乳房X光攝影腫瘤診斷量化報告系統研究

摘要：在台灣，乳癌已成為第二大類型的婦女癌變的疾病。發病率和死亡率都在增加，乳房X光檢查仍然是主要的篩選技術，能夠在早期階段檢測乳癌。基於乳房成像的標準方法，乳房成像報告和報告數據系統（BI-RADS）由美國放射學院提出，“腫瘤”被定義為一個佔位性病變可見於兩種不同的投影。腫瘤的解釋是基於三個方面。包含形狀，邊緣，和密度。在本文中，我們重點放在腫瘤關於形狀的電腦化解釋。腫瘤的形狀分解之一分為四個組成部分的形狀認知（SC）方法或用形態學方法的三個組成部分。基於這些組成部分，通過使用支持向量機方法，腫瘤被列為不規則，小葉橢圓形和圓形，這四個形狀分類是為了降低患癌症的風險。四個形狀中的不規則，小葉，橢圓形和圓形，可確定在訓練樣本與校正率達到100％。基於這個分類器系統上，41個臨床樣本被確定正確的識別率為不規則100％，小葉100％，橢圓100％，圓0％。平均正確識別率達到88.46％。

關鍵字

乳房X光攝影；腫瘤；形狀；電腦輔助診斷

並列摘要

In Taiwan, breast cancer has become the second leading type of cancerous disease among women. The rates of incidence and mortality are increasing, and mammography continues to be the main screening technique capable of detecting breast cancer at an early stage. Based on the standardized method for breast imaging, reporting of breast imaging and reporting data system (BI-RADS) proposed by the American College of Radiology, a “mass” is defined as a space-occupying lesion seen in two different projections. Interpretation of masses is based on three aspects. These are shape, margin, and density. In this paper, we focus on the computerized interpretation of masses with regards to shape. The shape of masses are decomposed either into four components using a shape cognitron (SC) method or three components using a morphological method. Based on these components and through the use of the supported vector machine method, masses are classified as irregular, lobular, oval, and round where these four shape classifications are in the order of decreasing risk of cancer. The four shapes of irregular, lobular, oval, and round can be identified in training samples with the correction rates of 100%. Based on this classifier system, 46 clinical samples were identified with a correct identification rate of 100%, 100%, 100%, and 0%. The average correct identification rate achieved being 88.46%.

並列關鍵字

mammography ； mass ； shape ； computer aided diagnosis

參考文獻

[1]Blanks, R.G. Wallis, M.G. & Moss, S. M. , 1998, “A comparison of cancer detection rates achieved by breast cancer screening programmes by number of readers, for one and two view mammography: Results from the UK National Health Service breast screening programme”, J. Med. Screening, 5(4):195–201.

[3]ACR BI-RADS Breast Imaging and Reporting Data System V4，2003，ACR.

[4]Akay, M.F., 2009, “Support vector machines combined with feature selection for breast cancer diagnosis”, Expert System With Applications, 36(2),3240–3247.

[5]Lo, C.S. & Wang, C.M. , 2012, “Support vector machine for breast MR image classification”, Computers and Mathematics with Applications, 64:1153–1162.

[8]Bettyann Holtzmann Kevles, 1997, Naked to the bone : medical imaging in the twentieth century, Rutgers University Press.

國際替代計量

乳房X光攝影腫瘤診斷量化報告系統研究

未授權

主題瀏覽