肝臟腫瘤是華人常見的疾病，早期肝腫瘤的診斷與治療是避免惡化很重要的措施。超音波掃描常被用來做第一線的篩選診斷，若病症需要治療，常用的肝腫瘤治療方式有酒精注射或RF 燒灼或動脈血管栓塞，嚴重者則進行(微創)肝臟切除手術。近年來快速發展的高強度聚焦超音波(High Intensity Focused Ultrasound, HIFU)以高溫方式破壞腫瘤細胞，此一非侵入式的熱治療方式極具發展潛力。 本研究目標是以超音波影像輔助機器手臂定位的HIFU 肝腫瘤治療系統，透過超音波影像導引技術，由超音波影像作三維重建並計算腫瘤尺寸與方位資訊，並依據HIFU燒灼方向對三維模型進行重切，方便後續的路徑規劃或是選取要燒灼治療的位置並記錄空間位置資料，再據以由電腦控制機械手臂移動HIFU 探頭至目標。 實驗部分主要是測量系統的準確度，分為超音波影像校正誤差量測、機械手臂結合光學定位器之空間定位誤差量測、整體系統誤差量測三個部分，實驗的結果顯示影像校正誤差平均誤差為1.02 ± 0.41mm，機械手臂空間定位誤差為0.08±0.06mm，整體系統誤差為1.06±0.38mm，可以順利的經由超音波影像導引機械手臂至目標位置。另外進行三維模型重建誤差量測，實驗結果位置誤差為2.06±0.45mm。最後實際以超音波影像導引機械手臂移動HIFU探頭燒灼蛋白假體，燒灼位置誤差在2.1mm以下。; Liver tumor (liver cancer) is a very common disease for Chinese people. Early diagnosis and treatment of liver tumor are important measures to avoid worsening. Ultrasound scanning is usually adopted for first-line screening and diagnosis. If the lesion needs treatment, percutaneous ethanol injection, radiofrequency ablation or arterial embolization is common treatment method. For serious cases, liver dissection treatment will be necessary. In recent years, noninvasive thermal treatment by High Intensity Focused Ultrasound (HIFU) is highly potential in tumor treatment. This project is aimed to study and develop an ultrasound assisted robotic positioning system for treating liver tumor by HIFU. With the assistance of ultrasound image guidance technology, ultrasound images can be reconstruct into the three-dimensional model of the tumor with size and location. We re-cut the three-dimensional model into parallel slices based on the HIFU probe direction, which is convenient for the follow path planning. Then, a computer controlled robot will move HIFU probe to follow in the planned path and execute thermal treatment. The experiments are mainly to measure the ultrasound imaging calibration error, robot spatial positioning error and overall system error. The results of several experiments show that distance errors are 1.02 ± 0.41mm, 0.08 ± 0.06mm, 1.06 ± 0.38mm, respectively. Finally, move the HIFU focus point to ablate polyacrylamide gel phantom. The experiment positioning error is less than 2.1mm.
