心肌的細胞外容積比(ECV)是計算在單位體積中細胞外體積占心肌全部體積的比例，在臨床上以ECV可用來評估心肌纖維化等許多心臟相關疾病的情況。本論文將提出一套以半自動化的方式，使用心臟CT影像計算左心室ECV的空間分佈圖(ECV mapping)，並且以立體曲面的方式顯示整個左心室的心肌ECV分佈情況。

本研究所採用的影像處理包含影像對位、影像分割、ECV 計算、以及ECV視覺化表示等四個主要步驟。影像對位是以一組在注射顯影劑後的angiographic 影像來當作對位中的參考影像，分別將注射顯影劑前和注射顯影劑後達到濃度平衡的影像對位到此組參考影像。而影像分割則是使用半自動演算法分割出左心室心肌和血液的區域，將其應用於ECV公式的計算。上述的影像對位和分割的步驟完成之後即可將結果帶入ECV公式得到ECV mapping的結果，最後以立體空間中的曲面表示左心室的ECV分佈。

在實驗結果中會展示四位不同受試者影像的對位、分割以及ECV mapping結果，並於討論影像分割分步驟中不同參數的結果比較，以及手動選取步驟對分割結果穩定度的影響。結果中顯示每位受試者影像都成功計算出左心室的心肌ECV ，但是由於對位不夠準確導致在多數受試者中，心肌某些區域ECV數值會出現異常偏高和偏低的情況，如能提高影像對位的正確性對此ECV mapping演算法的臨床參考價值將會很有幫助。

Extracellular volume (ECV) fraction is the volumetric ratio of extracellular space to the myocardium volume. ECV fraction has been used in the evaluation of many myocardiopathies diseases, such as myocardial fibrosis. In this thesis, a semi-automated method is proposed to obtain the ECV mapping of left ventricle by using cardiac computed tomography. A 3D demonstration of ECV on the ventricular surface is also shown for better visualization.

The image processing method of this study is consisted of four major parts, including image registration, tissue segmentation, ECV calculation and 3D visualization. In image registration, the non-contrast and late-phase enhanced equilibrium images would be registered to a high spatial-resolution and high contrast data set which is collected immediately after the administration of contrast agent, called angiographic images. Meanwhile, a semi-automated segmentation is applied to delineate the region of left ventricular myocardium and blood pool in the chamber. The signal change (ΔHU) of myocardium and blood is calculated to obtain the ECV mapping and show the ECV fraction of the whole left ventricle on the 3D surface.

The results of four subjects, including three normal and one patient with hypertrophic cardiomyopathy, are shown. The results of segmentation using different parameters are compared and discussed. In addition, the robustness of the semi-segmentation is examined by randomly selecting the seed point in repetitive experiments. It is shown that the ECV mapping is successfully done on all four subjects. However, the ECV value at the edge of ventricle may be over or under the normal range due to insufficient accuracy of registration. It is believed that the pixel-by-pixel ECV mapping would be of great clinical value if the accuracy of image registration could be improved in the near future.

論文審定書 i
致謝 ii
摘要 iii
Abstract iv
目錄 vi
圖目錄 viii
表目錄 x
第一章 簡介 1
1.1心肌細胞外容積比的介紹與臨床應用 1
1.2研究目的 5
第二章 研究方法 6
2.1 ECV公式介紹 6
2.2研究方法和流程 8
2.3影像對位 9
2.3.1 Rigid和Affine 10
2.3.2 Demons’ algorithm 11
2.3.3測試使用不同對位模型得到的對位結果 13
2.4影像分割 17
2.4.1分割左心室的血液區域 17
2.4.2分割左心室的心肌區域 21
2.5 ECV視覺化顯示 25
第三章 結果與分析 29
3.1影像資料 29
3.2影像對位結果 32
3.3影像分割結果 37
3.4 ECV mapping結果 40
3.5 ECV立體視覺化顯示 42
第四章 討論與結論 46
4.1心肌影像分割中layer growing方法的限制 46
4.2對心肌分割結果做更精確的分支去除 49
4.3使用不同種子點的影像分割結果比較 51
4.4 ECV mapping出現數值明顯偏高或偏低的區域 53
4.5結論 55
參考文獻 56

[1] C. Y. Ho, S. A. Abbasi, T. G. Neilan, R. V. Shah, Y. Chen, B. Heydari, et al., "T1 measurements identify extracellular volume expansion in hypertrophic cardiomyopathy sarcomere mutation carriers with and without left ventricular hypertrophy," Circulation: Cardiovascular Imaging, vol. 6, pp. 415-422, 2013.
[2] P. Kellman, J. R. Wilson, H. Xue, W. P. Bandettini, S. M. Shanbhag, K. M. Druey, et al., "Extracellular volume fraction mapping in the myocardium, part 2: initial clinical experience," Journal of Cardiovascular Magnetic Resonance, vol. 14, p. 64, 2012.
[3] S. K. White, D. M. Sado, M. Fontana, S. M. Banypersad, V. Maestrini, A. S. Flett, et al., "T1 mapping for myocardial extracellular volume measurement by CMR: bolus only versus primed infusion technique," JACC: Cardiovascular Imaging, vol. 6, pp. 955-962, 2013.
[4] M. S. Nacif, N. Kawel, J. J. Lee, X. Chen, J. Yao, A. Zavodni, et al., "Interstitial myocardial fibrosis assessed as extracellular volume fraction with low-radiation-dose cardiac CT," Radiology, vol. 264, pp. 876-883, 2012.
[5] M. S. Nacif, Y. Liu, J. Yao, S. Liu, C. T. Sibley, R. M. Summers, et al., "3D left ventricular extracellular volume fraction by low-radiation dose cardiac CT: assessment of interstitial myocardial fibrosis," Journal of cardiovascular computed tomography, vol. 7, pp. 51-57, 2013.
[6] T. A. Treibel, S. Bandula, M. Fontana, S. K. White, J. A. Gilbertson, A. S. Herrey, et al., "Extracellular volume quantification by dynamic equilibrium cardiac computed tomography in cardiac amyloidosis," Journal of cardiovascular computed tomography, vol. 9, pp. 585-592, 2015.
[7] A. S. Flett, M. P. Hayward, M. T. Ashworth, M. S. Hansen, A. M. Taylor, P. M. Elliott, et al., "Equilibrium contrast cardiovascular magnetic resonance for the measurement of diffuse myocardial fibrosis," Circulation, vol. 122, pp. 138-144, 2010.
[8] S. Bandula, S. K. White, A. S. Flett, D. Lawrence, F. Pugliese, M. T. Ashworth, et al., "Measurement of myocardial extracellular volume fraction by using equilibrium contrast-enhanced CT: validation against histologic findings," Radiology, vol. 269, pp. 396-403, 2013.
[9] X. Chen, M. S. Nacif, S. Liu, C. Sibley, R. M. Summers, D. A. Bluemke, et al., "A framework of whole heart extracellular volume fraction estimation for low-dose cardiac CT images," IEEE Transactions on Information Technology in Biomedicine, vol. 16, pp. 842-851, 2012.
[10] J.-P. Thirion, "Image matching as a diffusion process: an analogy with Maxwell's demons," Medical image analysis, vol. 2, pp. 243-260, 1998.
[11] P. Cachier, X. Pennec, and N. Ayache, "Fast non rigid matching by gradient descent: Study and improvements of the" demons" algorithm," INRIA, 1999.
[12] Insight Segmentation and Registration Toolkit. https://itk.org/
[13] MATITK Usage Guide. http://matitk.cs.sfu.ca/usageguide