臺灣博碩士論文加值系統

近來深度學習廣泛應用於各行各業中，已有許多成功的案例。根據世界衛生組織統計攝護腺癌在男性癌症排名第二名，因此本篇將深度學習應用於D'Amico Risk分類，期望能提供臨床醫師在進一步檢查前，就可先行知道D'Amico Risk等級，進而讓高風險群能提早進行治療。醫學上核磁共振成像（MRI）輔助有助於攝護腺癌的診斷， MRI主要是用於PI-RADS判定及術前的腫瘤位置探查。
本篇認為MRI對D'Amico Risk具有相關性，因此使用深度學習以MRI影像進行D'Amico Risk分類，並加入以指診資訊及攝護腺特異抗原（PSA）輔助提升辨識準確性，並以深度學習方式改善缺少指診資訊的問題提升辨識準確率。本篇使用衛生福利部台南醫院所蒐集的病患數據為訓練及測試，在測試準確率具有一定的成效。

Recently, deep learning has been widely used in all walks of life, and there have been many successful cases. cording to the World Health Organization, prostate cancer ranks second among men's cancers. Therefore, this article applies deep learning to D'Amico Risk classification. The hope is to provide clinicians with a priori knowledge of the D'Amico Risk level prior to further examination. This allows high-risk groups to receive early treatment. Magnetic resonance imaging (MRI) aids in the diagnosis of prostate cancer. MRI is mainly used for PI-RADS determination and preoperative tumor location exploration.
This article believes that MRI is relevant to D'Amico Risk, so deep learning is used to classify D'Amico Risk with MRI images, and digital diagnosis information and prostate-specific antigen (PSA) are added to help improve the accuracy of identification, and deep learning is used to classify D'Amico Risk The way to improve the accuracy of identification is to improve the problem of lack of digital diagnosis information. This article uses the patient data collected by the Tainan Hospital of the Ministry of Health and Welfare for training and testing, which has a certain effect on the accuracy of the test.

摘要......i
Abstract......ii
誌謝......iii
目錄......iv
表目錄......vi
圖目錄......vii
第一章 介紹......1
1.1 動機......1
1.2 攝護腺及現今醫療診療介紹......1
1.2.1 攝護腺特異抗原......2
1.2.2 肛門指診......3
1.2.3 直腸超音波......4
1.2.4 經直腸超音波導引下攝護腺針刺切片檢查......5
1.2.5 磁共振成像......6
1.2.6 風險判定......7
1.3 現今醫療檢查流程......7
1.4 文獻探討......8
1.5 本研究介紹......9
第二章 方法......12
2.1 數據蒐集與分類......12
2.1.1 MRI影像數據......15
2.1.2 PSA和Clinical stage數據......15
2.2 系統架構......16
2.3 前處理......16
2.3.1 數據量化......16
2.3.2 影像處理......17
2.4 模型......18
2.5 分類法......19
2.6 生成臨床分期（Clinical stage）......20
第三章 實驗分析......22
3.1 評比方法......22
3.1.1 混淆矩陣......22
3.1.2 Leave-One-Out Cross Validation......23
3.2 實驗......24
3.2.1 T2w影像數據......24
3.2.2 加入PSA數據......25
3.2.3 PSA量化......25
3.2.4 加入臨床分期（Clinical stage）量化數據......26
3.3 攝護腺區域感興趣（Region of Interest，ROI）......26
3.4 攝護腺及周圍脂肪區域感興趣（Region of Interest，ROI）......28
3.5 預測clinical stage......29
3.6 預測clinical stage結合ROI......30
第四章 結論......32
文獻參考......33
Extended Abstract......35

