跳到主要內容

臺灣博碩士論文加值系統

(35.172.136.29) 您好!臺灣時間:2021/07/29 08:44
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:林淳暐
研究生(外文):CHUN-WEI LIN
論文名稱:利用AI輔助判讀乳癌雙色雜交染色病理影像
論文名稱(外文):Use AI to assist interpretation Dual in situ hybridization staining pathological image of breast cancer
指導教授:趙載光趙載光引用關係
指導教授(外文):Tai-Kuang Chao
口試委員:高鴻偉王靖維
口試委員(外文):Hong-Wei GaoChing-Wei Wang
口試日期:2021-05-21
學位類別:碩士
校院名稱:國防醫學院
系所名稱:病理及寄生蟲學研究所
學門:醫藥衛生學門
學類:醫學學類
論文種類:學術論文
論文出版年:2021
畢業學年度:109
語文別:中文
論文頁數:81
中文關鍵詞:雙色原位雜交染色乳癌
外文關鍵詞:Dual in situ hybridizationbreast cancer
相關次數:
  • 被引用被引用:0
  • 點閱點閱:22
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
乳癌 (BC) 是女性常見的癌症之一,罹患乳癌的比例逐年攀升,如何有效的診斷及治療這是非常重要的。
本實驗利用收集到的乳癌病理切片,透過AI人工智慧加以分析可以有效提高病理診斷的正確性和預測預後,並可藉由HER2基因Dual in situ hybridization (DISH)雙色原位雜交的影像分析結果,可準確評估臨床病患HER2基因表達模擬兩可的病患是否確有HER2基因的過度表達和擴增,以利臨床準確評估是否可做抗HER2基因的標靶治療。
病理學家藉由在顯微鏡下研究染色後組織切片的方式來診斷癌症。在決定如何治療癌症時,得依靠病理學家的診斷,其中牽涉到極為主觀的意見,過程曠日費時。病理診斷中常會根據不同的顯微鏡下型態給予不同的病理診斷分類,但腫瘤內常出現不同分化的癌症,不過在病理學家的報告裡通常只會特別注意到最顯著的那一種。
但比例最高的腫瘤通常是屬於預後較好的類型,但占比較低的腫瘤分型預後較差,且常被忽略其重要性。因此,癌症療癒後又再次復發,若醫師能得知癌症可能會復發,便可提早使用化療或是標靶療法積極進行治療。但是傳統的顯微鏡觀察後給予不同分化的腫瘤評斷其比例,常不夠客觀且因人而異。若能藉由AI人工智慧加以判斷並分析其結果,可提供較為客觀的量化數據。
將染色後的乳癌病理組織切片影像利用高解析度相機拍照後,將數位化圖檔上傳至台灣科技大學王靖維教授團隊開發的 AIExplore 平台上,標記乳癌腫瘤位置後比較一般學習法及交叉驗證法的差異並利用 Cascade R-CNN 架構及弱監督學習法,訓練AI進行判讀得出快速客觀的臨床診斷。
透過人工智慧訓練學習之後,AI 能自主判讀出 DISH 病理影像,並且判斷出每個 DISH 染色切片的狀況且是否達到用藥的標準。而根據本實驗的結果得出一般學習法的準確度、精確度、召回率、F1分數、曲線下面積分別為0.6060、0.7492、0.6060、0.6576、0.6608,交叉驗證法得出的準確度、精確度、召回率、F1分數、曲線下面積分別為0.7441、0.6801、0.7441、0.6918、0.6800,由此可知交叉驗證法是比較好的方法,而後加入弱監督學習法之後比較有無弱監督學習法的差異,有加入的數據為0.73312,無加入的數據為0.60882,我們可以得知有加入弱監督學習法的數據比較高,所以能判斷此系統對於判讀乳癌的DISH染色片效果較好而且是有潛力的。

