial bridge between physiological signals and digital health monitoring. Based on their mechano-electrical signal transduction mechanisms, current wearable sensing technologies can be categorized into four main types: piezoresistive, capacitive, triboelectric, and piezoelectric sensors. The following sections provide an introduction to their fundamental principles (Fig. 4).

Fig. 4.

The working principles of four wearable sensor technologies: piezoresistive, capacitive, piezoelectric, and triboelectric sensors.

Piezoresistive Sensors are a class of sensors in which the material's electrical resistance changes in response to applied force. The core principle is that when the sensing element—typically composed of conductive polymers,⁷⁵ carbon-based materials,⁷⁶ Metal–organic frameworks (MOFs)⁷⁷—is subjected to external stress, strain, or pressure, its internal conductive pathways are altered, resulting in measurable changes in electrical resistance. These resistance changes primarily arise from the following mechanisms:

• 1)Geometric effect: External force alters the material's length and cross-sectional area; according to R = ρ⋅L/A, the resistance changes with the geometry.
• 2)Intrinsic piezoresistive effect: In certain semiconductors or sensitive materials, strain modifies the band structure or carrier mobility, leading to changes in resistivity.
• 3)Interfacial/tunneling effect: In composite conductive networks (e.g., carbon nanotubes (CNT),⁷⁸ graphene,⁷⁹ or Mxene⁸⁰), applied pressure increases particle contact or shortens tunneling distances, resulting in a decrease in overall resistance.

By correlating resistance changes with external mechanical stimuli, piezoresistive sensors can effectively convert force, pressure, or deformation into electrical signals.

Capacitive Sensors operate based on the parallel-plate capacitor principle: C = εrε0A/d, where C is the capacitance, εr is the dielectric constant, ε0 is the vacuum permittivity, A is the effective electrode area, and d is the electrode spacing. During sensing, external stimuli can change the electrode spacing, the overlapping electrode area, or the dielectric constant of the medium, leading to measurable variations in capacitance. Common mechanisms include:

• 1)Distance-change type⁸¹: External force reduces the spacing between electrodes, increasing the capacitance.
• 2)Dielectric-change type⁸²: Introduction of different materials or human tissue into the electric field changes the effective dielectric constant.

Triboelectric Sensor⁸³ operate based on the coupled principles of contact electrification and electrostatic induction. When two materials come into contact and then separate, electrons are transferred between their surfaces due to differences in electron affinity, generating equal and opposite charges at the interface—i.e., triboelectric charging. Subsequently, when electrodes are connected to form a circuit, external forces periodically alter the relative position or contact area between the materials, changing the electric field distribution. To maintain charge balance, electrons flow through the external circuit, producing measurable current or voltage signals.

Piezoelectric Sensors operate based on the piezoelectric effect. When materials with non-centrosymmetric crystal structures, such as quartz, lead zirconate titanate (PZT)⁸⁴ ceramics, or the polymer polyvinylidene fluoride (PVDF),⁸⁵ are subjected to external mechanical stress, such as pressure, tension, bending, or vibration, their internal dipole moments reorient or deform. This generates induced charges on the electrodes, producing a potential difference and an electrical signal between them. By measuring the output charge, voltage, or current, these sensors can quantify the magnitude, frequency, and direction of the applied force.

Overall, these four sensing mechanisms collectively form the core technological framework of novel wearable systems. However, their clinical applicability is strictly dictated by their fundamental signal generation mechanisms. Because piezoresistive and capacitive sensors function by modulating intrinsic material properties (resistance or capacitance) that remain stable under constant strain, they are inherently suitable for static or quasi-static applications. This non-decaying signal characteristic makes them the only viable option for tasks requiring sustained load quantification, such as ligament tension when the leg is held stationary. Conversely, because piezoelectric and triboelectric nanogenerators rely on transient charge generation driven by dynamic polarization or electrostatic induction, their output signals naturally leak and decay to zero when motion ceases. Consequently, while unsuitable for static weighing, they are uniquely advantageous for high-frequency dynamic scenarios, such as real-time gait analysis and energy harvesting, where detecting rapid motion events and converting kinetic energy are the primary objectives.

