骨科手術用的C-arm影像輔助手術導引系統已經在臨床上使用多年， 椎莖螺絲植入手術是重要的應用之一，然而導引定位的準確性欠缺定量的分析，且在前後向(Anterior–Posterior View, AP-View)的C-arm影像上很難決定植入的方向，所以本論文針對實驗室所研發的導引系統進行誤差分析與改善，並以DRR(Digitally Reconstructed Radiographs)資料歸納方式建立在AP影像上進行路徑規劃的依據。 在導引誤差部分，本論文採用三種方法進行驗證，包括1.計算手術器械在AP-View、LA-View上的投影，再比較其和器械影像的距離誤差；2.將器械影像進行逆投影得到空間座標，再比較其和實際量測的器械空間座標的誤差；3.設定植入的安全範圍，評估器械是否於安全範圍內。實驗結果顯示導引誤差主要是人體呼吸造成影像與影像投射方位資料的接收不同步，反光球的新舊對DRF(Dynamic Reference Frame)座標量測值影響相對微小。 在手術規劃方面，先在重建的3D CT模型上進行路徑規劃，之後進行DRR投影，得到AP影像上的路徑方位，並據以歸納出以脊椎特徵點為依據的路徑規劃原則。手術時醫生依此一原則在AP影像上進行手術路徑規劃，如此應可降低人為路徑規劃的不確定性，提高手術的成功率。 C-arm image assisted navigation system for orthopedic surgery has been applied clinically for several years. Pedicle screw implantation is one of the important applications. However, positioning accuracy is still lack of quantification analysis. Also, the implantation direction on the Anterior/Posterior-view image is difficult to plan. Therefore, in this research, we analyze and improve positioning errors of the C-arm assisted navigation system developed by our laboratory and establish a path planning reference on the AP image based on Digitally Reconstructed Radiographs of CT images. Three methods are applied to verify positioning errors. One is to calculate the position errors of calculated AP-View and LA-View images with the real image of the surgical instrument. Two is to calculate the errors of the inversely projected spatial position according to the AP-View image and LA-View image of the surgical instrument with its measured spatial position. Three is to define a safe zone and evaluate whether the instrument is within the zone. The experimental results show that the main factor of positioning error is due to respiration of the patient, which causes the asynchronous catch of C-arm images and the corresponding coordinates of the X-ray source. On the other hand, the new or used reflective balls only have minor effect in measuring the DRF coordinates. As to surgical path planning, the path is planned on the reconstructed 3D CT model. Then, it is projected onto so-called AP-view through Digitally Reconstructed Radiographs algorithm. The projected paths are defined according to the feature points of the spine. The average location of the paths is the reference to plan surgical path on the AP-view, which might reduce the uncertainty of path planning and promote success rate of operation.