Breast cancer (BC) is one of the common cancers in women, and the proportion of breast cancer is increasing year by year. How to effectively diagnose and treat this is very important.
This experiment uses the collected pathological slices of breast cancer and analyzes them through AI artificial intelligence, which can effectively improve the accuracy of pathological diagnosis and predict prognosis. It can also use the imaging analysis results of the HER2 gene Dual in situ hybridization (DISH). It can accurately assess whether the clinical patient's HER2 gene expression simulation is ambiguous and whether the patient does have the overexpression and amplification of the HER2 gene, so as to facilitate the clinical accurate assessment of whether it can be targeted for anti-HER2 gene therapy.
Pathologists diagnose cancer by studying stained tissue sections under a microscope. When deciding how to treat cancer, one has to rely on the diagnosis of a pathologist, which involves extremely subjective opinions, and the process is time-consuming and time-consuming. In pathological diagnosis, different types of pathological diagnosis are often given according to different types of microscopy, but differently differentiated cancers often appear in tumors, but in the pathologist's report, only the most prominent one is usually paid special attention.
However, the tumors with the highest proportion are usually the types with better prognosis, but the lower tumor types have poor prognosis, and their importance is often ignored. Therefore, if the cancer recurs after treatment, if the doctor knows that the cancer may recur, he can use chemotherapy or targeted therapy to actively treat it early. However, the ratio of differently differentiated tumors after traditional microscopic observation is often not objective enough and varies from person to person. If AI artificial intelligence can be used to judge and analyze the results, it can provide more objective quantitative data.
After the stained breast cancer pathological tissue slice image is taken with a high-resolution camera, the digital image file is uploaded to the AIExplore platform developed by the team of Professor Jingwei Wang from Taiwan University of Science and Technology. After marking the position of the breast cancer tumor, it is compared with the general learning method and the cross-validation method. Differences and using the Cascade R-CNN architecture and weakly-supervised learning method to train AI to perform interpretation and obtain fast and objective clinical diagnosis.
After learning through artificial intelligence training, AI can autonomously read out DISH pathological images, and determine the condition of each DISH stained section and whether it meets the medication standards. According to the results of this experiment, the accuracy, precision, recall rate, F1 score, and area under the curve of the general learning method are 0.6060, 0.7492, 0.6060, 0.6576, and 0.6608, respectively. The accuracy and precision obtained by the cross-validation method , Recall rate, F1 score, and area under the curve are 0.7441, 0.6801, 0.7441, 0.6918, 0.6800, respectively. It can be seen that the cross-validation method is a better method. After adding the weakly-supervised learning method, there is a difference between whether or not the weakly-supervised learning method is added. The data with added is 0.73312, and the data without added is 0.60882. We can know that the data with weakly supervised learning method is relatively high, so it can be judged that this system has a better effect on the interpretation of DISH stains for breast cancer and has potential.

目 錄
致 謝 I
中 文 摘 要 V
Abstract VIII
縮 寫 表 XII
第一章、 緒 論 1
第一節、乳房基本構造 2
第二節、乳癌之危險因子 4
第三節、乳癌的症狀與分期 7
第四節、TNM分期 9
第五節、乳癌病理學特徵 12
第六節、臨床診斷及治療 18
第七節、AI人工智慧 26
第八節、混淆矩陣與Cascade R-CNN架構 28
第九節、Cross validation及弱監督學習法 31
第二章、 研 究 動 機 34
第一節、 研究目的 34
第二節、 實驗設計 36
第三節、實驗架構 37
第三章、 材 料 與 方 法 38
第一節、檢體收集 39
第二節、石蠟組織切片 41
第三節、雙色原位雜交染色 42
第四章、實 驗 結 果 45
第五章、討 論 47
第六章、結 論 50
第七章、參 考 文 獻 51








圖 目 錄
圖 一、乳房構造 64
圖 二、乳癌分期與五年存活率 65
圖 三、TNM分期 66
圖 四、乳癌組織學分類 67
圖 五、乳癌的亞型 68
圖 六、HER2狀態指南 69
圖 七、HER2免疫染色的半定量評分系統 69
圖 八、乳房X光攝影 70
圖 九、混淆矩陣 71
圖 十、K折交叉驗證 72
圖 十一、AI判讀病理影像 73
圖 十二、DISH染色原理 74
圖 十三、FISH/DISH染色 75
圖 十四、圈選標記結果 76
圖 十五、AI判讀結果 77
圖 十六、圖片像素大小 78
圖 十七、不重疊補丁 79

