Lee Sunghoon I, Campion Andrew, Huang Alex, Park Eunjeong, Garst Jordan H, Jahanforouz Nima, Espinal Marie, Siero Tiffany, Pollack Sophie, Afridi Marwa, Daneshvar Meelod, Ghias Saif, Sarrafzadeh Majid, Lu Daniel C
College of Information and Computer Science, UMass Amherst, Amherst, USA.
Neuroplasticity and Repair Laboratory, UCLA, Los Angeles, USA.
J Neuroeng Rehabil. 2017 Jul 18;14(1):77. doi: 10.1186/s12984-017-0288-0.
Approximately 33% of the patients with lumbar spinal stenosis (LSS) who undergo surgery are not satisfied with their postoperative clinical outcomes. Therefore, identifying predictors for postoperative outcome and groups of patients who will benefit from the surgical intervention is of significant clinical benefit. However, many of the studied predictors to date suffer from subjective recall bias, lack fine digital measures, and yield poor correlation to outcomes.
This study utilized smart-shoes to capture gait parameters extracted preoperatively during a 10 m self-paced walking test, which was hypothesized to provide objective, digital measurements regarding the level of gait impairment caused by LSS symptoms, with the goal of predicting postoperative outcomes in a cohort of LSS patients who received lumbar decompression and/or fusion surgery. The Oswestry Disability Index (ODI) and predominant pain level measured via the Visual Analogue Scale (VAS) were used as the postoperative clinical outcome variables.
The gait parameters extracted from the smart-shoes made statistically significant predictions of the postoperative improvement in ODI (RMSE =0.13, r=0.93, and p<3.92×10) and predominant pain level (RMSE =0.19, r=0.83, and p<1.28×10). Additionally, the gait parameters produced greater prediction accuracy compared to the clinical variables that had been previously investigated.
The reported results herein support the hypothesis that the measurement of gait characteristics by our smart-shoe system can provide accurate predictions of the surgical outcomes, assisting clinicians in identifying which LSS patient population can benefit from the surgical intervention and optimize treatment strategies.
接受手术的腰椎管狭窄症（LSS）患者中，约33%对术后临床结果不满意。因此，确定术后结果的预测因素以及将从手术干预中获益的患者群体具有重大临床意义。然而，迄今为止，许多已研究的预测因素存在主观回忆偏差，缺乏精细的数字测量，且与结果的相关性较差。
本研究利用智能鞋在10米自定步速行走测试中术前采集步态参数，假设该测试能提供关于LSS症状导致的步态损害程度的客观数字测量，目的是预测接受腰椎减压和/或融合手术的LSS患者队列的术后结果。采用Oswestry功能障碍指数（ODI）和通过视觉模拟量表（VAS）测量的主要疼痛程度作为术后临床结果变量。
从智能鞋提取的步态参数对ODI术后改善情况（均方根误差=0.13，r=0.93，p<3.92×10）和主要疼痛程度（均方根误差=0.19，r=0.83，p<1.28×10）做出了具有统计学意义的预测。此外，与先前研究的临床变量相比，步态参数具有更高的预测准确性。
本文报告的结果支持以下假设，即我们的智能鞋系统对步态特征的测量可以准确预测手术结果，帮助临床医生确定哪些LSS患者群体可以从手术干预中获益，并优化治疗策略。
