UCL Discovery
UCL home » Library Services » Electronic resources » UCL Discovery

Deep learning applications in the prostate cancer diagnostic pathway

Mehta, Pritesh; (2022) Deep learning applications in the prostate cancer diagnostic pathway. Doctoral thesis (Ph.D), UCL (University College London). Green open access

[thumbnail of PhD_thesis_20221030.pdf]
PhD_thesis_20221030.pdf - Accepted Version

Download (6MB) | Preview


Prostate cancer (PCa) is the second most frequently diagnosed cancer in men worldwide and the fifth leading cause of cancer death in men, with an estimated 1.4 million new cases in 2020 and 375,000 deaths. The risk factors most strongly associated to PCa are advancing age, family history, race, and mutations of the BRCA genes. Since the aforementioned risk factors are not preventable, early and accurate diagnoses are a key objective of the PCa diagnostic pathway. In the UK, clinical guidelines recommend multiparametric magnetic resonance imaging (mpMRI) of the prostate for use by radiologists to detect, score, and stage lesions that may correspond to clinically significant PCa (CSPCa), prior to confirmatory biopsy and histopathological grading. Computer-aided diagnosis (CAD) of PCa using artificial intelligence algorithms holds a currently unrealized potential to improve upon the diagnostic accuracy achievable by radiologist assessment of mpMRI, improve the reporting consistency between radiologists, and reduce reporting time. In this thesis, we build and evaluate deep learning-based CAD systems for the PCa diagnostic pathway, which address gaps identified in the literature. First, we introduce a novel patient-level classification framework, PCF, which uses a stacked ensemble of convolutional neural networks (CNNs) and support vector machines (SVMs) to assign a probability of having CSPCa to patients, using mpMRI and clinical features. Second, we introduce AutoProstate, a deep-learning powered framework for automated PCa assessment and reporting; AutoProstate utilizes biparametric MRI and clinical data to populate an automatic diagnostic report containing segmentations of the whole prostate, prostatic zones, and candidate CSPCa lesions, as well as several derived characteristics that are clinically valuable. Finally, as automatic segmentation algorithms have not yet reached the desired robustness for clinical use, we introduce interactive click-based segmentation applications for the whole prostate and prostatic lesions, with potential uses in diagnosis, active surveillance progression monitoring, and treatment planning.

Type: Thesis (Doctoral)
Qualification: Ph.D
Title: Deep learning applications in the prostate cancer diagnostic pathway
Open access status: An open access version is available from UCL Discovery
Language: English
Additional information: Copyright © The Author 2021. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request.
UCL classification: UCL
UCL > Provost and Vice Provost Offices > UCL BEAMS
UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science
UCL > Provost and Vice Provost Offices > UCL BEAMS > Faculty of Engineering Science > Dept of Med Phys and Biomedical Eng
URI: https://discovery.ucl.ac.uk/id/eprint/10159230
Downloads since deposit
Download activity - last month
Download activity - last 12 months
Downloads by country - last 12 months

Archive Staff Only

View Item View Item