自動乳房超音波之腫瘤診斷

Automated Whole Breast Ultrasound Tumor Diagnosis

江立人(Li-Ren Chiang)

指導教授 ： 張瑞峰

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2010年)

https://doi.org/10.6342/NTU.2010.00126

摘要

乳癌一直是女性癌症中的主要死因，但是只要早期檢測與醫治就能大幅提高其治癒率。近年來，由於電腦輔助診斷系統發展快速，已經不只能偵測腫瘤，還能對腫瘤進行分析診斷，因此，對於病患的切片檢查次數就能相對減低。近年來，為了提供臨床上快速檢查的乳房攝影工具，因此研發新的自動全乳房超音波機器。在這篇論文中，使用了自動超音波影像來進行實驗。首先，先對腫瘤區域以等位函數法進行自動腫瘤切割，接著，依據切割出來的腫瘤輪廓進行灰階值共生矩陣的紋理分析，傳統形狀資訊分析，以及建立與腫瘤有最小距離相合的橢球模型，並作此橢球與腫瘤的異同點分析，我們即依據此三類的特徵分析來做為診斷的依據。本實驗中使用了147個經過病理驗證的腫瘤，其中包括了76個良性病例以及71個惡性病例，重複以不同種類組合的特徵，以邏輯回歸分析的方式，在留一交互驗證的規約下作準確性的測試。從實驗結果來看，橢球特徵和傳統形狀特徵相結合有較高的準確率，可以達到85.03%(125/147)，敏感性達到84.51%(60/71)，特異性85.53%(65/76)，ROC曲線面積0.9466，總合上述結果，我們相信自動全乳房超音波不僅能用來做乳房腫瘤的檢查，還能用來做已偵測出的腫瘤診斷。

關鍵字

超音波 ； 自動全乳房超音波 ； 診斷 ； 分類 ； 形狀 ； 橢圓

並列摘要

In the past, breast cancer is the major cause of death for women among all kinds of cancer. But the curability of breast cancer can be greatly improved if a proper treatment is adopted after an early detection. In recent years, the computer-aided diagnosis systems have been developed rapidly and they can not only detect the tumors but also differentiate malignant tumors from benign ones. Hence the demand of the breast biopsy of the detected tumors might be further reduced. Recently, the new automated whole breast ultrasound (ABUS) machines have also been developed in order to provide a fast screening tool as the routine clinical used mammography. In this paper, the ABUS images are used for the diagnosis of tumors. At first, the three-dimensional (3-D) tumor contour is segmented by using the automated level-set segmentation method. Then, the features including the texture information based on co-occurrence matrix, shape information, and ellipsoid fitting information are extracted based on the segmented 3-D tumor contour to classify the benign and malignant tumors. In the experiment, there are 147 pathologyproven cases, including 76 benign tumors and 71 malignant ones, are used to test the diagnosis performance of the logistic regression model with a leave-one-out cross validation based on the proposed features. From the experiment results, it is found that ellipsoid fitting features combined with traditional shape features can achieve a better performance with accuracy 85.03% (125/147), the sensitivity 84.51% (60/71), specificity 85.53% (65/76), and the area under the ROC curve Az 0.9466. Hence, the ABUS images could be used not only for screening the breast cancers but also diagnosing the detected tumors.

並列關鍵字

Ultrasound ； ABUS ； diagnosis ； classification ； shape ； ellipsoid

參考文獻

[1] W. A. Berg, J. D. Blume, J. B. Cormack, E. B. Mendelson, D. Lehrer, M. Bohm-Velez, E. D. Pisano, R. A. Jong, W. P. Evans, M. J. Morton, M. C. Mahoney, L. H. Larsen, R. G. Barr, D. M. Farria, H. S. Marques, and K. Boparai, "Combined screening with ultrasound and mammography vs mammography alone in women at elevated risk of breast cancer," JAMA, vol. 299, pp. 2151-63, May 14 2008.

[2] M. Nothacker, V. Duda, M. Hahn, M. Warm, F. Degenhardt, H. Madjar, S. Weinbrenner, and U. S. Albert, "Early detection of breast cancer: Benefits and risks of supplemental breast ultrasound in asymptomatic women with mammographically dense breast tissue. A systematic review," BMC Cancer, vol. 9, p. 335, 2009.

[3] P. B. Gordon and S. L. Goldenberg, "Malignant breast masses detected only by ultrasound. A retrospective review," Cancer, vol. 76, pp. 626-30, Aug 15 1995.

[4] T. M. Kolb, J. Lichy, and J. H. Newhouse, "Occult cancer in women with dense breasts: Detection with screening us--diagnostic yield and tumor characteristics," Radiology, vol. 207, pp. 191-9, Apr 1998.

[5] A. Izumori, K. Takebe, and A. Sato, "Ultrasound findings and histological features of ductal carcinoma in situ detected by ultrasound examination alone," Breast Cancer, vol. 17, pp. 136-41, Apr 2010.