臺灣博碩士論文加值系統

乳癌是近幾年來女性癌症攀升率第一位，也是女性十大癌症之一。因為現代人飲食西化與作息的不正常導致於乳房腺體的變異、良性囊腫的產生。一旦發生轉變後就可能進展至乳癌。
由於醫學影像的進步，乳房超音波掃瞄已成為篩檢乳癌的主要工具之一。有研究證明，如果能給予正確的乳癌種類與分期，則乳癌的治癒效果是更好的。因此近年來發展出以超音波影像特性為基礎的程式進行特徵分析預測乳癌進程程度，幫助醫生用有效的治療方式治療乳癌，降低乳癌的死亡率。
而本研究將量化乳房超音波影像，結合機器學習進行分類、預測模型來輔助診斷，期望能達到縮短診斷時間、提高診斷的一致性與準確度。
本研究回溯收集完全無病灶以及接受超音波導引穿刺粗針切片影像並將之帶入使用CNN auto-encoder所建立之模型，並與傳統醫生依據超音波影像所判定之BIRADS分期作統計比較，評估此模型對於醫生判定是否為乳癌惡性腫瘤有無幫助。
本研究所得之模型的AUC=0.802，相較於單就BIRADS其AUC=0.764有較好的判斷力。相信若能將模型結果讓醫生對照判斷，將能輔助醫生對於乳癌腫瘤的評估。

Breast cancer, one of the most common malignancies, is the second leading cause of cancer death in women. The chance that a woman will die from breast cancer is about 1 in 37 (about 2.7%).The change of modern people''s eating habit and work loading leads to the variation of breast glands and the development of benign cysts.
Due to the advantage of ultrasound in image, breast ultrasound has become one of the main tools for screen breast cancer. In addition to the detection of breast structure, glandular distribution, size and location of the lesion, ultrasound also has the ability to confirm axillary or clavicular lymph node metastases. If tests shows might have breast cancer may refer for a core needle biopsy.
Our study will quantify ultrasound images and combine machine learning with classification and prediction models to aid in diagnosis. The model obtained in our study AUC=0.802, which is better than BIRADS with AUC=0.764. We believe that the model will be able to assist doctors in the evaluation of breast cancer tumors. We hope to shorten the diagnosis time and improve diagnostic consistency and accuracy.

目錄
中文摘要 I
ABSTRACT II
表目錄 V
圖目錄 VI
第一章 緒論 1
1-1 超音波簡介 1
1-2 乳房外觀與構造 2
1-3 乳癌 4
1-4 乳房檢查的比較 9
1-5 乳癌的分期與報告 11
1-6 研究目的與目標 13
第二章 研究方法 15
2-1乳房超音波 15
2-2 收案與流程 18
2-3 腫瘤的圈選 19
2-4 CNN BASED AUTO-ENCODER 21
2-5 影像分析 24
2-6 統計分析 25
第三章 研究結果 26
3-1病患資料 26
3-2 各年齡層與腫瘤良惡性相關性 27
3-3 腫瘤面積與良惡性相關性 30
3-4 AUTO ENCODER各種統計 33
3-5 BIRADS分期各種統計 36
3-6 BIRADS與AUTO ENCODER比較 40
第四章 討論 42
第五章 結論 47
參考文獻 48

