Reliability of quadriceps and hamstring muscle stiffness measured by ultrasound shear wave elastography in patients with knee osteoarthritis Authors: Ping Liu, Mohamad Shariff Bin Abdul Hamid, Yueyue Shi, Xiaohan Fan, Hengzhi Deng, Eliza Hafiz, Xuelin Zhang, Yaru Hao Published: 2026 ABSTRACT ObjectiveThis study aimed to assess the intra- and inter-examiner reliability of shear wave elastography (SWE) in measuring the stiffness of quadriceps and hamstring muscles in individuals diagnosed with knee osteoarthritis (KOA).MethodsThirty-two individuals diagnosed with KOA underwent evaluations of the quadriceps and hamstring muscles. Shear-wave elastography (SWE) quantified shear-wave speed (m/s) in the rectus femoris, vastus lateralis, vastus medialis, quadriceps tendon, semimembranosus, hamstring tendon, and biceps femoris. Two independent examiners conducted SWE assessments to determine muscle stiffness, with measurements repeated after 6 to 8 days.ResultsIn the initial test session, mean shear-wave speeds ranged from 2.00 to 2.47 m/s for the quadriceps muscle and from 2.06 to 2.60 m/s for the hamstrings across testers 1 and 2. During the retest session, mean shear-wave speeds ranged from 1.98 to 2.48 m/s for the quadriceps and from 2.07 to 2.57 m/s for the hamstrings. The intraclass correlation coefficients (ICCs) indicated good to excellent reliability for both intra-examiner reliability (0.88–0.98) and inter-examiner reliability (0.78–0.93). Additionally, the ICCs and standard errors of measurement (SEM) for inter-examiner reliability varied from 0.78 to 0.93 and 0.04 to 0.14 for hamstring muscle stiffness, and from 0.84 to 0.91 and 0.08 to 0.12 for quadriceps muscle stiffness, respectively.ConclusionShear-wave elastography is a reliable technique for assessing muscle stiffness in the quadriceps and hamstrings of patients suffering from knee osteoarthritis. Introduction Knee osteoarthritis (KOA) is a leading cause of chronic pain and imposes a significant economic burden on healthcare systems and society [1]. The primary symptoms of KOA include tenderness to palpation, limited range of motion, persistent pain, muscle atrophy, and increased muscle stiffness [1–4]. Notably, muscle stiffness has been consistently reported to be associated with the pathophysiology and clinical presentation of KOA, thereby complicating accurate detection [4–6]. Alterations in muscle stiffness have been frequently observed as clinical features associated with KOA [7, 8]. Indeed, KOA patients generally exhibit higher quadriceps and hamstring stiffness than healthy individuals [4, 8, 9], and this increased stiffness has been reported to be associated with a higher incidence of KOA in older adults over a 24-month period [10]. Assessing thigh muscle stiffness is important [9, 11], as such stiffness may influence knee impact absorption and stability, both of which are crucial for reducing joint loading [12–14]. Nevertheless, current methodological studies report limited precision and consistency in assessing muscle stiffness in KOA [15, 16]. This is partly due to factors such as chronic stiffness, muscle inhibition, effusion, and compensatory movement patterns, which make measurements highly sensitive to testing conditions [6, 11, 17]. Muscle stiffness evaluation has evolved through various modalities, including myotonometry, magnetic resonance elastography (MRE), and ultrasound-based elastography, such as shear-wave elastography (SWE) and strain elastography [18–21]. However, the absence of reliable, objective, and reproducible methods for measuring muscle stiffness in KOA has limited rigorous evaluation of treatment effectiveness [22]. A standardized approach to stiffness measurement is essential for improving the monitoring of treatment response. SWE is a real-time, non-invasive, and low-risk imaging technique commonly used in musculoskeletal evaluations to assess soft tissues, such as muscles, nerves, tendons, and fascia [23]. It quantifies soft tissue stiffness by measuring the propagation speed of shear waves induced within the tissue [24]. Despite not being widely adopted in routine clinical practice, SWE shows promise in distinguishing musculoskeletal disorders, such as chronic low back pain, neck muscle stiffness, and KOA [10, 25, 26]. Previous studies have shown that SWE measurements are reliable across various muscles [27, 28]. Moreover, SWE quantifies muscle stiffness and elasticity, which have been widely considered in many clinical trials [29, 30]. This technique enables the quantification of structural changes, particularly during KOA postoperative rehabilitation [31, 32]. However, the reliability of SWE in evaluating quadriceps and hamstring muscle stiffness among KOA patients has yet to be documented [21]. This study, therefore, sought to evaluate the consistency of SWE measurements for quadriceps and hamstring muscle stiffness in KOA patients, focusing on both intra- and inter-examiner reliability. Methods Study design A prospective observational study was conducted to evaluate the reliability of stiffness measurements in both the quadriceps and hamstring muscles. Each session consisted of three measurement trials, followed by a follow-up retest conducted 6 to 7 days later. Approval of the study was granted by the Ethics and Research Committee of Qufu Normal University (No. 2024-061), and procedures adhered to the Declaration of Helsinki. Written