陳永耀Chen, Yung-Yaw臺灣大學:電機工程學研究所張鈺堂Chang, Yu-TangYu-TangChang2010-07-012018-07-062010-07-012018-07-062009U0001-2307200911433300http://ntur.lib.ntu.edu.tw//handle/246246/188136惡性腫瘤高居台灣十大死因之首，有效的治療方式中，超音波加熱治療較傳統治療方法有更低的非侵入性及副作用，而為極有潛力之腫瘤治療方法。進行超音波熱治療時，必須準確地聚焦在所要治療的患部以避免在正常的組織給予過多的熱劑量。如果治療的腫瘤是會隨著呼吸而產生週期性往復運動的動態腫瘤如肺癌、肝癌，必須準確的定位運動中的腫瘤，並驅動超音波探頭做追蹤加熱治療。本研究針將對肝腫瘤做探討。由於肝臟位於人體腹腔內，現有之醫學影像系統如電腦斷層影像和核磁共振影像可以直接準確的觀察到肝臟位置，但其速度仍遠低於即時控制所需。為了即時得知肝臟位置達到控制目的，本研究希望藉由間接量測與呼吸有關之替代訊號且建立和肝臟運動之關聯性以即時估測肝臟之運動。 本研究和台大醫學院外科醫師和技術員合作將位置感測器植入活體肝臟取得其位置資訊，同時量測胸腹起伏資訊。藉由分析肝臟運動和胸腹運動建立其關聯性以即時估測肝臟運動。在關聯性的分析中呼氣和吸氣運動對應到不同軌跡之遲滯現象會造成肝臟位置估測之不準確性，本研究提出橢圓模型來解決此一遲滯現象，比較文獻上各模型之估測誤差說明此橢圓擬合之優點。最後建立一套模型選擇準則來選取其適合之運動模型數。同時本研究也比較不同呼吸模式所帶來的影響和多點量測或單點量測之優劣。最後希望能找出肝臟之運動預測模型以準確的即時估計肝腫瘤位置。High Intensity Focused Ultrasound (HIFU) thermal therapy is regarded as one with great potential due to its non-invasion and low side-effect. In order to treat mobile tumors with periodic motions induced by respiration such as lung and liver tumor, the tumor positions have to be located precisely and transducer has to be driven to track mobile tumor. Some medical image system such as CT and MRI can locate liver position precisely. However, the frame rate of medical image system is not high enough to reach our controlling requirement. Our goal is building the correlation between liver motion and surrogated signals to locate liver tumor in real-time. In this study, the doctors and technicians from National Taiwan University Hospital will cooperate with us to conduct in vitro experiment for liver motion data acquisition and analysis. The relationship between liver motion and surrogated signals such as chest and abdomen motion is constructed. The ellipse model is proposed to solve the hysteresis. We also compare the influence of different breath mode and performance between single input and multiple inputs. Finally, the prediction model is expected to be constructed and estimating liver position in real-time.摘要 Ibstract IIontents IIIIist of Figures Vist of Tables VIIIhapter 1 Introduction 1.1 Background 1.2 Motivation 8.3 Organization 10hapter 2 Liver Motion and Surrogated Signals 11.1 Characteristic of Liver Motion and Surrogated Signals 11.2 Measurement System and Plant 15.2.1 Electromagnetic motion tracker 15.2.2 Respiratory motion sensor 19.2.2 Sensing System Integrationl 21hapter 3 Mapping Model Building 24.1 Related Work 24.2 Correlaiotn between Internal and External signals 28.2.1 Linear Correlation 28.2.2 Cross Correlaiotn 30.2.2 The Influence of Delay Time 32.3 Modeling 35.3.1 Linear Model 35.3.2 Quadratic Model 37.3.3 Ellipse Model 40.3.4 Error Rate 45.4 Multiple Signals Fitting 46hapter 4 Animal Experiment 48.1 Setup of Animal Experiment 48.1.1 Application 48.1.2 Operation 49.1.3 Breath Motion Sensing System Setup 50.1.4 Data Measure 51.2 Characteristic of Respiration and Liver Motion 56.2.1 Respiration 56.2.2 Liver Motion in Different Locations 57hapter 5 Analysis and Results 61.1 Models Analysis in Active and Passive Mode 61.1.1 Active Breath Model 62.1.2 Passive Breath Model 67.1.3 Model Choosing Criteria 72.2 Single Signal Fitting and Multiple signals Fitting 75.2.1 Single Signal Figgingl 75.2.2 Multiple Signals Figgingl 77.3 Comparison of Active and Passive Mode 79hapter 6 Conclusions ans Future Works 81eference 854968206 bytesapplication/pdfen-US高能聚焦超音波肝腫瘤追蹤關聯性分析遲滯橢圓模型預測模型HIFUliver tumor trackingcorrelationhysteresisellipse modelthesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/188136/1/ntu-98-R96921046-1.pdf