[1] Ka Sing Wong, Huan Li, Yu Leung Chan, Anil Ahuja, Wynnie W.M. Lam, Agatha Wong, Richard Kay. Use of Transcranial Doppler Ultrasound to Predict Outcome in Patients With Intracranial Large-Artery Occlusive Disease.Originally published November 1, (2000).
[2]W.Scott Burgin, Marc Malkoff, Robert A. Felberg, Andrew M. Demchuk, Ioannis Christou, James C. Grotta, Andrei V. Alexandrov. Transcranial Doppler Ultrasound Criteria for Recanalization After Thrombolysis for Middle Cerebral Artery Stroke. Originally published May 1, (2000).
[3]https://www.medindia.net/patients/patientinfo/breasts-structures-and-types-structure.htm(2018)
[4]http://www.radiologyassistant.nl/en/p53b4082c92130/bi-rads-for mammography-and-ultrasound-(2013).html
[5]國民健康署2017年健康年報
[6]衛生福利部2017年度癌症登記
[7]Chaoli X, Shuping Wei, Yingdong Xie, Xiaoxiang Guan, Ninghua Fu, Pengfei Huang and Bin Yang.Combined use of the automated breast volume scanner and the US elastography for the differentiation of benign from malignant lesions of the breast Xu et al. BMC Cancer( 2014), 14:798
[8]K. Kerlikowske , P.A. Carney , B. Geller , M.T. Mandelson , S.H. Taplin , K. Malvin , V. Ernster , N. Urban , G. Cutter , R. Rosenberg , R. Ballard-Barbash , Performance of screening mammography among women with and without a first-degree relative with breast cancer, Ann. Intern. Med. 133 (20 0 0) 855–863 .
[9]L. Berlin, Radiologic errors, past, present and future, Diagnosis (2014) 79–84, doi: 10.1515/dx- 2013- 0012 ..
[10]St.Gallen International Breast Cancer Conference Primary Therapy of Early Breast Cancer Evidence, Controversies, Consensus
[11]Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray, F. GLOBOCAN 2012 v1.1, Cancer Incidence and Mortality Worldwide: IARC Cancer Base No. 11 [Internet]. Lyon, France: International Agency for Research on Cancer; (2014).
[12]Chen-Pin Chou , Nan-Jing Peng, Tsung-Hsien Chang , Tsung-Lung Yang , Chin Hu ,Huey-Shyan Lin , Jer-Shyung Huang , Huay-Ben Pan:Clinical roles of breast 3T MRI, FDG PET/CT, and breast ultrasound for asymptomatic women with an abnormal screening mammogram. Int J Med Robot.( 2016) Sep 4. doi: 10.1002/rcs.1767.
[13]Chaoli Xu, Shuping Wei, Yingdong Xie, Xiaoxiang Guan, Ninghua Fu1, Pengfei Huang and Bin Yang:Combined use of the automated breast volume scanner and the US elastography for the differentiation of benign from malignant lesions of the breast. Xu et al. BMC Cancer (2014), 14:798
[14]Shwn-Huey Shieh, Vivian Chia-Rong Hsieh, Shu-Hui Liu , Chun-Ru Chien , Cheng-Chieh Lin , Trong-Neng Wu :Delayed time from first medical visit to diagnosis for breast cancer patients in Taiwan Journal of the Formosan Medical Association (2014) 113..
[15]J.S. Whang, S.R. Baker, R. Patel, L. Luk, A. Castro, The causes of medical mal- practice suits against radiologists in the United States, Radiology 266 (2013) 548–554, doi: 10.1148/radiol.12111119 .
[16]E.A. Sickles, Periodic mammographic follow-up of probably benign lesions: re- sults in 3,184 consecutive cases, Radiology 179 (1991) 463–46 8, doi: 10.114 8/ radiology.179.2.2014293.
[17]Elmore JG, Longton GM, Carney PA, Geller BM, Onega T, Tosteson AN, Nelson HD, Pepe MS, Allison KH, Schnitt SJ, O''Malley FP, Weaver DL. Diagnostic concordance among pathologists interpreting breast biopsy specimens. JAMA.(2015); 313(11):1122-32
[18]Hugo J. W. L. Aerts, Emmanuel Rios Velazquez, Ralph T. H. Leijenaar, Chintan Parmar, Patrick Grossmann, Sara Carvalho, et al.Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nature Communications , (2014),
[19]Y. Bengio, Learning deep architectures for AI, found, Trends??Mach. Learn 2 (2009) 1–127, doi: 10.1561/220000000.
[20]Y. Bengio, A. Courville, P. Vincent, Representation learning: a review and new perspectives, IEEE Trans. Pattern Anal. Mach. Intell. 35 (2013) 1798–1828, doi: 10.1109/TPAMI.2013.50.
[21]J. Schmidhuber, Deep learning in neural networks: an overview, Neural Netw. 61 (2015) 85–117, doi: 10.1016/j.neunet.2014.09.00.
[22]Y. LeCun, Y. Bengio, G. Hinton, Deep learning, Nature 521 (2015) 436–4 4 4, doi: 10.1038/nature14539.
[23]http://www.ultrasoundpaedia.com/normal-breast/(2018)
[24]C. Balleyguier, L. Ciolovana, S. Ammaria,S. Canalea, S. Sethoma, R. Al Rouhbanea,P.Vielh c,C.Dromaina:Breast elastography: The technical process and its applications. Diagnostic and Interventional Imaging (2013) 94, 503—513.
[25]Y. LeCun, K. Kavukcuoglu, C. Farabet, Convolutional networks and applications in vision, in: Proc. 2010 IEEE Int. Symp. Circuits Syst., (2010), pp. 253–256, doi: 10.1109/ISCAS.2010.5537907.
[26]H.C. Shin, H.R. Roth, M. Gao, L. Lu, Z. Xu, I. Nogues, J. Yao, D. Mollura, R.M. Summers, Deep convolutional neural networks for computer-aided de- tection: CNN architectures, dataset characteristics and transfer learning, IEEE Trans. Med. Imaging 35 (2016) 1285–1298, doi: 10.1109/TMI.2016.2528162 .
[27]G. Carneiro, J. Nascimento, A.P. Bradley, Unregistered multiview mammogram analysis with pre-trained deep learning models, in: Int. Conf. Med. Image Com- put. Comput.-Assist. Interv., Springer, (2015), pp. 652–660. http://link.springer. com/chapter/10.1007/978- 3- 319- 24574- 4 _ 78.