根據衛福部統計資料，癌症已連續35年蟬聯國人十大死因之首，其中以肺癌的死亡率最高。肺癌中又以肺腺癌為大宗，且好發於女性患者身上。鑑此，如何在早期篩檢的資料中，發現癌症特徵並進行治療，是對抗肺腺癌的不二法門。本研究與台大和榮總醫院合作取得病患的肺部低劑量電腦斷層影像(Low Dose CT)，透過3D影像重建軟體將2D影像重建成3D模型，並觀察與比較肺腺癌患者腫瘤3D重建後的各種外觀與質地特徵，以期找出能分辨早期惡性與良性腫瘤之關鍵特徵。我們研究指出除了腫瘤體積之外，腫瘤密度變化也與惡性程度高度相關。因此本論文設計了一測量腫瘤密度變化之新穎特徵，稱為3D腫瘤消失速率(Tumor Disappearance Rate)，能更準確區分良性與惡性腫瘤。 According to Ministry of Health and Welfare in Taiwan, Cancer remained the first rank of ten major death caused reasons in the last 35 years which Lung cancer has the highest mortality rate. Among all lung cancers, adenocarcinoma accounts for the majority of them and especially common in female patients. As a result, it is very important to discover the feature of adenocarcinoma and accept appropriate treatment in early stage. The purpose of this research is focused on discovering novel features via 3D reconstruction from 2D LDCT images. Our study indicated that, in addition to volume, the density within the tumor is also highly associated with malignancy of the tumor. This thesis proposed a novel feature called 3D Tumor Disappearance Rate (3D-TDR), which is able to accurately distinguish malignancy from benign tumors.