[1]C. Wild, E. Weiderpass, and B. W. Stewart, 2020, World cancer report: cancer research for cancer prevention. IARC Press.
[2]N. Mottet et al., 2017, "EAU-ESTRO-SIOG Guidelines on Prostate Cancer. Part 1: Screening, Diagnosis, and Local Treatment with Curative Intent", Eur Urol, vol. 71, no. 4, pp. 618-629, Apr.
[3]J. W. J. J. o. U. Moul, 2017, "Comparison of DRE and PSA in the Detection of Prostate Cancer", vol. 197, no. 2, pp. S208-S209,
[4]X. F. Huang, M. Chen, P. Z. Liu, and Y. Z. Du, 2020, "Texture Feature-Based Classification on Transrectal Ultrasound Image for Prostatic Cancer Detection", Computational and Mathematical Methods in Medicine, Article vol. 2020, p. 9, Oct.
[5]Z. Liu, C. Yang, J. Huang, S. Liu, Y. Zhuo, and X. Lu, 2021, "Deep learning framework based on integration of S-Mask R-CNN and Inception-v3 for ultrasound image-aided diagnosis of prostate cancer", Future Generation Computer Systems, vol. 114, pp. 358-367, 2021/01/01/.
[6]A. V. D'Amico et al., 1998, "Biochemical outcome after radical prostatectomy, external beam radiation therapy, or interstitial radiation therapy for clinically localized prostate cancer", Jama-Journal of the American Medical Association, Article vol. 280, no. 11, pp. 969-974, Sep.
[7]G. Rodrigues et al., 2012, "Pre-treatment risk stratification of prostate cancer patients: A critical review", Can Urol Assoc J, vol. 6, no. 2, pp. 121-7, Apr.
[8]D. J. Tamblyn, S. Chopra, C. Yu, M. W. Kattan, C. Pinnock, and T. Kopsaftis, 2011, "Comparative analysis of three risk assessment tools in Australian patients with prostate cancer", vol. 108, no. s2, pp. 51-56,
[9]D. Gabriele et al., 2016, "Beyond D'Amico risk classes for predicting recurrence after external beam radiotherapy for prostate cancer: the Candiolo classifier", Radiation Oncology, Article vol. 11, p. 10, Feb.
[10]M. R. Cooperberg et al., 2005, "The University of California, San Francisco Cancer of the Prostate Risk Assessment score: a straightforward and reliable preoperative predictor of disease recurrence after radical prostatectomy", J Urol, vol. 173, no. 6, pp. 1938-42, Jun.
[11]M. R. Cooperberg et al., 2006, "Multiinstitutional validation of the UCSF cancer of the prostate risk assessment for prediction of recurrence after radical prostatectomy", vol. 107, no. 10, pp. 2384-2391,
[12]M. W. Kattan, J. A. Eastham, A. M. F. Stapleton, T. M. Wheeler, and P. T. Scardino, 1998, "A Preoperative Nomogram for Disease Recurrence Following Radical Prostatectomy for Prostate Cancer", JNCI: Journal of the National Cancer Institute, vol. 90, no. 10, pp. 766-771,
[13]G. Cosma et al., 2021, "Prostate Cancer: Early Detection and Assessing Clinical Risk Using Deep Machine Learning of High Dimensional Peripheral Blood Flow Cytometric Phenotyping Data", Front Immunol, vol. 12, p. 786828,
[14]Y.-D. Zhang et al., 2016, "An imaging-based approach predicts clinical outcomes in prostate cancer through a novel support vector machine classification", vol. 7, no. 47,
[15]A. Algohary et al., 2020, "Combination of Peri-Tumoral and Intra-Tumoral Radiomic Features on Bi-Parametric MRI Accurately Stratifies Prostate Cancer Risk: A Multi-Site Study", vol. 12, no. 8, p. 2200,
[16]R. Algarra et al., 2014, "Optimizing D'Amico risk groups in radical prostatectomy through the addition of magnetic resonance imaging data", Actas Urologicas Espanolas, Article vol. 38, no. 9, pp. 594-599, Nov.
[17]Y. K. Tsehay et al., 2017, "Convolutional neural network based deep-learning architecture for prostate cancer detection on multiparametric magnetic resonance images", Medical Imaging 2017: Computer-Aided Diagnosis, p. 1013405: International Society for Optics and Photonics.
[18]A. R. Rastinehad et al., 2011, "D'Amico risk stratification correlates with degree of suspicion of prostate cancer on multiparametric magnetic resonance imaging", J Urol, vol. 185, no. 3, pp. 815-20, Mar.
[19]S. Liu, H. Zheng, Y. Feng, and W. Li, 2017, "Prostate cancer diagnosis using deep learning with 3D multiparametric MRI", Medical imaging 2017: computer-aided diagnosis, p. 1013428: International Society for Optics and Photonics.
[20]Y. Yuan et al., 2019, "Prostate cancer classification with multiparametric MRI transfer learning model", Med Phys, vol. 46, no. 2, pp. 756-765, Feb.
[21]J. C. Seah, J. S. Tang, and A. Kitchen, 2017, Detection of prostate cancer on multiparametric MRI (SPIE Medical Imaging). SPIE.
[22]A. Mehrtash et al., 2017, "Classification of Clinical Significance of MRI Prostate Findings Using 3D Convolutional Neural Networks", Proceedings of SPIE--the International Society for Optical Engineering, vol. 10134, p. 101342A,
[23]C. Vente, P. Vos, M. Hosseinzadeh, J. Pluim, and M. Veta, 2021, "Deep Learning Regression for Prostate Cancer Detection and Grading in Bi-Parametric MRI", IEEE Trans Biomed Eng, vol. 68, no. 2, pp. 374-383, Feb.
[24]K. He, X. Zhang, S. Ren, and J. Sun, 2016, "Deep Residual Learning for Image Recognition", 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770-778.