3-D能量都卜勒乳房超音波影像的血管化特徵電腦輔助診斷之研究

Tunghai University Institutional Repository > 工學院 > 資訊工程學系所 > 碩士論文 > Item 310901/2667

Please use this identifier to cite or link to this item: http://140.128.103.80:8080/handle/310901/2667

Title:	3-D能量都卜勒乳房超音波影像的血管化特徵電腦輔助診斷之研究
Other Titles:	Computer-aided Diagnosis for Breast Tumors by Using Vasculization of 3-D Power Doppler Ultrasound
Authors:	許嘉家 Hsu, Chia-Chia
Contributors:	黃育仁 Huang, Yu-Len 東海大學資訊工程學系
Keywords:	3-D能量都卜勒超音波影像;腫瘤血管化;乳房腫瘤;電腦輔助診斷系統;支援向量機;等位函數法 three-dimensional power Doppler;tumor vascularity;volume of interest;computer-aided diagnosis;support vector machine;level set method
Date:	2008
Issue Date:	2011-03-23T05:56:38Z (UTC)
Abstract:	血管化是腫瘤產生、生長與擴散時公認的基本程序，腫瘤內血管化的程度與腫瘤的良惡性有絕對的相關性，許多已發表的研究使用彩色都卜勒超音波來估算乳房腫瘤的血管化特徵，結合彩色都卜勒超音波與傳統的灰階超音波影像能夠有效地定位腫瘤血管，並區分乳房腫瘤的良惡性，然而，彩色都卜勒超音波難以偵測到與主軸方向垂直的血管。近年來的研究指出，能量都卜勒超音波對血流偵測之敏感度較彩色都卜勒超音波高，且能量都卜勒超音波在血管角度、深度的影響與血管尺度限制較小，因此能量都卜勒超音波用於偵測乳房腫瘤的血管化特徵以及分類乳房腫瘤之良惡性上，都優於彩色都卜勒超音波。三維的能量都卜勒超音波影像較二維的超音波影像包含了更多的血流資訊，能夠偵測整個乳房腫瘤的血管化，且惡性腫瘤的體積通常較大，三維灰階超音波影像能夠估算出非常近似於真實乳房腫瘤的體積，可作為區分乳房腫瘤良惡性之有用的特徵。此外，一份統計的資料指出乳癌與病人的年齡有關，通常40歲以上的婦女得到乳癌的機率較高，因此，本論文提出一乳癌三維能量都卜勒超音波影像血流特徵電腦輔助診斷系統，利用從三維能量都卜勒超音波影像中所取得的腫瘤血流特徵，再加入腫瘤體積與病人年齡，來區分乳房腫瘤之良惡性。首先，請醫師針對三維灰階超音波影像上的乳房腫瘤繪出有興趣的區域(Volume of Interest, VOI)，依照此區域計算近似於真實腫瘤大小的體積，接著估算三維能量都卜勒超音波中的血流特徵，再加入病人的年齡形成特徵向量，最後使用支援向量機(Support Vector Machine, SVM)來分類乳房腫瘤。支援向量機的學習速度快且較穩定，能夠減少訓練的時間並得到良好的分類結果，大大的提升了診斷速度與正確性。實驗結果證明，本論文利用支援向量機，結合腫瘤血流、體積與病人年齡等特徵，來分類乳房腫瘤的良惡性，得到令人滿意的準確度。此外，本論文也使用等位函數法(Level set method)自動切割乳房腫瘤，使用自動切割所得到的輪廓區域去估算以上所提的特徵，並與原始方法比較診斷之結果。 Rationale and Objectives: To evaluate the accuracy of three-dimensional (3-D) power Doppler ultrasound in differentiating benign and malignant breast tumors by using support vector machine (SVM).Materials and Methods: 3-D power Doppler ultrasonography were obtained from 164 patients (age range, 17-80 years; mean age, 44 years) with 86 benign and 78 malignant breast tumors. The volume of interest (VOI) in 3-D ultrasound images was automatic generated by using three rectangular region of interest. The vascularization index (VI), flow index (FI) and vascularization-flow index (VFI) on 3-D power-Doppler ultrasound images were evaluated for the entire volume area, the computer extracted VOI area and the opposite area of VOI. Patients’ age and the volume of VOI were also applied to identify breast tumors. Besides, this study also proposed an automatic contouring method by using level set method (LSM) to extract the similar 3-D volume and improve the diagnostic accuracy. The level set segmentation automatically extracted 3-D contours of breast tumors from ultrasound images after the image preprocessing. In this study, each ultrasonography classified as benign or malignant based on 11 features by using the SVM model. In the experiment, all the breast tumors were sampled with k-fold cross-validation (k = 10) to evaluate the performance with receiver operating characteristic (ROC) curve.Results: The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of SVM for classifying malignancies were 94%, 67%, 72%, 92% and 80%, respectively. The classification performance in terms of Az value for the ROC curve of the features derived from 3-D power Doppler is 0.91. Comparing to this CAD system, the performance of the CAD system with automatic contouring method is better.Conclusions: The vascular features of 3-D power Doppler ultrasound combined with patients’ age and volume of VOI are effective in differentiating benign and malignant breast tumors by using SVM.
Appears in Collections:	[資訊工程學系所] 碩士論文

Files in This Item:

File	Size	Format
096THU00394013-001.pdf	747Kb	Adobe PDF	750	View/Open

Loading...