| 摘要: | 微創椎莖螺釘植入手術為高風險與高難度的手術，傳統手術在過程中醫師須拍攝大量C-arm X光影像，以確認手術器械是正確的，因此造成病患與醫護人員吸收許多的輻射劑量，此外醫師徒手進行手術，手部晃動易造成的手術器械方位誤差，因此發展微創椎莖螺釘植入用的機械臂輔助導引系統是需要的。
本研究參考實驗室既有的機械臂定位技術與3D C-arm影像輔助導航技術，發展一微創脊椎手術用3D C-arm影像輔助機械臂導引系統。研究項目主要包括機械臂定位系統與導航系統的軟體整合並改良導航系統鋼珠辨識方法、導引輔助器械設計、病患與影像的註冊、手術路徑規劃視窗的路徑平移與旋轉拉霸、及機械臂導引定位誤差實驗。另針對註冊板上的鋼珠影像，發展3D建模參數自動設定方法，期望在不同3D C-arm設備與X光強度設定下皆可自動設定；又透過連通區域分析方法與球體特性，建立鋼珠模型辨識方法，以利不同C-arm設備下使用。
機械臂導引定位系統的定位誤差實驗，以驗證板與驗證器械的鋼珠特徵點作為定位目標，將特徵點空間座標作為對照組(C)；特徵點影像座標作為實驗組(T)；機械臂導引器械至規劃路徑後的器械空間方位座標作為實驗組(T^′′)，將實驗組(T)的影像座標轉換至空間座標後計算註冊誤差(|C-T|)與機械臂導引定位誤差(|C-T^′′ |)。經數次實驗，驗證板之|C-T|=0.86±0.10mm；|C-T^′′ |=1.27±0.47mm。驗證器械之|C-T|：結束點0.89±0.40mm、進入點0.72±0.29mm、方向0.43±0.13°；|C-T^′′ |：結束點1.67±0.36mm、進入點1.51±0.28mm、方向0.54±0.24°。上述的誤差值都小於椎莖螺釘植入手術的2mm臨床需求範圍。
關鍵字：3D C-arm、手術導引、機械手臂、脊椎手術
;Pedicle screw implantation is a high-risk operation. In the conventional operation, the surgeon must take many C-arm images to confirm that the surgical instruments will not injure the spinal cord by mistake, which will cause patients and medical staff to expose high radiation risk. In addition, the operation is highly relying on the surgeon’s skill and clinic experiences. Therefore, it is needed to develop a robotic navigation system that can assist surgeons to position and hold surgical tools for spine surgery.
This research refers the existing robotic positioning and 3D C-arm navigation technology in the laboratory to develop a 3D C-arm image assisted robotic navigation system. The main studies include the software integration of robotic system and navigation system with segmentation method improvement of steel marker images, design of assistive navigation devices, patient and image coordinate registration, design of path planning window with translation and rotation buttons, and experiments of robotic positioning accuracy. Further, an automatic image threshold setting for separating steel markers from the C-arm image is developed, which should be applicable for different 3D C-arm and X-ray intensity setting. Also, connecting area analysis and characteristics of sphere are used to identify features of steel sphere models so that intuitive facilitate adjustment can be done when different C-arm equipment is used.
The average positioning distance and direction error of the system were tested by using the verification board (V) and the verification device (SV). The coordinates of physical features of the testing devices are defined as control group(C)；the transferred coordinates of image features of the testing devices are defined as group(T)；the position and direction of the positioning probe navigated by the robot to the planned path are defined as group(T^′′). After at least three-time testing, the registration error(|C-T|) and system positioning error (|C-T^′′ |) are |C-T|=0.86±0.10mm, |C-T^′′ |=1.27±0.47mm based on V board；|C-T|：0.89±0.40mm (end point), 0.72±0.29mm (entry point), 0.43±0.13° (direction)；|C-T^′′ |：1.67±0.36mm (end point), 1.51±0.28mm (entry point), 0.54±0.24° (direction) based on SV device. The overall position errors are lower than 2mm, a requirement for pedicle screw implantation.
Keyword：C-arm Image, Surgical Navigation, Robot, Spine Surgery |
