Explainable AI for Prognostic and Predictive Digital Histopathology in Endometrial Cancer

About the Project

This PhD project aims to develop explainable artificial intelligence (AI) models to analyse digital histopathology whole-slide images (WSIs) of endometrial cancer and its precursor, endometrial hyperplasia, with the goal of improving prognosis prediction, molecular subtyping (ER, PR, P53, MMR, POLE), and therapy response evaluation.

Leveraging deep learning techniques, the project will segment and classify histomorphological features from WSIs and fuse these with clinical and molecular data using multi-modal AI models. Publicly available datasets (e.g. TCGA-UCEC) and local institutional cohorts will be utilised for training and validation.

The project also emphasises explainability to ensure clinical interpretability, using techniques such as attention mapping and SHAP to visualise morphological drivers of AI decisions. Outcomes will include the development of an open-source AI pipeline, identification of novel prognostic markers, and potential translation into clinical decision support systems in gynaecological oncology.

Techniques associated with the project:

• Deep learning (CNNs, vision transformers, survival models)
• Whole-slide image processing (tiling, segmentation, normalisation)
• Multi-modal data fusion (genomic, clinical, and image data)
• Explainability methods
• Digital pathology software (e.g. QuPath)
• Survival and risk modelling
• Access and analysis of large datasets (e.g. TCGA, CPTAC)
• Laboratory testing (e.g. immunohistochemistry)
• Clinical profiling (e.g. electronic healthcare record interrogation, case sourcing)
• Histological evaluation (e.g. identification of hyperplasia/carcinoma)
• Slide scanning

This project is available as part of the International PhD Academy: Medical Research

Eligibility

You should hold a first degree equivalent to at least a UK upper second class honours degree in a relevant subject.

Candidates whose first language is not English must provide evidence that their English language is sufficient to meet the specific demands of their study. The Faculty of Medicine and Health minimum requirements are:

• British Council IELTS - score of 7.0 overall, with no element less than 6.5
• TOEFL iBT - overall score of 95 with not less than 22 in listening, 22 in reading, 24 in speaking and 22 in writing.

How to apply

Applications can be made at any time. To apply for this project applicants should complete an online application form and attach the following documentation to support their application.

• a full academic CV
• degree certificate and transcripts of marks
• Evidence that you meet the University's minimum English language requirements (if applicable)

To help us identify that you are applying for this project please ensure you provide the following information on your application form;

• Select PhD in Medicine, Health and Human Disease as your programme of study
• Give the full project title and name the supervisors listed in this advert

Any queries regarding the application process should be directed to fmhpgradmissions@leeds.ac.uk

Funding Notes

This project is aimed at International applicants who are able to self fund their studies or who have a sponsor who will provide their funding.