表 目 錄
表 一、一般學習法與交叉驗證法比較 80
表 二、有無弱監督學習法差異 81

















Veronesi, U., et al., Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer. New England Journal of Medicine, 2002. 347(16): p. 1227-1232.
2.Torre, L.A., et al., Global cancer statistics, 2012. CA: a cancer journal for clinicians, 2015. 65(2): p. 87-108.
3.Chen, W., et al., Cancer statistics in China, 2015. CA: a cancer journal for clinicians, 2016. 66(2): p. 115-132.
4.Siegel, R.L., K.D. Miller, and A. Jemal, Cancer statistics, 2016. CA: a cancer journal for clinicians, 2016. 66(1): p. 7-30.
5.Bray, F., et al., Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: a cancer journal for clinicians, 2018. 68(6): p. 394-424.
6.Sun, Y.-S., et al., Risk factors and preventions of breast cancer. International journal of biological sciences, 2017. 13(11): p. 1387.
7.Breast360.org. Breast Anatomy. 2019; Available from: https://breast360.org/topic/2017/01/01/breast-anatomy/.
8.Institute, N.C. 29 June 2014; Available from: https://www.cancer.gov/types/breast.
9.Waks, A.G. and E.P. Winer, Breast cancer treatment: a review. Jama, 2019. 321(3): p. 288-300.
10.衛生福利部國民健康署. 乳癌防治. 2019/10/16; Available from: https://www.hpa.gov.tw/Pages/Detail.aspx?nodeid=614&pid=1124.
11.Feng, Y., et al., Breast cancer development and progression: Risk factors, cancer stem cells, signaling pathways, genomics, and molecular pathogenesis. Genes & diseases, 2018. 5(2): p. 77-106.
12.Kamińska, M., et al., Breast cancer risk factors. Przeglad menopauzalny= Menopause review, 2015. 14(3): p. 196.
13.Koh, Y.W., et al., Dual-color silver-enhanced in situ hybridization for assessing HER2 gene amplification in breast cancer. Modern Pathology, 2011. 24(6): p. 794-800.
14.Pegram, M.D., et al., Phase II study of receptor-enhanced chemosensitivity using recombinant humanized anti-p185HER2/neu monoclonal antibody plus cisplatin in patients with HER2/neu-overexpressing metastatic breast cancer refractory to chemotherapy treatment. Journal of clinical oncology, 1998. 16(8): p. 2659-2671.
15.Baselga, J., et al., Phase II study of weekly intravenous recombinant humanized anti-p185HER2 monoclonal antibody in patients with HER2/neu-overexpressing metastatic breast cancer. Journal of clinical oncology, 1996. 14(3): p. 737-744.
16.Baselga, J., et al., Recombinant humanized anti-HER2 antibody (Herceptin™) enhances the antitumor activity of paclitaxel and doxorubicin against HER2/neu overexpressing human breast cancer xenografts. Cancer research, 1998. 58(13): p. 2825-2831.
17.Rubin, I. and Y. Yarden, The basic biology of HER2. Annals of oncology, 2001. 12: p. S3-S8.
18.Program, J.B.C. BREAST CANCER STAGES. 2019; Available from: http://www.jbcp.jo/understandingbreastcancer/33.
19.Akram, M., et al., Awareness and current knowledge of breast cancer. Biological research, 2017. 50(1): p. 1-23.
20.Ljuslinder, I., Studies of LRIG1 and the ERBB receptor family in breast and colorectal cancer. 2009, Umeå universitet.
21.Giuliano, A.E., S.B. Edge, and G.N. Hortobagyi, of the AJCC cancer staging manual: breast cancer. Annals of surgical oncology, 2018. 25(7): p. 1783-1785.
22.Jemal, A., et al., Global cancer statistics. CA Cancer J Clin, 2011. 61(2): p. 69-90.
23.Dai, X., et al., Breast cancer cell line classification and its relevance with breast tumor subtyping. Journal of Cancer, 2017. 8(16): p. 3131.
24.Yin, L., et al., Triple-negative breast cancer molecular subtyping and treatment progress. Breast Cancer Research, 2020. 22(1): p. 1-13.
25.Wolff, A.C., et al., American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. J Clin Oncol, 2007. 25(1): p. 118-45.
26.Tormo, E., et al., The role of miR-26a and miR-30b in HER2+ breast cancer trastuzumab resistance and regulation of the CCNE2 gene. Scientific reports, 2017. 7(1): p. 1-9.