consent was obtained from all subjects prior to evaluation. The protocol complied with the Standard Protocol Items guidelines. Data collection took place over 15 days, with five participants assessed per day under controlled conditions. To minimize environmental variation, the room temperature was maintained at 24 °C. The examiner team consisted of the principal investigator (a specialist in therapy and rehabilitation) and two experienced sonographers from a rehabilitation hospital, each with 3 to 8 years of professional experience. To ensure measurement consistency, all examiners completed a seven-day intensive training workshop on ultrasound elastography led by the principal investigator. Participants Sample size estimation The sample size was estimated using the STATA software (Stata Corp LP, College Station, Texas), applying the “sampsi_rho” command with a one-sided α of 0.05, a statistical power of 0.90, a null coefficient of 0.00, and an alternative coefficient of 0.50. This yielded an estimated sample size of 32 [33, 34]. As a result, 32 individuals with KOA participated in the study, with assessments performed on 32 quadriceps and 32 hamstring muscles. Older KOA patients: inclusion criteria and exclusion criteria Participants were included if they had been diagnosed with KOA within the past six months, based on the American College of Rheumatology (ACR) criteria [35] and if radiographic imaging confirmed mild, moderate, or severe KOA according to the Kellgren and Lawrence scale [36]. Eligible individuals were between 45 and 75 years old, able to walk independently for at least 10 m, and had stable health conditions. They were also required to have normal vision and language comprehension, no severe organ dysfunction, and a KOA symptom duration of more than three months. Participants were excluded if they did not meet KOA diagnostic criteria, had neurological conditions, or suffered from severe cardiovascular diseases. Additional exclusion criteria included significant lumbar disc herniation, uncontrolled hypertension or diabetes, recent smoking or alcohol use, or medical contraindications to physical activity. Those taking beta-blockers with visual field deficits were also excluded. Outcome measures Demographic information Participants’ demographic data was recorded, including their age (years), gender, educational status, marital status, socioeconomic status, and employment. A research assistant conducted anthropometric measurements, including height (m) and weight (kg), using a mechanical weighing scale and stadiometer (SECA, Germany). Participants stood barefoot and wore light clothing. All anthropometric assessments, including weight and height, were conducted following the guidelines of the International Society for the Advancement of Kinanthropometry (ISAK) [37]. Visual analogue scale (VAS) Pain levels were assessed using the Visual Analogue Scale (VAS) to rate average pain over the past week, spanning from 0 (no discomfort) to 10 (extreme pain) [38]. Among individuals with chronic musculoskeletal issues, test-retest outcomes showed moderate to strong reliability [39]. The vertical and horizontal VAS formats have shown a correlation as high as 0.99 [40]. Muscle stiffness Muscle stiffness measurements of the quadriceps and hamstrings were obtained using an ACUSON Oxana ultrasound system (Siemens Medical Solutions USA, Mountain View, CA) with a 9L4 linear probe (7–9 MHz). The equipment was set to SWE using a customized musculoskeletal setting with penetration mode, smoothing level 5, and persistence turned off. To ensure optimal visualization of muscle architecture, the probe was aligned parallel to the muscle fibres. The muscle was stratified into upper, middle, and lower layers, with a fixed-size square region (SWE Box) positioned in the middle layer to delineate the elastography imaging field [41, 42]. After stabilizing the SWE map for approximately 5 s, the image was frozen, and three square regions of interest (ROIs; 5 × 5 mm) were placed within the muscle belly. Each test was performed three times, with the mean value calculated and documented. Surface landmarks guided measurements: The elastic modulus of the quadriceps tendon (QT) was measured at the superior edge of the patella [43]. The elastic modulus of the rectus femoris, vastus medialis, and vastus lateralis muscles was assessed at the midpoint along the line connecting the anterior superior iliac spine and the upper edge of the patella [30, 44, 45]. Hamstring measurements were taken at the midpoint between the ischial tubercle and popliteal fossa, including the biceps femoris (BF) and semimembranosus (SM) [46]. Hamstring tendon (HT) stiffness was assessed at the proximal hamstring tendon origin near the ischial tuberosity [47, 48]. Muscle stiffness was determined by using shear wave mode, capturing velocity values (m/s) per muscle and averaging for the final shear wave velocity (Figs. 1 and 2). Fig. 1Typical elastography maps of the quadriceps showing the vastus lateralis (left) and vastus medialis (right) Typical elastography maps of the quadriceps showing the vastus lateralis (left) and vastus medialis (right) Fig. 2Typical elastography maps of the hamstrings showing the biceps femoris (left) and semimembranosus (right) Typical elastography maps of the hamstrings showing the biceps femoris (left) and semimembranosus (right) Participants were scanned