人工膝關節置換術中 髕骨未置換病患之膝關節振動訊號分析

Abstract

摘 要 人體之關節如同機器一般，在過度使用或使用不當下會造成功能之損害或退化，而在骨科臨床上發現，當膝關節發生病變時，其活動時會產生異常的聲音，亦即膝關節在擺動下所產生的振動訊號，本研究即是希望藉由分析此一振動訊號來診斷區分不同之膝關節病症。 本研究運用自迴歸模型(Autoregressive Model)與希爾柏-黃轉換(Hilbert-Huang transform)--適應性訊號時頻分析處理方法於膝關節振動訊號分析，研究對象為:正常者、前十字韌帶重建術後病變患者及退化性膝關節炎患者三群組間之分析，及正常者、髕骨有置換病患與髕骨未置換病患三群組間之分析，並從訊號的時域、頻域、時頻域分析中提取出可用來區分不同病症的訊號特徵參數。 研究結果顯示，希爾柏-黃轉換能對本質為非線性、非穩態的膝關節振動訊號有較佳之解析，能在訊號的時頻分佈中提取出較具區分能力的特徵參數，另外在時頻分佈瞬時特性所得出之特徵參數，可以針對某時間區段的訊號時頻分佈作分析，進而更深入的檢視膝關節在什麼角度情況下會有差異性存在，如能在日後用來當作偵測膝關節病變的工具，能讓醫師臨床診斷更加的精準，因此對特性為非線性非穩態的膝關節振動訊號而言，使用時頻分析方法來解析膝關節振動訊號，是相當適合且具有發展潛力的。 由於關節振動訊號測量術(Vibration Arthrometry)為一非侵襲性、簡單方便且低成本的膝關節病變診斷工具，選擇適當的診斷治療方式，能使病人免除痛苦並避免浪費醫療資源，如能持續發展，未來將能成為骨科醫師用來診斷病人的重要工具。

Abstract The joint of human body is just like the machine.It liable to cause the damage and degeneration when use it excessively or incorrectly.There is a phenomenon which abnormal joint sound arises from knee joint pathology during knee motion can be detected in the clinical diagnosis.The knee joint could produce vibration signals during normal flexion-extension motion.The objective of this study is to diagnose and discriminate the pathologies of the knee joint by analyzing these vibration signals. This study applied Autoregressive Model and Hilbert-Huang transform--adaptive signal-analysis methodology to the knee joint vibration signal analysis.The methods were tested on the signals of the groups of normal、anterior cruciate ligament(ACL) reconstruction patients and degenerative osteoarthristis patients,and the groups of normal、patella-resurfacing and patella-nonresurfacing TKA patients.The objective of this study is to extract and identify the relevant features in the time domain、frequency domain and time-frequency domain which could differentiate the varieties of pathologies. We have found that application of Hilbert-Huang transform in knee joint vibration signals which are nonstationary and nonlinear in nature gives better ability of analysis.The features extracted from time-frequency distribution show better discriminative ability.We can estimate the knee joint angle which has discrimination between the different knee joint pathologies by analyzing the time-frequency distribution of signal.If this method can be applied to examine the knee joint pathologies in the future,the doctor will make more precise clinical diagnosis. Therefore, time-frequency analysis methods are suitable and have shown good potential for analysis of knee joint vibration signals which are nonstationary and nonlinear in nature. Vibration Arthrometry provided a noninvasive、convenient and cheap clinic tool for diagnosing knee joints. Patients can avoid aches,and the medical resources can be reserve by appropriate therapy.If this research can be continued advancing,the achievements will be the significant clinical diagnosis tools in the future .