椎體成形術中，通常只使用兩張X光影像醫師診斷及進行術前規劃，但由於脊椎的幾何結構複雜，上述資料無法提供醫師足夠的資訊。在術中，為了補強脊椎的結構強度，必須將骨水泥打入脊椎中進行加強，但其作業上必須一直拍攝多張C-Arm來觀看目前骨水泥的灌注情形，導致醫療人員的輻射量大增。為了讓醫師在術前對於骨水泥灌注量有一個大致上的掌握，本研究發展自動塌陷椎體復位技術，其概念為使病患塌陷之椎體經由變形技術復原成破損前的樣子，最後利用變形前後之體積差異提供給醫師作為骨水泥灌注量之參考，本椎體自動復位技術重點包含：(1)病患本身三維椎體模型重建輸入，(2)自動尋找塌陷椎體的表面位置，(3)自動尋找塌陷椎體於變形復位後的表面位置，(4)保留椎體原有特徵之限制，藉此完成椎體復位變形。本研究藉由數個模擬塌陷椎體模型驗證自動塌陷椎體復位技術之可靠度，並以實際病患案例進行變形復位，針對體積變化的合理性進行探討與分析，說明本研究提出之方法的可行性。 In vertebroplasty surgery, the surgeons usually use two pieces of X-ray images for diagnosis and surgical planning. Because of the complicated geometry in spine, such information is usually not rich enough for decision making. During surgery, the surgeons inject bone cement into the center of the collapsed spinal vertebra to stabilize and strengthen the crushed bone. It is necessary for the surgeons to take plenty of C-arm images to realize the real time position and orientation of the tools as well as the cement. The increasing radiation exposure for this kind of surgery will probably bring the surgeons some diseases in the future. To assist the surgeons knowing the volume of the bone cement injected, this research intends to develop an auto-reduction of subsided vertebral. The idea is to deform collapsed vertebra into its normal shape, and then to use the volume difference as a bone cement perfusion reference. The proposed auto-reduction technique mainly contains the following features : (1)input patient 3D vertebral model, (2)compute the surface of vertebral body, (3)evaluate the objective surface of the deformed vertebral body, and (4)retain the origin al feature of the vertebra. In this study, several examples, including artificial models and real patient models, are employed to demonstrate the feasibility of the proposed method.