in a relaxed supine position for quadriceps assessment and in a prone position for hamstring evaluation (Fig. 3A–B). In the supine position, knee flexion was maintained at approximately 30 degrees using a standardized pillow height(Fig. 3C). Before measurement, participants rested for 5 min with the limb supported. They were asked to keep their legs relaxed and breathe normally. If the examiner saw obvious contraction or guarding, the measurement was stopped and repeated after a short rest. The same positioning and instructions were used at both sessions (6–8 days apart). The examiner positioned the ultrasound probe on the anterior and posterior aspects of the thigh, perpendicular to its long axis, as established in a prior study [49]. Participants were instructed to wear shorts to ensure full exposure of the thigh region and to remove their shoes to avoid interference during the examination. During the assessment, only the examiner and the participant were present in the room, and participants were instructed to remain silent to minimize potential movement and distraction. To minimize image distortion, an excessive amount of ultrasound gel was applied to the skin. Image acquisition was performed on the affected limb or on the dominant side in cases of bilateral knee involvement. Fig. 3Surface landmarks and joint positioning for SWE measurements. A Anatomical landmarks for quadriceps muscle–tendon assessment. B Anatomical landmarks for hamstring muscle–tendon assessment. C Standardized knee-joint positioning for SWE acquisition; a towel roll supported the lower leg to maintain uniform alignment across participants Surface landmarks and joint positioning for SWE measurements. A Anatomical landmarks for quadriceps muscle–tendon assessment. B Anatomical landmarks for hamstring muscle–tendon assessment. C Standardized knee-joint positioning for SWE acquisition; a towel roll supported the lower leg to maintain uniform alignment across participants All measurements were completed within two hours to minimize physiological variability in muscle stiffness. To evaluate intra- and inter-rater reliability, two examiners independently performed three repeated SWE measurements per muscle during each session, with a second session held 6–8 days later. The measurements were taken after the exhalation phase of respiration for both the subjects and the assessors [49]. After each session, the participant was repositioned to the initial state, and the skin was thoroughly cleaned to remove any remaining gel or markings. This procedure ensured that each image and dataset were obtained independently, thereby minimizing the risk of measurement bias, such as anchoring effects [50]. Statistical analysis STATA software (StataCorp LP, College Station, Texas) was utilized for data analysis. A paired t-test was applied to assess systematic bias between test and retest scores. Intra- and inter-examiner reliability within and between sessions was evaluated by averaging three muscle stiffness measurements from two examiners per session [51]. A two-way random effects model was employed to calculate intraclass correlation coefficients (lCCs) to evaluate the reliability of muscle stiffness measurements both within and across sessions. The measured variables exhibited normal distribution, as confirmed by the Shapiro-Wilk test. Paired t-tests and ICCs were deemed suitable for these variables [52]. Absolute reliability was assessed using the standard error of measurement (SEM), which is derived from the square root of the error variance, to estimate measurement error. Based on established thresholds, ICCs of 0.75 or higher indicated good reliability, whereas ICCs of 0.90 or higher signified excellent reliability [53]. Additionally, reliability was further evaluated using Bland–Altman plots. To evaluate whether there were meaningful changes between the two measurement occasions, the minimal detectable change at the 95% confidence level (MDC95) was calculated as MDC95 = 1.96 × SEM × √2. Additionally, the coefficient of variation (CV) was employed to compare the absolute reliability characteristics of the different assessments. The formula used was [(standard deviation/mean) x 100%]. All analyses deemed outcomes statistically significant when p-values were below 0.05. Results Characteristics of the overall sample (N = 32) The study involved 32 participants (18 women and 14 men) with a mean age of 52.7 ± 4.7 years, a mean body weight of 62.6 ± 8.5 kg, and a mean height of 169 ± 7.5 cm, assessing 32 quadriceps and 32 hamstring muscles in total (Table 1). Table 1Overall sample’s socio-demographic characteristics (N = 32)VariablesKnee OA (n = 32)FrequencyPercentageGender Woman1856.3% Man1443.7%Affected side Left825.0% Right825.0% Bilateral1650.0%KOA Severity Mild1031.3% Moderate1443.7% Severe825.0% Table 1Overall sample’s socio-demographic characteristics (N = 32)VariablesKnee OA (n = 32)FrequencyPercentageGender Woman1856.3% Man1443.7%Affected side Left825.0% Right825.0% Bilateral1650.0%KOA Severity Mild1031.3% Moderate1443.7% Severe825.0% Overall sample’s socio-demographic characteristics (N = 32) Compare the differences in quadriceps and hamstring muscle stiffness between the test and retest conditions As assessed by two or three evaluators, the mean stiffness of the quadriceps and hamstring muscles exhibited no significant statistical variation between the test and