To gain a more comprehensive understanding of the quantitative indicators of sensors based on different mechanisms in the treatment of osteoarthritis, we have created Table 1, which includes key performance metrics such as sensitivity, response time, and linearity. Beyond electromechanical specifications, assessing the translational maturity of these technologies is critical for understanding their current feasibility in orthopedic practice. Therefore, we evaluated the development status of the cited studies based on the standard technology readiness level (TRL) framework.⁸⁶

Table 1.

Comparison of portable electronic devices for OA applications.

Mechanism	Materials	Application	Sensitivity	Response Time	Linearity/Rang	Fatigue Life/Stability	Biocompatibility	Development Status (TRL)	Ref.
Piezoresistive	Electric conductive rubber sensors	Gait rehabilitation training					Wearable	TRL 6-7	87
	Hartmann hydrocoll	Monitoring system for knee extensor training			0-90°	>40 compression cycles	Wearable	TRL 4-5	88
	Silica nanoparticle-reinforced polydimethylsiloxa-ne	Implantable pressure sensors for TKA	Pressure<3 MPa:5.22 MPa−1 Pressure>3 MPa:70 MPa−1	30 ms	45 Pa - 4.1 MPa	>5000 compression cycles	<24 h in vivo	TRL 4-5	89
Capacitive	Polyacrylamide gel(PAAm)/segmented embedded hydrogels (SEHs)	Cartilage-inspired strain sensor for motion	Strain<150% GF = 6.92, Strain>150% GF = 21.98	93 ms	0–350% strain	>600 compression cycles	Good	TRL 4-5	90
Capacitive	P(EA-co-AN) ionogel	Implantable pressure sensors for TKA	2.43 – 2.48 kPa−1		R2 > 0.998, 0 – 2.0 MPa	43 years	Excellent	TRL 5-6	91
Triboelectric	Styrene-ethylene-butylene-styrene (SEBS)/CNT/BaTiO3 (BTO)	Rehabilitation assistance (quadriceps, pulse, plantar)	0.068 kPa−1 (<53 kPa); 0.013 kPa−1 (>53 kPa)	25 ms	0–660 kPa		Skin-inspired, breathable, permeable	TRL 4	92
	CNT/acetylene black/polydimethylsilox-ane (PDMS)	Smart Insole for gait monitoring and recognition	Nonlinear synergistic strategy		R2 > 0.999 (0–225 kPa)	>180,000 compression cycles	Wearable	TRL 4-5	93
	PVA-borax hydrogel, polydopamine (PDA)-CNTs	Muscle motion monitoring, photothermal treatment				Self-Healing: recovers in 10 min	Wearable	TRL 4	94
Triboelectric	Conductive sponge, polyurethane sponge	Muscle telescopic monitoring	Adjustable sensitivity	2 ms	0.001-0.010 N	>360,000 compression cycles	Wearable	TRL 4	95
	Nylon/Eco-flex	Metaverse human-machine interaction	0-5 Hz		1.0-4.5 mm		Wearable	TRL 4-5	96
	Polytetrafluoroethy-lene (PTFE), copper foil	Smart healthcare	15°		15°–120°	>5000 compression cycles	Wearable	TRL 4-5	97
	Porous PDMS/Copper	Rehabilitation, Muscle function	1.76 V/kPa	∼50 ms	R2 = 0.092,0-150 kPa	>100,000 compression cycles	Wearable	TRL 4-5	98
	PVDF/BTO/CNT	Gait analysis, motion monitoring		66 ms	<100 kPa	>2000 compression cycles	Wearable	TRL 5-6	99
Triboelectric	Thermoplastic polyurethane (TPU)	Implantable pressure sensors for TKA			R2 = 0.99, 0-350N	>30,000 compression cycles		TRL 5-6	100
Piezoelectric	Multi-layered PVDF	Implantable pressure sensors for TKA	8.5 mV/N					TRL 5-6	101
	Boron Nitride Nanotubes (BNNTs)/PDMS	Joint torque monitoring, knee load	∼120 mV/(kPa·wt%)	<50 ms		>5000 compression cycles	Wearable	TRL 4-5	102
Inertial	Commercial wearable sensors	Rehabilitation exercise assessment (KOA)					Wearable	TRL 6-7	103
	Multi-sensor accelerometer units	Predict improvement after exercise intervention					Wearable	TRL 6-7	104
Inertial	Commercial Pressure Mat	Pelvis-trunk motion and OA detection					External use	TRL 6-7	105
Piezoelectret	Expanded polytetrafluoroethylene (e-PTFE)	Implantable pressure sensors for TKA	9 mV N−1	93 ms	R2 = 0.992 (1.4 - 13.6 N)	>4000 compression cycles		TRL 5-6	106

