乳房動態磁振造影之生物標記分析

Abstract

根據美國癌症協會最新的統計，乳癌是導致女性癌症死亡的第二大主因，且預期在未來是女性中最常見的癌症。動態磁振造影以非電離輻射的方式產生高解析度的三維影像，被廣泛利用在乳癌第二線診斷評估及追蹤。臨床上，一旦確診為乳癌後，醫師將安排一系列的評估以製定合適的治療計畫。隨著乳癌治療方式的發展，愈來愈多針對特定生物標記的新型治療方法被提出並達到有效的治療效果。因此，在進行治療之前識別相關的生物標記將有助於醫師製定客製化的治療方式。相較於免疫組織化學法與脫氧核糖核酸測序，透過影像分析識別這些生物標記可達到非侵入式且整體性的分析。本論文提出在乳癌中識別荷爾蒙受體狀態與體細胞突變的影像分析方法，第一個主題提出一個量化乳房腫瘤異質性的方法，並且進一步識別動激素受體、人類表皮生長因子受體以及三陰性乳癌的狀態。第二個主題透過數值排序紋理分析與形態學分析識別乳癌中體細胞突變的鑑定。實驗結果顯示量化乳房腫瘤內部異質性可反映不同分子標記的血管複雜性，基於數值排序的乳房腫瘤紋理分析可有效識別TP53突變與PIK3CA突變。

According to the estimations of the American Cancer Society, breast cancer is the second leading cause of cancer death in women, which is expected as the most frequently diagnosed cancer (30%) of all new cancer diagnoses. Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) is one of the second-phase examinations, which adopts non-ionizing radiation to create high resolution three dimensional (3-D) images and achieves high sensitivity in breast cancer diagnosis. Once breast cancer is diagnosed in clinical, the healthcare provider will begin to develop a treatment plan under different circumstances. With the growth of development in breast cancer treatment, novel treatments were invented for effective response on specific biomarkers. Therefore, identifying the biomarkers which associated with therapeutic response before treatment is necessary in customizing treatment plans. Furthermore, identification of these therapeutic biomarkers by image analysis techniques provides a non-invasive and overall quantification. This study proposed methods of identifying hormone receptor (ER, HER2, and TNBC) statuses and somatic mutations (TP53 and PIK3CA) in breast cancer. The first sub-study proposed a technique of quantifying intra-tumoral heterogeneity, and further recognizing the ER, HER2, and TNBC statuses. The second sub-study recognized TP53 and PIK3CA mutations by texture analysis based on ranklet transform and morphological analysis. The results show that quantification of intra-tumoral heterogeneity can be used to reflect the vasculature complexity of different molecular markers and texture analysis of breast tumor based on ranklet transform is potential in recognizing the presence of TP53 and PIK3CA mutations.