電腦輔助人工膝關節術後定位與磨損分析之技術發展

Development of Computer-Aided Technique to Postoperatively Evaluate Registration and Damage for Total Knee Replacement

許啟彬(Hsu, Chi Pin)

指導教授 ： 賴景義

國立中央大學/工學院/機械工程學系/博士(2011年)

摘要

人工膝關節置換已是相當普及的手術，而人工膝關節會因長期使用產生塑膠墊片或金屬元件的磨損，如產生過多的磨屑可能會導致骨溶解，因此手術後的追蹤觀測金屬元件對位情況與塑膠墊片磨損情形是相當重要的。傳統觀測大多使用X光影像，而二維影像難以精確判斷出元件對位或墊片磨損情形，因此發展電腦輔助的X光影像三維定位技術來協助分析。傳統X光定位於硬體方面必頇使用昂貴的垂直正交X光機或特殊拍攝姿勢來固定膝關節取得X光影像，軟體方面使用的粗定位方法為手動調整或資料庫比對，甚至對於精定位的影像夾角也必頇預先取得，因此對於效率與臨床應用還有很大的進步空間。

本研究針對現有X光定位之缺點，於硬體方面設計膝關節固定器與旋轉平台來拍攝不同方位之X光影像，於效率提升方面以網格簡化與特徵分離方法減少光學掃描的網格模型資料，X光定位技術方面則提出自動化的特徵辨識粗定位方法提升定位效率，精定位則使用輪廓比對技術，並以獨立與合併比對之流程計算正確的影像夾角提升定位精確性，最後將定位結果反推計算塑膠墊片的磨損區域。實驗驗證方面除針對演算法正確性的模擬測詴外，還進行體外與人體實驗，利用體外夾治具與旋轉平台分別進行之，而實驗結果發現整體的精確度約0.3 mm，對於僅使用肉眼觀察的臨床醫師或長期的追蹤診斷已相當足夠，同時也能提升X光定位技術於臨床適用性。

關鍵字

術後評估 ； 影像處理 ； 影像定位 ； 人工膝關節 ； 電腦輔助診斷

並列摘要

After total knee replacement, the monitoring of the prosthetic performance is often done by roentgenographic examination. However, the two-dimensional (2D) roentgen images only provide information about the projection onto the anteroposterior (AP) and medrolateral (ML) planes. In particular, the radiolucent insert is not obvious and the wear of the insert is difficult to exactly define. The model-based roentgen stereo-photogrammetric analysis (RSA) has recently been developed to in vivo estimate the prosthetic pose (i.e. location and orientation). For hardware of model-based RSA, specific X-ray equipment (e.g. dual) or photography posture (e.g. supine) is used to synchronize the capture of two knee images and incline angle. For software of model-based RSA, the excessive elements of mesh model and the process of manual adjustment inevitably leads to low efficiency and practicability.

In this study, one hardware design, two mesh-manipulating methods, and two RSA algorithms were developed to improve the clinical applicability of the RSA method. Firstly, this study designed the rotation platform to avoid the use of the dual synchronized X-ray system. In addition, two mesh-manipulating methods (QEM simplified method and feature-segmented method) for knee model were developed to decrease the element number of the models and thus increase the execution efficiency. Finally, the feature-recognized and outline-optimized algorithms were further used to automatically estimate the rough registration and measure the exact inclined angle of two X-ray photos. The simulative and experimental tests were used to validate clinical applicability, robustness, and accuracy of the aforementioned methods.

並列關鍵字

Postoperatively evaluate ； Image processing ； Image registration ； Knee prosthesis ； Computer aided diagnosis

參考文獻

[65] J. Y. Lai and Y. C. Tsai, “A data segmentation technique for triangular meshes in reverse engineering”, Journal of the Chinese Society of Mechanical Engineers, Vol. 28, No. 3, pp. 289-300, 2007.

[1] G. C. Sutton, “Computer‐aided diagnosis: A review”, British Journal of Surgery, Vol. 76, No. 1, pp. 82-85, 1989.

[2] G. C. Sutton, “How accurate is computer-aided diagnosis?”, The Lancet, Vol. 334, No. 8668, pp. 905-908, 1989.

[3] J. Karrholm, “Roentgen stereophotogrammetry. Review of orthopaedic applications”, Acta Orthopaedica Scandinavica, Vol. 60, No. 4, pp. 491-503, 1989.

[5] G. Selvik, “Roentgen stereophotogrammetry- a method for the study of the kinematics of the skeletal system”, Acta Orthopaedica Scandinavica, Vol. 232, pp. 1-51, 1989.