Gayo, Iani Jirah Mae Bulatao;
(2025)
Deep Reinforcement Learning for Prostate Cancer Biopsy Planning.
Doctoral thesis (Ph.D), UCL (University College London).
Preview |
Text
Gayo_Thesis.pdf - Submitted Version Download (16MB) | Preview |
Abstract
Prostate Cancer (PCa) is the most common cancer in men, and is a leading cause of death worldwide. Targeted biopsy, which utilises pre-procedure magnetic resonance images (MRI) to identify suspected lesions, offers a promising approach to precisely target cancer during an ultrasound (US) guided biopsy procedure. An MR-targeted approach reduces patient discomfort as fewer number of needles are used, and leads to reduced risk of infection. However, this method still fails to detect some clinically significant cancer, due to limited investigation into optimal targeting strategies, intra-procedure errors and the presence of MR-invisible lesions. To address the challenges of MR-targeted biopsy, deep reinforcement learning is proposed for biopsy planning, a paradigm which learns the optimal sequence of actions to maximise a reward. This framework has potential to learn novel biopsy targeting strategies and learn dynamically adapting biopsy needle positions that can mitigate intra-procedure errors that cause missed or underdiagnosed prostate cancer. The key contributions of this thesis are as follows: (i) Novel sampling strategies are learned for patient-specific cases which adapt to the size of the lesion, (ii) Targeting strategies are learned through intra-procedure planning, that adapt to observed intra-procedure changes such as deformation and registration, (iii) US images are also incorporated for intra-procedure planning, to closely reflect a real US-guided biopsy, and finally (iv) initial results are presented to learn biopsy strategies that can target MR-invisible lesions. The presented works identify optimal biopsy positions that can be used to improve the biopsy procedure, reduce patient discomfort and increase the detection of clinically significant cancer.
| Type: | Thesis (Doctoral) |
|---|---|
| Qualification: | Ph.D |
| Title: | Deep Reinforcement Learning for Prostate Cancer Biopsy Planning |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2025. 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 > Dept of Med Phys and Biomedical Eng |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10208197 |
Archive Staff Only
 |
View Item |
