Active-Matrix Sensing Array Assisted with Machine-Learning Approach for Lumbar Degenerative Disease Diagnosis and Postoperative Assessment

Lumbar degenerative disease (LDD) refers to the nerve compression syndrome such as neurogenic intermittent claudication and lower limb pain, which disturbs people's daily life, and its incidence increases with age. Traditional diagnosis often employs magnetic response imaging or other imaging examinations. But the radiological data have uncertain clinical correlation and often be overemphasized in clinical decision making. Here, an active-matrix sensing array (AMSA) is proposed to measure plantar pressure during walking, in order to improve LDD diagnostic processes. An array of piezoelectric sensors with high robustness are assembled. Combined with a support vector machine (SVM) supervised learning algorithm, the system can classify the common human motions of half-squat, squat, jump, walk and jog with an accuracy up to 99.2%, demonstrating its capability of recognizing personal activities. More importantly, in 62 clinical samples of lumbar degenerative patients, the system can perform an artificial intelligence diagnosis, achieving an accuracy of 100% with an area under receiver operating characteristic curve of 0.998, and also gives out recovery assessments after surgery. Since the personal plantar pressure is also indicative of other diseases, such as diabetes and fasciitis, the system can be extended to other medical aspects, showing a broad impact in biomedical engineering.

Automated summary

This study proposes the use of an active-matrix sensing array to measure plantar pressure and improve the diagnosis of lumbar degenerative disease (LDD). By combining piezoelectric sensors and a support vector machine learning algorithm, the system can accurately classify human motions and perform artificial intelligence diagnosis of LDD with high accuracy. The system also has the potential to be applied in other medical areas, showing a broad impact in biomedical engineering.