retest assessments, suggesting no systematic measurement bias (Table 2). Table 2Descriptive data for test-retest measures of quadriceps and hamstring muscle stiffness in 64 musclesSonographic MeasuresFirst ExaminerSecond ExaminerTestRetest P TestRetest P MeanSDMeanSDMeanSDMeanSDSM stiffness (m/s)2.560.212.550.200.202.580.342.570.350.22HT stiffness (m/s)2.060.142.070.150.292.070.132.080.110.09BF stiffness (m/s)2.600.252.590.220.562.560.102.570.080.15RF stiffness (m/s)2.470.202.450.220.112.460.212.480.240.05VL stiffness (m/s)2.000.341.980.370.342.060.252.050.470.88VM stiffness (m/s)2.400.272.410.280.332.410.242.380.190.13QT stiffness (m/s)2.410.312.420.330.252.420.322.440.290.22SD Standard deviation, SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon. Values are mean ± SD Table 2Descriptive data for test-retest measures of quadriceps and hamstring muscle stiffness in 64 musclesSonographic MeasuresFirst ExaminerSecond ExaminerTestRetest P TestRetest P MeanSDMeanSDMeanSDMeanSDSM stiffness (m/s)2.560.212.550.200.202.580.342.570.350.22HT stiffness (m/s)2.060.142.070.150.292.070.132.080.110.09BF stiffness (m/s)2.600.252.590.220.562.560.102.570.080.15RF stiffness (m/s)2.470.202.450.220.112.460.212.480.240.05VL stiffness (m/s)2.000.341.980.370.342.060.252.050.470.88VM stiffness (m/s)2.400.272.410.280.332.410.242.380.190.13QT stiffness (m/s)2.410.312.420.330.252.420.322.440.290.22SD Standard deviation, SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon. Values are mean ± SD Descriptive data for test-retest measures of quadriceps and hamstring muscle stiffness in 64 muscles SD Standard deviation, SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon. Values are mean ± SD Reliability of quadriceps and hamstring muscle stiffness across and within sessions Table 3 outlines the intra-examiner reliability across within- and between-session assessments of quadriceps and hamstring muscle stiffness. The ICC values indicated a range from good to excellent reliability for intra-examiner measurements of hamstring muscle stiffness across within- and between-session evaluations. The ICC for the first examiner ranged from 0.83 to 0.94, while for the second examiner, it varied from 0.81 to 0.98. Table 3Within and between session reliability of the quadriceps and hamstring muscle stiffness in 64 muscles with KOAExaminerMeasuresWithin-Session IntraexaminerBetween-Session IntraexaminerTestRetestICCSEMMDCCVFirst examinerSM0.880.930.940.050.147.98HT0.830.900.910.040.127.03BF0.880.940.940.060.169.07RF0.790.860.880.070.208.55VL0.900.920.940.090.2417.86VM0.860.910.930.070.2011.44QT0.970.980.980.050.1313.26Second examinerSM0.810.900.900.110.3013.40HT0.970.960.980.020.055.80BF0.860.890.920.030.073.53RF0.950.960.970.040.119.13VL0.830.900.900.120.3318.32VM0.880.920.930.060.169.04QT0.850.890.920.090.2412.57ICC Intraclass correlation coefficients, SEM Standard error of measurement, MDC Minimal detectable change, CV Coefficient of variation, SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon. Values are mean ± SD Table 3Within and between session reliability of the quadriceps and hamstring muscle stiffness in 64 muscles with KOAExaminerMeasuresWithin-Session IntraexaminerBetween-Session IntraexaminerTestRetestICCSEMMDCCVFirst examinerSM0.880.930.940.050.147.98HT0.830.900.910.040.127.03BF0.880.940.940.060.169.07RF0.790.860.880.070.208.55VL0.900.920.940.090.2417.86VM0.860.910.930.070.2011.44QT0.970.980.980.050.1313.26Second examinerSM0.810.900.900.110.3013.40HT0.970.960.980.020.055.80BF0.860.890.920.030.073.53RF0.950.960.970.040.119.13VL0.830.900.900.120.3318.32VM0.880.920.930.060.169.04QT0.850.890.920.090.2412.57ICC Intraclass correlation coefficients, SEM Standard error of measurement, MDC Minimal detectable change, CV Coefficient of variation, SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon. Values are mean ± SD Within and between session reliability of the quadriceps and hamstring muscle stiffness in 64 muscles with KOA ICC Intraclass correlation coefficients, SEM Standard error of measurement, MDC Minimal detectable change, CV Coefficient of variation, SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon. Values are mean ± SD Additionally, the ICC values for intra-examiner quadriceps muscle stiffness were good within and between sessions. Values for the first examiner varied between 0.79 and 0.98, while those for the second ranged from 0.83 to 0.97. These findings demonstrate high consistency in measurements, confirming the reliability of the methods used to evaluate muscle stiffness. Reliability of quadriceps and hamstring muscle stiffness among examiners Table 4 presents the values of ICC, SEM, MDC, and CV concerning inter-examiner reliability. The ICC and SEM values for the inter-examiner reliability of hamstring muscle stiffness ranged from 0.78 to 0.93 and from 0.04 to 0.14., respectively (MDC: 0.10 to 0.38; CV: 6.56% to 11.25%). Similarly, the ICC and SEM values for the inter-examiner reliability of quadriceps muscle stiffness varied between 0.84 and 0.91 and 0.08 to 0.12, respectively (MDC: 0.21 to 0.34; CV: 8.72% to 17.49%). These results indicate a high level of measurement consistency, reinforcing the reliability of the applied methodology. Table 4Inter examiner reliability of the quadriceps and hamstring muscle stiffness in 64 muscles with KOAReliabilityMeasuresICCSEMMDCCVInter-examiner (test