Reflecting the translational focus of this review, we categorized the technologies into three pivotal development stages: (i) TRL 4 (laboratory validation): Representing early-stage prototypes tested on benchtop setups or healthy human volunteers (e.g., Refs ^92,94); (ii) TRL 5 (relevant environment validation): Referring to advanced prototypes validated in vivo animal models (e.g., porcine or ovine knees) or ex vivo cadaveric studies (e.g., Refs 91, 100, 106) to simulate physiological loading; (iii) TRL 6–7 (clinical pilot demonstration): Denoting system prototypes that have advanced to pilot clinical trials involving actual OA patients or intraoperative surgical demonstrations, representing the critical leap from laboratory to clinic.

5.2. Non-surgical external sensors

In the stages of preoperative disease management and postoperative rehabilitation, wearable sensing technologies offer new possibilities for monitoring joint function and evaluating rehabilitation training. Patients with osteoarthritis often exhibit altered movement patterns, abnormal joint loading, and reduced muscle strength during daily activities, and their postoperative recovery relies on continuous tracking of joint range of motion, gait parameters, and muscle function.¹⁰⁷ Compared with traditional imaging and clinical scoring methods, wearable sensors enable continuous, dynamic, and noninvasive monitoring, capturing gait cycle,¹⁰⁸ joint angles,¹⁰⁹ muscle activity,¹¹⁰ and pain-related biosignals.¹¹¹ This capability provides patients with more refined functional assessment and rehabilitation guidance. With advances in flexible materials and artificial intelligence algorithms, these devices are transitioning from laboratory validation to everyday wearable platforms, offering high-precision and low-cost solutions for digital rehabilitation and home-based management of osteoarticular disorders.^112,113

New wearable devices can continuously capture joint motion trajectories in a non-invasive manner. For example, Chen et al.¹¹⁴ developed a rehabilitation assessment system based on a tri-sensor array. By placing sensors on the chest, thigh, and lower leg, the system enables quantitative analysis of knee flexion–extension angles and movement stability, allowing patients to manage their rehabilitation independently at home. Chang et al.¹⁰² designed a sensor capable of real-time knee torque monitoring (Fig. 5a). With AI-based estimation and overload-warning functions, the system achieves 99.8% accuracy in torque measurement and is suitable for home-based rehabilitation.

Fig. 5.

a) Schematic diagram of AI-assisted knee joint monitoring.¹⁰² Copyright 2025, Springer Nature. b) Diagram and physical prototype of the self-powered smart insole.⁹³ Copyright 2025, AAAS. c) Preparation process of the biomimetic gradient-fiber skin pad and its application in rehabilitation guidance.⁹² Copyright 2025, Springer Nature. d) Schematic diagram of machine-learning-assisted muscle monitoring.⁹⁸ Copyright 2024, ELSEVIER.

Advances in flexible pressure-sensing technology have provided new approaches for dynamic plantar pressure monitoring in gait analysis. The self-powered smart insole system developed by Wang's team⁹³ (as shown in Fig. 5b) uses a nonlinear synergistic pressure-sensing mechanism to capture real-time pressure variations across different plantar regions during walking, with data transmitted wirelessly to mobile devices. This system can identify abnormal gait patterns, such as the compensatory foot eversion commonly observed in patients with knee osteoarthritis, thereby offering support for tailoring rehabilitation training plans. In addition, the wearable training system developed by He et al.⁸⁷ which integrates motion sensors with pressure-sensitive conductive rubber, enables real-time extraction of gait parameters and provides feedback cues to guide patients in adjusting their walking posture. This effectively reduces the knee adduction moment and alleviates joint loading.

Innovations in muscle-strength signal monitoring stem from the integration of high–signal-to-noise-ratio materials with adaptive algorithms. Jiang et al.⁹² developed a skin-mimicking fibrous pad featuring a gradient porous architecture. Its tissue-like modulus ensures intimate contact between the electrode and skin, markedly reducing motion artifacts during electromyography acquisition (Fig. 5c). This sensor can accurately capture activation timing and intensity variations of key muscle groups, such as the quadriceps and hamstrings, during rehabilitation training, providing objective indicators for muscle-function assessment. Liu et al.⁹⁸ combined a porous triboelectric nanogenerator (TENG) array with machine-learning algorithms to construct an intelligent rehabilitation-monitoring system. The platform achieves high accuracy in classifying movement patterns, demonstrating the potential of multimodal data fusion for enhancing monitoring precision (Fig. 5d).

