利用四維動態超音波資料定量分析左心室功能及型態

Quantitative Analysis of Left Ventricular Dynamic and Shape Using 4D Echocardiography

周全勳(Chuan-Hsun Chou)

指導教授 ： 胡威志

中原大學/工學院/醫學工程研究所/碩士(2004年)

https://doi.org/10.6840/cycu200400711

摘要

本研究目的在於發展一套軟體分析工具，利用食道超音波影像重組後之四維網格模型（三維＋時間變化）觀察心內膜壁運動與心室運動形態。分析工具包含利用動態邊緣圈選 (Active Contour)所有時序、角度邊緣，並將空間點轉換後所有角度左心室邊緣組成三維立體模型，再將長軸與短軸等份分割構成含有時間的四維立體網格模型，藉由不同時序的參數變化來觀察左心室有問題區域。
由於超音波影像沒有標籤(tag)來做為空間上定位，本研究利用多時序、多角度的影像特性，藉由最小搜尋角度與最近距離的方法來追蹤心內膜壁可能的移動距離與方向，並且利用曲率值來觀察三維立體左心室形態上的變化。由於搜尋到的邊緣對於網格重組與參數計算有顯著的影響，因此更利用Mitral Annulus Line來代替影像之腔室瓣膜改善Snake演算法所造成圈選上的失誤，觀察Mitral Annulus Line 之定位可使邊緣更貼附於影像邊緣，提供更準確的容積運算。利用正確邊緣搜尋的優點，本研究提供左心室形態分析參數與心室壁運動參數分析，藉由牛眼圖（Bull’sEye）的方式來觀察，並將有問題區域以顏色貼圖於立體模型上，標示左心室運動異常之部位。
由統計XY軸在不同時序下移動距離分佈在-0.3cm~0.3cm，相較於整個心臟來說其空間變動很小所以可追蹤，再藉由角度差異性比較可得最佳追蹤角度為12度，利用最佳角度可將每個邊緣分割成30等分。因此其四維心室形態模型可視為連續之心室收縮模型並可追蹤其收縮動量後的位移。另外，外型上藉由五種數學幾何假體驗證了曲率計算對於外觀上變化的靈敏性，同時也得出心室舒張時的曲率值大於收縮時的曲率值，結果顯示了分析工具具有高度可行性及可靠性。

關鍵字

超音波 ； 追蹤 ； 空間解析度 ； 瓣膜線 ； 關鍵字:左心室 ； 動態圈選

並列摘要

The objective of this study was to provide methods to identify the area of abnormal movement at LV. The study has developed a software tool that was to analyze the shape of left ventricle (LV) and the motion of endocardium using reconstructed 4D LV (three-dimension + time changes) mesh model. The series of high spatial and temporal resolution transesophageal echocardiograph (TEE) image data was used in the reconstruction of LV model. The method of an active contour algorithm was employed to extract the endocardial boundaries of a complete cardiac cycle. The delineated endocardial contours were reconstructed into 3D mesh LV model in Cartesian coordinates. The local parameters which were extracted from segmented 3D model along the long-axis and the short-axis of LV. The difficulty of TEE images was tracing the movement of myocardium without tissue tagging. The proposed method of cardiac wall motion estimation was used the criteria of the minimum searching degree and the shortest distance of movement to track the dense cardiac displacement from 2D ultrasound images with high temporal resolution. The result showed that the majority of displacement was ±0.3 cm between two time frames. This result indicated that the displacement was relative small which meant that the motion registered in the displacement of reconstructed mesh could be considered as continuous motion. Furthermore, the majority of nodes (vertices of reconstructed mesh) at next time frame could be found within 12°of previous node. Therefore, tracing the motion of reconstructed mesh was possible. The result also indicated that the self-developed software and analyzing tools was practical and stable in analyzing LV motion and shape. With the excellence of LV border delineating, the thesis also reported the capability of calculating LV volume, parameters of LV shape analysis and wall motion analysis. The Bull’s Eye display method was used to illustrate the results of shape analysis and wall motion analysis.

並列關鍵字

3D+1D ； Active Contour ； Track ； LV Shape Analysis ； Keyword: Left Ventricular function ； LV wall motion analysis

參考文獻

[15] 李佳融 , ”半自動式左心室三維動態模型之運動和形態分析系統,” 中原大學 , July 2002

[20] S. Richard Wright, Jr. Michael Sweet, ”OpenGL超級手冊, ” 碁峯,2000.

[5] G. Meyer Francois, Todd Constable, Albert J. Sinusas and James S. Duncan, ”Tracking Myocardial Deformation Using Phase Contrast MR Velocity Fields: A Stochastic Approach,” IEEE Transactions on Medical Imaging, vol. 15, no. 4, August 1996

[6] J. C. McEachen, and J. S. Duncan, ”Shape-based Tracking of Left Ventricular Wall Motion,” IEEE Transactions in Medical Imaging,

[7] Maria J. Ledesma-Carbayo, Jan Kybic, Manuel Desco, Andres Santos, and Michael Unser, ”Cardiac Motion Analysis from Ultrasound Sequences Using Non-Rigid Registration, ” Proceedings of MICCAI, October, 2001.