Artificial intelligence (AI) and machine learning (ML) are rapidly changing the landscape of radiology. AI and ML have the potential to aid tumour detection across the body, provide segmentation of disease and normal anatomy for treatment planning, as well as the automatic tracking of tumours during treatment.
In diagnostic imaging, MR images provide both anatomical and functional information, which can yield diagnostic, response or prognostic biomarkers. However, the use of quantitative data from MRI depends on accurate tumour identification and segmentation, but manual image segmentation is prohibitively time-consuming and is associated with inter-observer variations. Automatic or semi-automatic tools that can aid radiologists to rapidly identify and perform tumour segmentation would increase the usage of quantitative imaging in oncology and improve the curation of prospective imaging data for developing future AI and ML-based imaging tools.
In radiotherapy, recent developments in MRI for radiotherapy planning, as well as the advent of combined MRI-Linac systems, necessitate developments in tumour definition and segmentation using MR images to optimise treatment planning. Furthermore, there is a need to identify organs-at-risk comprising normal anatomical structures to minimise toxicity to normal tissues. Hence, the development of novel algorithms and tools that would facilitate the rapid assessment tumour volume and organs at risk using MR images would be welcomed.
In this proposed PhD project, we will investigate, develop and apply computational imaging algorithms that will identify normal pelvic anatomical structures and detect pelvic tumours (including cervical and rectal cancers), as well as perform volumetric tumour segmentations for use in radiological and radiotherapy clinical applications.
• To design and optimize ML algorithms for identifying normal pelvic organs and anatomy on MRI
• To design and optimize ML algorithms to detect and segment cervical cancers on MRI, and to generate volumes of interest for radiological and radiotherapy applications
• To design and optimize ML algorithms to detect and segment rectal cancers on MRI, and to generate volumes of interest for radiological and radiotherapy applications
This studentship will provide the following translational training experience
The student will work within a multidisciplinary team in a hospital/ scientific environment, which will provide ample opportunities for the student to learn about the relevance of their work and research towards patient care and disease outcomes. The student will acquire new skills in AI/ ML and learn how to apply these practically and in a clinically meaningful way to build new tools that can help radiologists and radiation oncologists to work more effectively.
This project provides an excellent opportunity for the student to translate science to practice; and student learning will be supported by a strong supervisory team that comprises experts in physics, computational analysis, radiologists and radiation oncologists. In addition, this research will be developed in collaboration with Chang Gung Memorial Hospital (Taipei) and Queen Mary Hospital (Hong Kong), providing the candidate exposure to multi-centre collaborations.
Candidates must have a first class or upper second class honours BSc Honours/ MSc in Physics or Engineering; Maths; Computer Science
How to apply
Full details about these studentship projects, and the online application form, are available on our website, at: http://www.icr.ac.uk/phds
Applications for all projects should be made online. Please ensure that you read and follow the application instructions very carefully.
Closing date: Monday 6th May
Applicants should be available for interview end of May 2019.
Please apply via the ICR vacancies web portal https://apply.icr.ac.uk/