Emerging wearable systems can adapt rehabilitation programs based on patients’ daily physiological responses. Evidence shows that although strength training benefits most individuals with knee osteoarthritis, responses vary markedly across patients.¹¹⁵ By integrating wearable-derived acceleration data with patient-reported outcomes, data-driven models can identify responders to specific interventions and predict rehabilitation trajectories, enabling more precise allocation of clinical resources. Wearable feedback platforms also support gait correction. He et al. demonstrated that real-time feedback effectively guides users to adjust gait patterns and reduces knee adduction moment, thereby lowering joint load.⁸⁷ Likewise, Chen et al. developed a three-sensor rehabilitation system that allows patients to regulate training intensity at home, preventing both over- and under-loading.¹¹⁴ Overall, continuous monitoring combined with adaptive feedback enhances training safety, promotes personalized progression, and improves long-term adherence in musculoskeletal rehabilitation.

Despite these advancements and innovations, current technologies still face challenges such as signal drift and insufficient adaptability to individual differences. Future research needs to further optimize sensor calibration procedures and develop intelligent algorithms capable of adaptively adjusting gain and filtering parameters¹¹⁶ to meet the dynamic needs of different patient groups during rehabilitation.

5.3. Implantable intraoperative sensors

In recent years, advances in flexible electronics and implantable sensing technologies have driven increasing interest in intraoperative stress and displacement monitoring. These sensors can be embedded at the prosthesis–bone interface or within surrounding soft tissues to enable highly sensitive, real-time, and visualized detection of mechanical loads. By providing dynamic feedback during surgery, they assist surgeons in optimizing osteotomy angles, implant alignment, and soft-tissue tension, offering new avenues for precision arthroplasty and individualized rehabilitation.

The evolutionary trajectory of intraoperative monitoring began with the pioneering work of Kaufman et al., in 1996,¹¹⁷ who established the foundational concept of mechano-electrical transduction by embedding piezoresistive units into a tibial spacer to quantify prosthetic loads. By measuring the flexural deformation of four strain-sensing diaphragms arranged beneath the spacer, mechanical stresses were converted into electrical signals, enabling real-time assessment of vertical load and pressure center. This prototype marked the inception of intraoperative stress sensors and established the foundational concept of mechano-electrical transduction. Building upon this groundwork, in 2006, D'Lima et al.¹¹⁸ first developed an implantable sensor integrated into a total knee arthroplasty for in vivo data collection. This device comprised four embedded load sensors and a miniature transmitter for wireless communication. Subsequently, researchers began exploring mechanical sensing structures centered on piezoelectric ceramics, such as PZT. A representative advancement is Safaei's 2018 study,⁹ in which piezoelectric ceramics were embedded into a high–molecular-weight polyethylene spacer to directly convert stress changes into electrical charges. Compared with earlier piezoresistive strain-based sensors, this type of device offers several advantages: it is self-powered, exhibits rapid response, provides stable signal output, and enables multi-axis detection. Moreover, by integrating signal conditioning and charge amplification modules, the system achieves real-time monitoring of stress distribution and dynamic loads. Thus, intraoperative sensing technology has evolved from externally wired strain-based sensors to self-generating piezoelectric systems, offering new pathways for long-term prosthetic load monitoring and intelligent feedback.

