利用電腦斷層影像紋理特徵辨識胰臟癌區域

Pancreatic Cancer Detection by Patch-Based Computed Tomography Radiomics

顏惠萱(Hui-Hsuan Yen)

指導教授 ： 王偉仲

共同指導教授 ： 陳素雲

國立臺灣大學/電機資訊學院/資料科學學位學程/碩士(2019年)

https://doi.org/10.6342/NTU201900731

摘要

胰腺癌(PC)是致命率極高的癌症，也是美國癌症死亡的第四大原因。 影像組學是一種從醫學圖像中提取定量統計和特徵以解碼組織表型的方法。 本研究的目的是開發一種機器學習模型，使用放射學特徵在有打顯影劑的電腦斷層影像(CT)上區分PC與正常胰臟，並且找出重要的影像組學特徵。對於感興趣的區域(ROI)，我們會對幾個重疊的小區塊進行採樣。每個小區塊提取總共91個影像組學特徵並且以機器學習模型訓練而進行分類，最後我們選出重要的11個影像組學特徵。我們的模型可以利用這11個影像組學特徵，準確地檢測胰腺癌，為一種潛在的計算輔助診斷工具。

關鍵字

胰腺癌 ； 影像組學 ； 電腦斷層影像 ； 區塊取樣 ； 機器學習

並列摘要

Pancreatic cancer (PC) is the most lethal cancer and the fourth leading cause of cancer deaths in the United States. Radiomics is a methodology that extracts quantitative statistics and features from medical images to decode the phenotype of tissues. The purpose of this study is to develop a machine learning model to differentiate PC from healthy pancreas on contrast-enhanced computed tomography (CT) using radiomic features and then investigate the important features. With a region in interest (ROI), we sample several overlapping patches. A total of 91 radiomic features were extracted of each patch and subject to a machine learning model to perform classification. We select 11 important features at last. Our model can accurately detect PC by using these 11 important features and is a potential computer-aided diagnosis tool.

並列關鍵字

pancreatic cancer ； radiomics ； computed tomography ； patch sampling ； machine learning

參考文獻

[1] B. Agarwal, A. M. Correa, and L. Ho. Survival in pancreatic carcinoma based on tumor size. Pancreas, 36(1):e15–20, Jan. 2008.

[2] K. Armanious, C. Jiang, M. Fischer, T. K¨ustner, K. Nikolaou, S. Gatidis, and B. Yang. MedGAN: Medical image translation using GANs. June 2018.

[3] M. A. Attiyeh, J. Chakraborty, A. Doussot, L. Langdon-Embry, S. Mainarich, M. G¨onen, V. P. Balachandran, M. I. D’Angelica, R. P. DeMatteo, W. R. Jarnagin, T. P. Kingham, P. J. Allen, A. L. Simpson, and R. K. Do. Survival prediction in pancreatic ductal adenocarcinoma by quantitative computed tomography image analysis. Ann. Surg. Oncol., 25(4):1034–1042, Apr. 2018.

[4] M. A. Attiyeh, J. Chakraborty, C. A.McIntyre, R. Kappagantula, Y. Chou, G. Askan, K. Seier, M. Gonen, O. Basturk, V. P. Balachandran, T. P. Kingham, M. I. D’Angelica, J. A. Drebin, W. R. Jarnagin, P. J. Allen, C. A. Iacobuzio-Donahue, A. L. Simpson, and R. K. Do. CT radiomics associations with genotype and stromal content in pancreatic ductal adenocarcinoma. Abdom Radiol (NY), June 2019.

[5] M. A. Attiyeh, J. Chakraborty, C. A.McIntyre, R. Kappagantula, Y. Chou, G. Askan, K. Seier, M. Gonen, O. Basturk, V. P. Balachandran, T. P. Kingham, M. I. D’Angelica, J. A. Drebin, W. R. Jarnagin, P. J. Allen, C. A. Iacobuzio-Donahue, A. L. Simpson, and R. K. Do. CT radiomics associations with genotype and stromal content in pancreatic ductal adenocarcinoma. Abdom Radiol (NY), June 2019.