The spine journal : official journal of the North American Spine Society · 2026
Wearable sensor-based gait evaluation in cervical spondylotic myelopathy: detecting asymptomatic gait dysfunction and heterogeneous gait impairment.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Li Zhenxu, Tie Dazhao, Bai Wenbin, Xia Tian, Pan Shengfa, Chen Xin, Zhao Yanbin, Zhao Minwei, Zhou Feifei
Аннотация
BACKGROUND CONTEXT: Current severity assessment methods for cervical spondylotic myelopathy (CSM) overlook asymptomatic gait dysfunction and potential heterogeneous gait impairment. Also, gait alterations for disease progression remain inadequately characterized.
PURPOSE: This study aimed to detect asymptomatic gait dysfunction, achieve accurate detection of CSM, and discover heterogeneous gait impairment in CSM using wearable sensor-based gait evaluation.
STUDY DESIGN: Cross-sectional study.
PATIENT SAMPLE: A total of 45 CSM patients and 39 age-matched healthy controls (HC) were recruited for this study.
OUTCOME MEASURES: Waist and lower extremity kinematics and spatiotemporal parameters.
METHODS: This study utilized a 7-sensor wearable inertial measurement unit (IMU) system to obtain spatiotemporal and kinematic (waist and lower extremity joint angle time series in triaxial planes) during 2-minute walk tests (2MWT). Group comparisons were performed using the Kruskal-Wallis test to assess differences in gait parameters between lower extremity asymptomatic CSM patients CSM-ALE, cervical spondylotic myelopathy asymptomatic lower extremity (CSM-ALE) and healthy controls (HC), and between the overall CSM cohort and HC. Machine learning was used for asymptomatic gait dysfunction and CSM detection. Graph convolutional network (GCN)-based unsupervised clustering was used for heterogeneous gait impairment recognition. The Pearson correlation coefficient was used to examine the relationships between gait parameters and Japanese Orthopedic Association (JOA) scores in order to evaluate the applicability of gait metrics for monitoring CSM severity.
RESULTS: For gait parameters, CSM patients showed significantly longer turn time (3.06 vs 2.60 seconds), longer gait cycle time (1.28 vs 1.21 seconds), greater waist flexion angle (4.00° vs 2.95°), shorter step length (0.66 vs 0.83 m), slower single-step speed (1.03 vs 1.38 m/s), and reduced hip (76.31 vs 105.19 °/s) and knee angular velocities (146.98 vs 180.77 °/s) compared with healthy controls (HC). Using fused kinematic time-series data, the SVM classifier achieved excellent discrimination between CSM-ALE patients and HC (AUC=0.996) as well as between all CSM patients and HC (AUC=0.997), outperforming analyses based on individual gait parameters. Unsupervised GCN clustering revealed 2 distinct gait impairment patterns. Pattern I was characterized by prolonged turning time (4.03 vs 2.71 seconds) and increased waist flexion angle (9.27° vs 2.99°) compared with Pattern II. In contrast, Pattern II was marked by reduced ankle velocity (81.05 vs 100.82 °/s) but preserved turning time (2.71 vs 2.60 seconds vs HC), suggesting distal motor impairment with relatively maintained balance control. JOA scores was correlated with gait speed (r=0.431, p=.003), swing phase duration (r=0.444, p=.002), and knee angular velocity (r=0.373, p=.012).
CONCLUSION: Wearable sensor-based gait evaluation can effectively distinguish CSM patients, including CSM-ALE patients, from HC, and reveal distinct patterns of gait impairment. These findings highlight its clinical value not only for early diagnosis but also for potentially enabling intervention within a therapeutic window before significant functional decline occurs.
PURPOSE: This study aimed to detect asymptomatic gait dysfunction, achieve accurate detection of CSM, and discover heterogeneous gait impairment in CSM using wearable sensor-based gait evaluation.
STUDY DESIGN: Cross-sectional study.
PATIENT SAMPLE: A total of 45 CSM patients and 39 age-matched healthy controls (HC) were recruited for this study.
OUTCOME MEASURES: Waist and lower extremity kinematics and spatiotemporal parameters.
METHODS: This study utilized a 7-sensor wearable inertial measurement unit (IMU) system to obtain spatiotemporal and kinematic (waist and lower extremity joint angle time series in triaxial planes) during 2-minute walk tests (2MWT). Group comparisons were performed using the Kruskal-Wallis test to assess differences in gait parameters between lower extremity asymptomatic CSM patients CSM-ALE, cervical spondylotic myelopathy asymptomatic lower extremity (CSM-ALE) and healthy controls (HC), and between the overall CSM cohort and HC. Machine learning was used for asymptomatic gait dysfunction and CSM detection. Graph convolutional network (GCN)-based unsupervised clustering was used for heterogeneous gait impairment recognition. The Pearson correlation coefficient was used to examine the relationships between gait parameters and Japanese Orthopedic Association (JOA) scores in order to evaluate the applicability of gait metrics for monitoring CSM severity.
RESULTS: For gait parameters, CSM patients showed significantly longer turn time (3.06 vs 2.60 seconds), longer gait cycle time (1.28 vs 1.21 seconds), greater waist flexion angle (4.00° vs 2.95°), shorter step length (0.66 vs 0.83 m), slower single-step speed (1.03 vs 1.38 m/s), and reduced hip (76.31 vs 105.19 °/s) and knee angular velocities (146.98 vs 180.77 °/s) compared with healthy controls (HC). Using fused kinematic time-series data, the SVM classifier achieved excellent discrimination between CSM-ALE patients and HC (AUC=0.996) as well as between all CSM patients and HC (AUC=0.997), outperforming analyses based on individual gait parameters. Unsupervised GCN clustering revealed 2 distinct gait impairment patterns. Pattern I was characterized by prolonged turning time (4.03 vs 2.71 seconds) and increased waist flexion angle (9.27° vs 2.99°) compared with Pattern II. In contrast, Pattern II was marked by reduced ankle velocity (81.05 vs 100.82 °/s) but preserved turning time (2.71 vs 2.60 seconds vs HC), suggesting distal motor impairment with relatively maintained balance control. JOA scores was correlated with gait speed (r=0.431, p=.003), swing phase duration (r=0.444, p=.002), and knee angular velocity (r=0.373, p=.012).
CONCLUSION: Wearable sensor-based gait evaluation can effectively distinguish CSM patients, including CSM-ALE patients, from HC, and reveal distinct patterns of gait impairment. These findings highlight its clinical value not only for early diagnosis but also for potentially enabling intervention within a therapeutic window before significant functional decline occurs.