In recent years, rapid advances in smart materials have enabled intraoperative joint-cavity stress monitoring to move beyond simply embedding a few sensors in spacers, offering a wider range of material options and structural configurations. Shi et al.⁹¹ developed a pressure sensor array based on a humidity-insensitive ionic gel, capable of stable operation in high-humidity environments. The sensor achieves an angular resolution of 0.1° and a pressure resolution of 0.1%, making it suitable for use in the synovial fluid–filled environment of the knee joint (Fig. 6a). Li et al.⁸⁹ proposed a strain-effect–based piezoresistive sensor in which an interlaced micro-dome structure induces tensile strain in a conductive thin film. This design achieves an ultra-wide detection range of 45 Pa to 4.1 MPa and a sensitivity of 5.22–70 MPa⁻¹, effectively unifying high sensitivity with a broad monitoring range (Fig. 6b). Chahari et al.¹⁰¹ systematically compared the performance of piezoelectric nanogenerators (PENG) and triboelectric nanogenerators (TENG) in knee joint implants. The study showed that multilayer PVDF-based PENG sensors can accurately track dynamic force distributions, while micro-cube patterned SR/BT@PDA TENGs demonstrate superior energy-harvesting performance, achieving a maximum output power of 6 μW. Wu et al.¹⁰⁶ developed a fully encapsulated flexible sensor based on a piezoelectric electret, featuring a piezoelectric coefficient of 23.8 pC/N and a fast response time of 93 ms. The sensor demonstrated remarkable linearity (R² = 0.992) between force and voltage, which significantly minimized the need for complex on-chip signal conditioning. This signal fidelity enabled the system to realize real-time wireless transmission via a standard low-power Bluetooth module, successfully capturing dynamic pressure maps in porcine knee models (Fig. 6c). Beyond material innovations, knee joint stress sensors have also entered the era of wireless transmission. Luo et al.¹⁰⁰ proposed an innovative signal transmission mechanism based on Maxwell displacement current, enabling implanted TENG sensors to wirelessly transmit data without relying on integrated electronics or external power sources (Fig. 6d). This mechanism exhibits lower signal attenuation in lossy media—such as saline, blood, or animal tissue—than in air or even vacuum. Crucially, this “electronic-free” architecture fundamentally circumvents conventional bottlenecks in bandwidth and on-board signal processing by shifting the computational burden entirely to the external receiver.

Fig. 6.

a) Flexible iontronic pressure sensor.⁹¹ Copyright 2024, Oxford University Press. b) Flexible piezoresistive pressure sensors based on the strain effect.⁸⁹ Copyright 2024, Wiley. c) Flexible electret-based pressure sensors and their dynamic pressure monitoring during porcine knee arthroplasty.¹⁰⁶ Copyright 2024, Wiley. d) Developing self-powered, electronic-free metamaterial implants for wireless force sensing.¹⁰⁰ Copyright 2025, ELSEVIER.

However, these novel materials and sensors have been developed primarily in the context of TKA, and to date, there has been no research focusing on intra-articular stress sensors for HTO. Unlike total knee arthroplasty (TKA), which creates ample space for prosthetic components and focuses on implant-bone interface forces, high tibial osteotomy (HTO) redirects load from the damaged medial compartment to the relatively healthy lateral compartment by adjusting the tibial mechanical axis.¹¹⁹ Consequently, monitoring requirements differ fundamentally: the success of the procedure depends heavily on the precise postoperative distribution of knee joint loads on the native cartilage surfaces within a highly constrained joint space, making real-time monitoring of intra-articular stress critical for evaluating surgical outcomes and preventing complications.¹²⁰

Currently, there are three fundamental reasons why the new flexible sensors are not suitable for HTO (Fig. 7). The first is biocompatibility. The core monitoring requirement for HTO is the physiological distribution of intra-articular stress, which is fundamentally different from load monitoring at the prosthesis–bone interface in TKA. To prevent acute inflammatory responses, synovitis, or cytotoxicity, the electronics must be hermetically sealed in biocompatible encapsulation materials. Solutions include united states pharmacopeia (USP) Class VI medical-grade silicones,¹²¹ Parylene C conformal coatings,¹²² or fluorinated elastomers.¹²³ The encapsulation layer is both biocompatible and chemically inert, ensuring that the sensor remains stable during surgery and prevents fluid ingress, which could otherwise lead to measurement drift.

Fig. 7.

Schematic diagram of flexible intra-articular stress sensor for HTO.