session)SM0.830.110.3211.00HT0.950.030.086.54BF0.680.110.307.38RF0.950.050.138.32VL0.960.060.1714.70VM0.880.090.2510.62QT0.890.100.2913.05Inter-examiner (retest session)SM0.910.090.2411.13HT0.880.030.086.34BF0.710.090.296.42RF0.920.070.189.34VL0.940.100.2920.99VM0.920.070.199.99QT0.970.050.1512.78Inter-examiner (test and retest sessions)SM0.780.140.3811.25HT0.930.040.106.56BF0.860.070.186.81RF0.870.080.218.72VL0.880.120.3417.49VM0.840.090.259.49QT0.910.100.2812.99ICC Intraclass correlation coefficients, SEM Standard error of measurement, MDC Minimal detectable change, CV Coefficient of variation, SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon. Values are mean ± SD Table 4Inter examiner reliability of the quadriceps and hamstring muscle stiffness in 64 muscles with KOAReliabilityMeasuresICCSEMMDCCVInter-examiner (test session)SM0.830.110.3211.00HT0.950.030.086.54BF0.680.110.307.38RF0.950.050.138.32VL0.960.060.1714.70VM0.880.090.2510.62QT0.890.100.2913.05Inter-examiner (retest session)SM0.910.090.2411.13HT0.880.030.086.34BF0.710.090.296.42RF0.920.070.189.34VL0.940.100.2920.99VM0.920.070.199.99QT0.970.050.1512.78Inter-examiner (test and retest sessions)SM0.780.140.3811.25HT0.930.040.106.56BF0.860.070.186.81RF0.870.080.218.72VL0.880.120.3417.49VM0.840.090.259.49QT0.910.100.2812.99ICC Intraclass correlation coefficients, SEM Standard error of measurement, MDC Minimal detectable change, CV Coefficient of variation, SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon. Values are mean ± SD Inter examiner reliability of the quadriceps and hamstring muscle stiffness in 64 muscles with KOA ICC Intraclass correlation coefficients, SEM Standard error of measurement, MDC Minimal detectable change, CV Coefficient of variation, SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon. Values are mean ± SD The Bland–Altman plots illustrating intra-rater reliability values for the SM, HT, BF, RF, VL, VM, and QT are shown in Fig. 4A–G. The mean differences were 0.012 m/s for SM, -0.007 m/s for HT, -0.003 m/s for BF, 0.001 m/s for RF, 0.012 m/s for VL, 0.011 m/s for VM, and -0.015 m/s for QT. The corresponding 95% limits of agreement were -0.246 to 0.270 m/s (SM), -0.088 to 0.073 m/s (HT), -0.133 to 0.126 m/s (BF), -0.158 to 0.159 m/s (RF), -0.259 to 0.282 m/s (VL), -0.135 to 0.156 m/s (VM), and -0.151 to 0.121 m/s (QT). Fig. 4Bland–Altman plots illustrating intra-rater reliabilities for the shear-wave speed of the quadriceps muscles and hamstring muscles. The difference in shear-wave speed between the test and retest sessions is plotted against the mean shear-wave speed of each participant for the quadriceps muscles (D, E, F, and G) and hamstring muscles (A, B, C). In the images, the dotted line represents the mean difference, and the solid lines above and below indicate the limits of agreement Bland–Altman plots illustrating intra-rater reliabilities for the shear-wave speed of the quadriceps muscles and hamstring muscles. The difference in shear-wave speed between the test and retest sessions is plotted against the mean shear-wave speed of each participant for the quadriceps muscles (D, E, F, and G) and hamstring muscles (A, B, C). In the images, the dotted line represents the mean difference, and the solid lines above and below indicate the limits of agreement The inter-rater reliability values are shown in Fig. 5A–G. The mean differences were -0.019 m/s for SM, -0.008 m/s for HT, 0.033 m/s for BF, -0.011 m/s for RF, -0.060 m/s for VL, 0.009 m/s for VM, and -0.011 m/s for QT. The corresponding 95% limits of agreement were -0.340 to 0.302 m/s (SM), -0.114 to 0.098 m/s (HT), -0.308 to 0.374 m/s (BF), -0.183 to 0.162 m/s (RF), -0.201 to 0.080 m/s (VL), -0.209 to 0.227 m/s (VM), and -0.222 to 0.200 m/s (QT). Fig. 5Bland–Altman plots illustrating inter-rater reliability for the shear-wave speed of the quadriceps muscles and hamstring muscles. The difference in shear-wave speed between the two testers is plotted against the mean shear-wave speed of each participant for the quadriceps muscles (D, E, F, and G) and hamstring muscles (A, B, C). In the images, the dotted line represents the mean difference, and the solid lines above and below indicate the limits of agreement Bland–Altman plots illustrating inter-rater reliability for the shear-wave speed of the quadriceps muscles and hamstring muscles. The difference in shear-wave speed between the two testers is plotted against the mean shear-wave speed of each participant for the quadriceps muscles (D, E, F, and G) and hamstring muscles (A, B, C). In the images, the dotted line represents the mean difference, and the solid lines above and below indicate the limits of agreement Reliability of quadriceps and hamstring muscle stiffness across KOA stages Table 5 presents intra-examiner reliability across the different stages of KOA. Mild KOA showed the highest reliability (Examiner 1: 0.81–0.98; Examiner 2: 0.84–0.98). In moderate KOA, the ICCs were lower and more dispersed (Examiner 1: 0.58–0.98; Examiner 2: 0.56–0.98). Severe KOA exhibited even greater variability, especially for SM and HT (Examiner 1: 0.65–0.98; Examiner 2: 0.52–0.97). Table 5Intra-examiner reliability of SWE measurements across different stages of KOAKOA SeverityExaminerSM ICCHT ICCBF ICCRF ICCVL ICCVM ICCQT ICCMild KOAExaminer 10.890.870.960.810.930.960.98Examiner 20.900.970.840.980.850.960.92Moderate KOAExaminer 10.800.980.820.960.910.580.97Examiner 