體內不連貫運動高光譜成像技術對乳腺腫瘤的檢測和分類

Abstract

乳腺癌是全球女性疾病和死亡的主要原因。由於傳統的乳腺X線攝影和超聲檢查的局限性，在過去的幾十年中，磁共振成像（MRI）已逐漸成為乳腺癌評估的一種重要的放射學方法。MRI不會出現與放射線暴露相關的問題，並具有出色的圖像分辨率和對比度。但是，缺點是需要注射顯影劑，該顯影劑對某些患者（例如患有慢性腎臟疾病的患者或孕婦和哺乳期婦女）有毒。大腦中釓沉積物的最新發現也令人關注。為了解決這些問題，本文開發了一種基於體素不相干運動（IVIM-）MRI的直方圖分析方法，該方法利用波段擴展過程（BEP）等幾種高光譜技術將多光譜圖像擴展為高光譜圖像並創建一個自動目標生成過程（ATGP）。在自動找到可疑目標後，使用內核約束能量最小化（KCEM）進行了進一步檢測。決策樹和直方圖分析透過對檢測到的病變進行定量分析來對乳腺組織進行分類，用於區分乳腺組織的三類：惡性腫瘤（即中央和周圍區域），囊腫和正常乳腺組織。實驗結果表明，基於IVIM-MRI的直方圖分析方法可以有效地區分這三種乳腺組織。

Breast cancer is a main cause of disease and death for women globally. Due to limitation of traditional mammography and ultrasonography, magnetic resonance imaging (MRI) has gradually become an important radiological method for breast cancer assessment over the past decades. MRI is free of problems related to radiation exposure and provides excellent image resolution and contrast. However, one disadvantage is the injection of contrast agent, which is toxic for some patients (such as patients with chronic renal disease or pregnant and lactating women). Recent findings of gadolinium deposits in the brain are also a concern. To address these issues, this dissertationdevelops an intravoxel incoherent motion- (IVIM-) MRI-based histogram analysis approach, which takes advantage of several hyperspectral techniques, such as band expansion process (BEP)which expands multispectral images to hyperspectral images and automatic target generation process (ATGP) which generates desired targets to be used for follow-upsupervised detection and classification. After automatically finding suspected targets by ATGP, target detection was further performed by kernel constrained energy minimization (KCEM). A decision tree and histogram analysis are then applied to classifying breast tissue via quantitative analysis for detected lesions, which can be classified into three categories of breast tissue: malignant tumors (i.e., central and peripheral zone), cysts, and normal breast tissues. The experimental results demonstrate that the proposed IVIM-MRI-based histogram analysis approach can effectively differentiate amongthese three breast tissue types.