Second challenge is spatial constraints and shape compatibility. The anatomical space of the HTO joint cavity is extremely limited, with a joint gap of only approximately 3 mm,¹²⁴ whereas existing sensors are significantly larger than this allowance. The PVDF sensor designed by Chahari et al. has a thickness of 19 mm, which is entirely incompatible with the physiological joint cavity. While the 3 mm joint gap in HTO presents a structural challenge, it is by no means an insurmountable barrier when compared to the dimensional constraints successfully navigated in other implantable domains, such as intracranial and intravascular monitoring. Starr et al. developed a capacitive pressure transducer for cardiovascular catheters with a functional profile of only 15 μm.¹²⁵ By utilizing surface-micromachining to transfer a polyimide diaphragm onto a flexible substrate, they achieved a device so thin that it does not disrupt hemodynamic flow. Similarly, in the realm of biofluid pressure sensing, Tang et al. introduced an ultraminiature sensor based on interior corner flow principles with a total thickness of merely 8 μm.¹²⁶ This device leverages capillary action within micro-channels rather than bulky electronic components, demonstrating that physical sensing structures can be reduced to the cellular scale.

Finally, there is the issue of mechanical incompatibility. Unlike TKA sensors that contact metal and bone, HTO sensors interface directly with native cartilage. A rigid sensor would create stress concentration, distorting measurement data and potentially injuring the delicate chondral surface. Therefore, the next generation of sensors must employ soft, stretchable substrates (such as thermoplastic polyurethane (TPU)¹²⁷ or hydrogels¹²⁸) with an elastic modulus (∼1–10 MPa) comparable to that of articular cartilage.¹²⁹ This “mechanical impedance matching” ensures conformal contact and accurate transmission of contact forces without artifacts.

Synthesizing these requirements, the ideal HTO sensor is a wireless, paper-thin (<0.5 mm), and mechanically compliant patch. Functionally, it should be capable of being rolled for arthroscopic delivery, deploying autonomously across the medial compartment, and providing a high-resolution, real-time “pressure map.” This evolution will transform HTO from a geometrically approximated procedure into a force-guided precision surgery.

6. Outlook

6.1. Critical barriers to clinical translation

The transition from laboratory prototypes to clinical reality faces distinct engineering hurdles tailored to the specific application scenarios. For non-surgical wearable devices, the primary challenge lies in maintaining longitudinal signal fidelity during daily rehabilitation, where sensors must resist mechanical fatigue and signal drift caused by sweat and motion artifacts. In the context of intraoperative sensing, while long-term biocompatibility is less critical given the short surgical window (<24 h), the acute physiological environment poses immediate risks to signal integrity. To address this, Fu et al. ¹³⁰ emphasized that next-generation intelligent sensors must evolve beyond simple transducers to incorporate “self-compensation” and “self-calibration” capabilities. Instead of relying solely on passive encapsulation, they propose integrating microprocessors to execute real-time algorithms—such as zero-drift correction and non-linear compensation—thereby actively neutralizing environmental interference to ensure measurement accuracy.

Beyond engineering optimization, the translation of sensor-based technologies into standard orthopedic practice faces significant regulatory and economic barriers that have been largely overlooked. For wearables, the challenge lies in distinguishing “general wellness devices” from regulated “medical devices.” To be used for clinical decision-making, rehabilitation sensors must demonstrate substantial equivalence to gold-standard gait labs. Moreover, as these devices continuously transmit data outside hospitals, data privacy will also become a key focus of regulation. For intraoperative sensors, maintaining sterility and biocompatibility (ISO 10993) is the baseline. However, the inclusion of wireless data transmission introduces new scrutiny regarding cybersecurity and electromagnetic compatibility (EMC). According to recent food and drug administration (FDA) guidance, manufacturers must strictly validate cybersecurity to prevent unauthorized access to the embedded device, as mandated by the latest FDA Guidance (2025).¹³¹

The adoption of sensor technology also hinges on its economic viability. For wearable sensors, the value proposition lies in remote therapeutic monitoring. By enabling home-based quantitative assessment, these systems can reduce the burden of frequent in-clinic physical therapy visits. Future economic models must prove that sensor-guided home rehab reduces total episode-of-care costs by minimizing hospital readmissions and improving patient adherence. For TKA/HTO procedures, the use of stress sensors is justified only when their cost is significantly lower than the expenses associated with the costly revision surgeries they can help avoid. Even if only a small fraction of mechanical failures can be prevented, this technology remains cost-effectiveness.

Finally, there remains a significant gap between laboratory research and clinical validation. To date, the vast majority of sensor-related studies are still confined to in vitro performance testing or small-scale feasibility trials (sample size <50),¹¹⁵ which lack the statistical power to predict long-term clinical outcomes. To achieve widespread adoption and application of this technology, the field must accelerate the advancement of large-scale randomized controlled trials.

