6533b872fe1ef96bd12d2ae2

RESEARCH PRODUCT

Artificial intelligence for image-guided prostate brachytherapy procedures

subject

description

Radiotherapy procedures aim at exposing cancer cells to ionizing radiation. Permanently implanting radioactive sources near to the cancer cells is a typical technique to cure early-stage prostate cancer. It involves image acquisition of the patient, delineating the target volumes and organs at risk on different medical images, treatment planning, image-guided radioactive seed delivery, and post-implant evaluation. Artificial intelligence-based medical image analysis can benefit radiotherapy procedures. It can help to facilitate and improve the efficiency of the procedures by automatically segmenting target organs and extrapolating clinically relevant information. However, manual delineation of target volumes is still the standard routine for most clinical centers, which is time-consuming, challenging, and not immune to intra- and inter-observer variations. In this thesis, we aim to develop medical image processing solutions to automate various components of the current image-guided prostate brachytherapy procedures, including radioactive seeds identification from CT images and clinical target volume segmentation from different medical images. In the first application, we developed and evaluated a new technique for detecting and identifying implanted radioactive seeds on post-implant CT scans of prostate brachytherapy. This allows experts to evaluate the quality of the image-guided radioactive seed delivery by computing the delivered dosimetric parameters, specifically to compute the post-implant dosimetry of salvage prostate brachytherapy performed years after primary brachytherapy in the treatment of relapsed prostate cancer. The second application involved the development of deep learning methods to delineate clinical target volumes automatically. We evaluated the proposed methods on a clinical database of intraoperative transrectal ultrasound and post-implant CT images of image-guided prostate brachytherapy. The evaluation is then extended to other medical image analysis applications. Our methods yielded promising results and opening important perspectives towards efficient and accurate medical image analysis tasks. They can be applied to automate the management of image-guided prostate brachytherapy procedures.