20.880.980.870.960.880.860.56Severe KOAExaminer 10.980.650.690.770.800.890.95Examiner 20.520.960.960.940.740.810.97KOA severity was classified as mild, moderate, and severe according to the Kellgren–Lawrence (KL) grading system. SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon Table 5Intra-examiner reliability of SWE measurements across different stages of KOAKOA SeverityExaminerSM ICCHT ICCBF ICCRF ICCVL ICCVM ICCQT ICCMild KOAExaminer 10.890.870.960.810.930.960.98Examiner 20.900.970.840.980.850.960.92Moderate KOAExaminer 10.800.980.820.960.910.580.97Examiner 20.880.980.870.960.880.860.56Severe KOAExaminer 10.980.650.690.770.800.890.95Examiner 20.520.960.960.940.740.810.97KOA severity was classified as mild, moderate, and severe according to the Kellgren–Lawrence (KL) grading system. SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon Intra-examiner reliability of SWE measurements across different stages of KOA KOA severity was classified as mild, moderate, and severe according to the Kellgren–Lawrence (KL) grading system. SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon Table 6 summarizes inter-examiner reliability for different stages of KOA. Mild KOA showed good to excellent reliability, with ICC values ranging from 0.82 to 0.97. Moderate KOA had slightly lower and more variable ICCs, ranging from 0.75 to 0.98. Table 6Inter-examiner reliability of SWE measurements across different stages of KOAKOA SeveritySM ICCHT ICCBF ICCRF ICCVL ICCVM ICCQT ICCMild KOA0.920.940.820.960.950.950.97Moderate KOA0.900.980.750.970.950.850.87Severe KOA0.740.880.570.880.910.920.97KOA severity was classified as mild, moderate, and severe according to the Kellgren–Lawrence (KL) grading system. SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon Table 6Inter-examiner reliability of SWE measurements across different stages of KOAKOA SeveritySM ICCHT ICCBF ICCRF ICCVL ICCVM ICCQT ICCMild KOA0.920.940.820.960.950.950.97Moderate KOA0.900.980.750.970.950.850.87Severe KOA0.740.880.570.880.910.920.97KOA severity was classified as mild, moderate, and severe according to the Kellgren–Lawrence (KL) grading system. SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon Inter-examiner reliability of SWE measurements across different stages of KOA KOA severity was classified as mild, moderate, and severe according to the Kellgren–Lawrence (KL) grading system. SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon In contrast, severe KOA showed the greatest variability, as evidenced by the lowest ICC for BF at 0.57, with a broader range of 0.57 to 0.97. Associations between pain intensity and quadriceps and hamstring muscle stiffness Table 7 displays Spearman correlation results showing the relationship between pain intensity (VAS) and muscle stiffness. Pain intensity had positive correlations with SM stiffness (ρ = 0.46, p = 0.008), BF stiffness (ρ = 0.41, p = 0.019), VL stiffness (ρ = 0.39, p = 0.027), VM stiffness (ρ = 0.43, p = 0.015), and QT stiffness (ρ = 0.40, p = 0.022). Other stiffness measurements, such as HT and RF, showed similar patterns but were not statistically significant (p > 0.05). Table 7Associations between pain intensity (VAS) and quadriceps and hamstring muscle and tendon ultrasound parametersMeasuresSpearman correlation coefficient (ρ) P SM0.460.008HT0.330.065BF0.410.019RF0.280.127VL0.390.027VM0.430.015QT0.400.022Correlation analyses were based on measurements obtained during the first assessment by the first examiner. SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon Table 7Associations between pain intensity (VAS) and quadriceps and hamstring muscle and tendon ultrasound parametersMeasuresSpearman correlation coefficient (ρ) P SM0.460.008HT0.330.065BF0.410.019RF0.280.127VL0.390.027VM0.430.015QT0.400.022Correlation analyses were based on measurements obtained during the first assessment by the first examiner. SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon Associations between pain intensity (VAS) and quadriceps and hamstring muscle and tendon ultrasound parameters Correlation analyses were based on measurements obtained during the first assessment by the first examiner. SM Semimembranosus, HT Hamstrings tendon, BF Biceps femoris, RF Rectus femoris, VL Vastus lateralis, VM Vastus medialis, QT Quadriceps tendon Discussions The study demonstrated good to excellent intra-examiner reliability in assessing quadriceps and hamstring muscle stiffness using shear wave elastography (SWE). Furthermore, inter-examiner reliability for these variables was also good to excellent. To date, the reliability of SWE measurements for quadriceps and hamstring muscles in participants with KOA has not been reported [54]. However, SWE has shown good to excellent intra- and inter-rater reliability in the popliteus and gastrocnemius of KOA patients (ICC 0.766–0.986) [15], and moderate to high intra- and inter-session reliability in the vastus lateralis and biceps femoris of healthy adults (ICC 0.58–0.80) [55]. Notably, in a comparative study with healthy participants, Bravo-Sánchez et al. (2021) reported that Myotonometry had the highest reliability for assessing vastus lateralis stiffness (ICC = 0.93), followed by tensiomyography (ICC = 0.89), while SWE exhibited moderate reliability (ICC = 0.62) [56]. Nevertheless, applying SWE in this study offers distinct advantages for individuals with KOA. It enables direct, non-invasive