27.Fragomeni, S.M., A. Sciallis, and J.S. Jeruss, Molecular subtypes and local-regional control of breast cancer. Surgical Oncology Clinics, 2018. 27(1): p. 95-120.
28.Araújo, T., et al., Classification of breast cancer histology images using convolutional neural networks. PloS one, 2017. 12(6): p. e0177544.
29.Liu, Z.H., et al., Impact of the updated 2018 ASCO/CAP guidelines on HER2 FISH testing in invasive breast cancer: a retrospective study of HER2 fish results of 2233 cases. Breast Cancer Res Treat, 2019. 175(1): p. 51-57.
30.Rosen, P.P., et al., Immunohistochemical detection of HER2/neu in patients with axillary lymph node negative breast carcinoma. A study of epidemiologic risk factors, histologic features, and prognosis. Cancer, 1995. 75(6): p. 1320-6.
31.Pathmanathan, N. and A.M. Bilous, HER2 testing in breast cancer: an overview of current techniques and recent developments. Pathology, 2012. 44(7): p. 587-95.
32.Tubbs, R.R., et al., Fluorescence in situ hybridization (FISH) as primary methodology for the assessment of HER2 Status in adenocarcinoma of the breast: a single institution experience. Diagn Mol Pathol, 2007. 16(4): p. 207-10.
33.Couturier, J., et al., Strong correlation between results of fluorescent in situ hybridization and immunohistochemistry for the assessment of the ERBB2 (HER-2/neu) gene status in breast carcinoma. Mod Pathol, 2000. 13(11): p. 1238-43.
34.Lebeau, A., et al., Her-2/neu analysis in archival tissue samples of human breast cancer: comparison of immunohistochemistry and fluorescence in situ hybridization. J Clin Oncol, 2001. 19(2): p. 354-63.
35.Yaziji, H., et al., HER-2 testing in breast cancer using parallel tissue-based methods. JAMA, 2004. 291(16): p. 1972-7.
36.Arafah, M., H.K. Kfoury, and S.N. Zaidi, HER2/neu Immunostaining in Invasive Breast Cancer: Analysis of False Positive Factors. Oman Med J, 2010. 25(4): p. 261-3.
37.Gajaria, P.K., et al., Dual-color dual-hapten in situ hybridization (D-DISH) - Comparison with fluorescence in situ hybridization (FISH) for HER2/neu testing in breast cancer. Indian J Pathol Microbiol, 2020. 63(2): p. 194-199.
38.Nishimura, R., et al., Bright-field HER2 dual in situ hybridization (DISH) assay on breast cancer cell blocks: a comparative study with histological sections. Breast Cancer, 2016. 23(6): p. 917-921.
39.Mahoney, L. and A. Csima, Efficiency of palpation in clinical detection of breast cancer. Canadian Medical Association Journal, 1982. 127(8): p. 729.
40.三軍總醫院. 乳癌之診斷. 2020/11/4; Available from: https://wwwv.tsgh.ndmctsgh.edu.tw/unit/10203/15630.
41.Hoorntje, L.E., et al., Vacuum-assisted breast biopsy: a critical review. European Journal of Cancer, 2003. 39(12): p. 1676-1683.
42.Laufer, U., et al., Minimally invasive stereotactic breast biopsy with the" mammotome". Methodology and personal experiences. Zentralblatt fur Chirurgie, 1998. 123: p. 63.
43.Dhillon, M., S. Bradley, and D. England, Mammotome biopsy: impact on preoperative diagnosis rate. Clinical radiology, 2006. 61(3): p. 276-281.
44.Wolff, A.C., et al., American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer. Arch Pathol Lab Med, 2007. 131(1): p. 18-43.
45.Kaptain, S., L.K. Tan, and B. Chen, Her-2/neu and breast cancer. Diagn Mol Pathol, 2001. 10(3): p. 139-52.
46.Slamon, D.J., et al., Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science, 1987. 235(4785): p. 177-82.
47.Tandon, A.K., et al., HER-2/neu oncogene protein and prognosis in breast cancer. J Clin Oncol, 1989. 7(8): p. 1120-8.
48.Slamon, D.J., et al., Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med, 2001. 344(11): p. 783-92.
49.Piccart-Gebhart, M.J., et al., Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. N Engl J Med, 2005. 353(16): p. 1659-72.
50.Burstein, H.J., et al., Trastuzumab and vinorelbine as first-line therapy for HER2-overexpressing metastatic breast cancer: multicenter phase II trial with clinical outcomes, analysis of serum tumor markers as predictive factors, and cardiac surveillance algorithm. J Clin Oncol, 2003. 21(15): p. 2889-95.