6.2. Roadmap toward integrated intelligent systems

Building upon the comprehensive review of sensor technologies in knee osteoarthritis (OA) management, a clear technological trajectory has emerged. Current systems have successfully established the foundational pillars for continuous data acquisition in both ambulatory and surgical settings. Wearable devices enable the monitoring of rehabilitation, while intelligent implants provide intraoperative feedback. However, these elements often operate in isolation. The future of OA care lies not in further incremental improvements within these silos, but in their strategic integration into a coherent, patient-centric feedback framework. This integration necessitates coordinated advancements across four interdependent domains, evolving hierarchically from signal acquisition to clinical application, as visualized in Fig. 8.

• (1)Foundation—Sensor Performance & Physiological Compatibility: The basis of this ecosystem depends on the quality of the signal source. Current sensors still face limitations in mechanical compliance, long-term stability, and power autonomy. Future developments should focus on ultra-thin, flexible sensors with tissue-like modulus, as well as self-powered mechanisms such as piezoelectric or triboelectric nanogenerators, to enable seamless integration with biological tissues.

• (2)Transmission—Wireless, Integrated Data Platforms: Once reliable data acquisition is ensured, it must be effectively transmitted. Future systems must evolve from single-channel or wired setups toward multi-modal, synchronized, and wireless platforms capable of real-time data fusion, visualization, and closed-loop feedback between patients and clinicians.
• (3)Intelligence—AI-Driven Adaptive Analytics: With the influx of multidimensional data from the wireless platforms, advanced processing capabilities become essential. While AI algorithms show promise in motion classification and torque estimation, they have not yet been fully integrated into dynamic clinical decision-making. Future frameworks should incorporate adaptive and reinforcement learning models to calibrate sensor parameters and intervention strategies in real time, according to individual patient progression.
• (4)Implementation—Large-Scale Clinical Deployment: Finally, the validated intelligent system must be integrated into the healthcare network. Widespread adoption will require validation through large-scale randomized controlled trials and the development of viable reimbursement models. A holistic “hospital–community–home” network should be established to enable equitable resource distribution and continuous patient monitoring across care settings.

Fig. 8.

Future directions for sensor-enabled integrated OA management.

In summary, the integration of flexible electronics, multimodal sensing, AI analytics, and scalable telehealth infrastructure is poised to unify preoperative assessment, intraoperative guidance, and postoperative rehabilitation into a continuous, patient-specific care cycle. Overcoming these interdisciplinary challenges will pave the way for a closed-loop OA management system that sustains joint homeostasis and improves long-term functional outcomes.

7. Conclusion

This review has systematically delineated the pivotal role of portable electronic devices across the entire spectrum of knee osteoarthritis (OA) management. The evidence presented underscores a fundamental clinical paradigm shift: moving from subjective, episodic observation toward objective, continuous, and mechanobiology-driven precision medicine.

Current sensing technologies have successfully addressed the distinct biomechanical needs of the two main therapeutic stages. In non-surgical conservative care, wearable sensors have transformed rehabilitation from a “black box” into a quantifiable process, enabling clinicians to monitor neuromuscular control and gait compliance in real-world settings. In surgical intervention, instrumented implants and intraoperative sensors have evolved to provide the precise load quantification required to restore physiological joint kinematics, particularly in complex procedures like TKA and HTO.

Ultimately, the value of these technologies lies not just in data collection, but in their ability to operationalize the principles of joint mechanobiology. By converting invisible mechanical stresses into actionable clinical metrics, these devices empower physicians to preserve joint homeostasis more effectively. As these sensing modalities mature and overcome technical hurdles, they could become the cornerstone of a new standard of care, advancing the goal of lifelong joint preservation.

CRediT authorship contribution statement

Ran Xu: Writing – review & editing, Writing – original draft. Xu Jiang: Writing – original draft. Yushun Tao: Writing – original draft. Shikun Fang: Writing – original draft. Jie Li: Writing – original draft. Fan Zhao: Writing – review & editing, Funding acquisition, Conceptualization. Fujun Wang: Funding acquisition. Liao Wang: Funding acquisition, Conceptualization. Jun Zhang: Conceptualization.

Ethical approval

This article is a review article and does not contain any studies with human participants or animals performed by any of the authors. Therefore, ethical approval and informed consent are not required.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.