quantification of tissue elasticity with high spatial resolution. This is valuable for assessing deep or anatomically complex structures, such as the quadriceps and hamstrings tendons and their peritendinous musculature. In patients with KOA, these regions are often among the first to show degenerative changes [11]. The reliability of SWE measurements in this study, both intra- and inter-examiner, varied from good to excellent (ICC = 0.78–0.98). These findings, along with low MDC and SEM values, confirm the accuracy of our measurements and are consistent with previous reports of strong SWE reliability in healthy individuals and other musculoskeletal populations [57, 58]. Zardi et al. demonstrated excellent intra-observer SWE reliability for quadriceps-related structures in healthy adults, reporting ICCs of 0.969 for the vastus medialis, 0.995 for the quadriceps tendon, and 0.989 for the patellar tendon [43]. Their findings suggest that SWE provides highly reliable measurements for the quadriceps tendon, followed by the patellar tendon. However, in our KOA cohort, we did not observe a consistent advantage for tendons. Both muscles and tendons showed similarly high ICCs and low SEMs under the standardized protocol. We infer that standardizing acquisition parameters, including joint angle, preload control, and consistent ROI depth, minimizes tissue-specific variability and enables good to excellent reliability for both tissue types [59, 60]. Similarly, Phan et al. demonstrated good intra-operator reliability (ICC > 0.75) for SWE-based quadriceps elasticity measurements, with the knee in both passive flexion (90°) and extension (180°) positions. They suggested that selecting the quadriceps scanning site based on joint positions, along with standardized imaging parameters, could lead to more consistent and reliable muscle stiffness measurements [61]. While some studies have noted variability between examiners when using SWE, others have confirmed strong reliability across and within examiners [21, 62, 63]. Interestingly, our KL-stratified analysis revealed varying patterns of measurement consistency across different stages of KOA. Participants were categorized by KOA severity (mild, moderate, severe), and reliability was assessed both within and between examiners for each subgroup. Both intra-examiner and inter-examiner reliability were generally higher in mild KOA (ICC: 0.81–0.98 and ICC: 0.82–0.97, respectively), whereas greater variability was observed in severe KOA (ICC: 0.52–0.98 and ICC: 0.57–0.97, respectively). One possible reason is that later-stage KOA is more often accompanied by guarding and heterogeneous tissue changes (e.g., fibrosis), which can make SWE readings less stable [64–66]. Given the subgroup sizes, this observation should be treated as preliminary. The SWE values for the quadriceps and hamstrings in this study are relatively low, ranging from 2.00 to 2.60 m/s across examiners and sessions. However, compared with previously reported values for the vastus lateralis (1.48 ± 0.12 m/s) and biceps femoris (1.40 ± 0.17 m/s) in healthy older adults [67], our findings indicate a relative increase in muscle stiffness in individuals with KOA. Increased quadriceps and hamstring stiffness in KOA may reflect chronic muscle guarding and fibrosis from persistent pain and low-grade inflammation, leading to lasting changes in passive elasticity [4]. Additionally, compensatory movement patterns adopted to stabilize the degenerated knee often increase co-contraction and mechanical load in surrounding muscles, thereby contributing to higher and more variable apparent stiffness [17]. It is worth noting that these KOA-specific features may increase muscle stiffness sensitivity to testing conditions, which could limit the generalizability of reliability data from healthy cohorts. However, with standardized acquisition parameters, this study demonstrated good to excellent reliability, supporting the use of SWE for clinical monitoring and disease progression tracking in KOA. Notably, the strict protocol standardization used in this study (positioning, probe placement, and acquisition conditions) may partly explain the higher ICCs relative to some healthy-cohort studies [56, 68], rather than KOA pathology alone. In the Bland–Altman plots analysis, the intra-rater reliability values showed that the upper and lower limits of agreement were narrow across the seven assessed structures (ranging from -0.259 to 0.282 m/s), and the mean differences were very small (-0.015 to 0.012 m/s). These results indicate excellent intra-rater reliability. On the other hand, in the inter-rater reliability plots, the agreement between the upper and lower limits of the means slightly decreased (approximately -0.340 to 0.374 m/s), and the mean differences increased modestly (-0.060 to 0.033 m/s). Visual inspection suggested slightly larger differences at higher mean stiffness values (SM, VL, VM, and QT), indicating a possible mild proportional bias, although no formal test was performed. This observation may reflect minor between-examiner procedural variation and greater sensitivity at the upper stiffness range. While the level of agreement remained acceptable, these findings demonstrate that intra-rater reliability surpassed inter-rater reliability, consistent with the findings reported by Dickson [69]. The observed discrepancies may result from minor