51.Press, M.F., et al., Assessment of ERBB2/HER2 Status in HER2-Equivocal Breast Cancers by FISH and 2013/2014 ASCO-CAP Guidelines. JAMA Oncol, 2019. 5(3): p. 366-375.
52.Schmidhuber, J., Deep learning in neural networks: An overview. Neural networks, 2015. 61: p. 85-117.
53.孔英會 and 景美麗, 基於混淆矩陣和集成學習的分類方法研究. 计算機工程與科學, 2012. 34(6): p. 111-117.
54.Shin, H.-C., et al., Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning. IEEE transactions on medical imaging, 2016. 35(5): p. 1285-1298.
55.Cai, Z. and N. Vasconcelos. Cascade r-cnn: Delving into high quality object detection. in Proceedings of the IEEE conference on computer vision and pattern recognition. 2018.
56.Allen, D.M., The relationship between variable selection and data agumentation and a method for prediction. technometrics, 1974. 16(1): p. 125-127.
57.Stone, M., Journal of the royal statistical society. Series B (Methodological), 1974. 36(2): p. 111-147.
58.Stone, M., An asymptotic equivalence of choice of model by cross‐validation and Akaike's criterion. Journal of the Royal Statistical Society: Series B (Methodological), 1977. 39(1): p. 44-47.
59.Chang, J.-S., Y.-F. Luo, and K.-Y. Su. GPSM: A generalized probabilistic semantic model for ambiguity resolution. in 30th Annual Meeting of the Association for Computational Linguistics. 1992.
60.Devijver, P.A. and J. Kittler, Pattern recognition: A statistical approach. 1982: Prentice hall.
61.Mahobia, N.K., et al., Validation method used in quantitative structure activity relationship. Der Pharma Chemica, 2010. 2(5): p. 260-271.
62.Cannaday II, A.B., et al. Decision-level fusion of DNN outputs for improving feature detection performance on large-scale remote sensing image datasets. in 2019 IEEE International Conference on Big Data (Big Data). 2019. IEEE.
63.Ramos-Vara, J. and M. Miller, When tissue antigens and antibodies get along: revisiting the technical aspects of immunohistochemistry—the red, brown, and blue technique. Veterinary pathology, 2014. 51(1): p. 42-87.
64.Nitta, H., et al., Development of automated brightfield double In Situ hybridization (BDISH) application for HER2 g ene and chromosome 17 centromere (CEN 17) for breast carcinomas and an assay performance comparison to manual dual color HER2 fluorescence In Situ hybridization (FISH). Diagnostic pathology, 2008. 3(1): p. 1-12.
65.Pedersen, M. and B.B. Rasmussen, The correlation between dual-color chromogenic in situ hybridization and fluorescence in situ hybridization in assessing HER2 gene amplification in breast cancer. Diagnostic Molecular Pathology, 2009. 18(2): p. 96-102.
66.McKinney, S.M., et al., International evaluation of an AI system for breast cancer screening. Nature, 2020. 577(7788): p. 89-94.
67.Pisano, E.D., AI shows promise for breast cancer screening. 2020, Nature Publishing Group.
68.Chan, H.-P., R.K. Samala, and L.M. Hadjiiski, CAD and AI for breast cancer—recent development and challenges. The British journal of radiology, 2019. 93(1108): p. 20190580.
69.Arlot, S. and A. Celisse, A survey of cross-validation procedures for model selection. Statistics surveys, 2010. 4: p. 40-79.
70.Browne, M.W., Cross-validation methods. Journal of mathematical psychology, 2000. 44(1): p. 108-132.
71.Shao, J., Linear model selection by cross-validation. Journal of the American statistical Association, 1993. 88(422): p. 486-494.
72.Budak, Ü., et al., Computer-aided diagnosis system combining FCN and Bi-LSTM model for efficient breast cancer detection from histopathological images. Applied Soft Computing, 2019. 85: p. 105765.
73.pathology-ai-histopaththolog-vision. https://wwwv.cureintel.com/.

電子全文 電子全文(網際網路公開日期:20260526)
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
無相關論文
 
無相關期刊
 
無相關點閱論文