variations in probe placement, manual pressure, and tissue-related factors such as skin elasticity and subcutaneous fat thickness, all of which can influence shear-wave propagation and contribute to inter-examiner variability [70]. Overall, the limits of agreement and measurement biases in this study remained narrow, indicating excellent intra- and inter-rater reproducibility of SWE measurements that were comparable to, or in some cases exceeded, findings reported in healthy cohorts [71, 72]. Beyond reproducibility, SWE-derived shear wave velocity provides an index of mechanical stiffness under the tested conditions and helps interpret KOA-related changes in muscle mechanical properties [73, 74]. In this study, we further observed weak-to-moderate associations between VAS and muscle stiffness (SM, BF, VL, VM, and QT; p < 0.05), suggesting that SWE measurements may capture clinically relevant, pain-related alterations in muscle properties. Although these findings are exploratory and unadjusted for potential confounders (age, BMI, and KL grade), they provide initial clinical context for SWE beyond reproducibility. Because pain is only one aspect of KOA, future studies should further examine how SWE measures relate to functional and mechanical indicators (muscle strength and passive torque) and to structural alterations in muscle tissue, to clarify better what SWE-derived stiffness represents in this population. It appears that the reliability of SWE depends on several factors. Location and muscle depth are important. Superficial muscles, such as the vastus lateralis, have greater reliability than deeper muscles like the vastus intermedius because shear wave propagation is less efficient in deeper tissues [73]. Moreover, improper sampling box placement, imaging settings, or ROI size can negatively affect stiffness measurements in SWE. Recent studies have shown that such factors reduce measurement repeatability and interfere with shear wave propagation [75]. Physiological factors are also important considerations in the reliability of SWE measurements. Muscle stiffness varies throughout the day and responds to recent physical activity. Dankel et al. reported that stiffness increases in the morning or after exercise due to residual muscle tension [76]. Bernabei et al. observed that higher muscle temperatures reduce stiffness by increasing tissue elasticity [74]. To control these variables, all measurements in this study were conducted within a consistent time window in a temperature-controlled room, and participants refrained from intense physical activity for at least 12 h prior to testing. However, this strict control may not be suitable for short-interval post-intervention reassessments (e.g., immediately after exercise or therapy), where activity restrictions and fixed scheduling are impractical. Future studies should consider protocol variants for rapid retesting (e.g., recording time-since-activity, allowing shorter washouts, or modeling temperature and preload as covariates). Future directions and clinical implications Our findings demonstrate that SWE is a reliable method for measuring quadriceps and hamstring stiffness in patients with knee osteoarthritis. Its portability, affordability, and short acquisition time support its use in routine clinical practice to objectively assess the effects of interventions and monitor progress. Clinicians can use a quantified measurement-error threshold to determine whether changes reflect true biological response and to adjust rehabilitation intensity, exercise progression, or follow-up intervals. Future research should establish KOA-specific reference values and minimal detectable change, evaluate responsiveness to rehabilitation and pharmacological treatments, assess the prognostic value for function and pain, and standardize acquisition parameters, such as knee angle, region of interest, and probe pressure, across centers and devices. Limitations This study presents several inherent limitations. First, neuromuscular activation was not assessed, which limits our ability to account for muscle activity and potential quadriceps–hamstring co-activation during SWE. Although standardized procedures were used to minimize active tension, subtle activation cannot be fully ruled out without EMG monitoring. Second, all stiffness measurements were collected in static, resting positions and may therefore not reflect muscle behavior during clinically relevant functional or weight-bearing tasks in KOA. Finally, although most quadriceps and hamstring muscles were assessed, the generalizability of these findings is restricted to participants with a clinical diagnosis of KOA. Conclusion The assessment of quadriceps and hamstring muscle stiffness using SWE shows good to excellent intra-examiner and inter-examiner reliability in individuals with KOA. Its strong reproducibility supports the use of SWE as a non-invasive tool for monitoring muscle changes related to joint dysfunction in KOA. The MDC values obtained in this study can help clinicians assess whether observed stiffness changes are real. By detecting stiffness variations early, SWE can help clinicians track disease progression and predict functional recovery. Future research with larger, more diverse populations is warranted to validate these results and to explore the longitudinal implications of changes in muscle stiffness for joint health management.