Project Details
Description
Background Slow growing brain tumours (such as meningiomas or low-grade gliomas) are usually managed with repeated follow up imaging. This is both expensive for the NHS and causes anxiety to the patient who lives knowing their brain tumour may grow. It is not possible to predict speed of growth or really measure this in a clinical setting. Aims and Objectives The aim of this project is to be able to tell patients if their tumour will grow and how quickly. The objective of the project is to develop an AI software prototype for efficient brain tumour monitoring and progression prediction. The prototype will re-align heterogeneous MRI scans, segment the tumour MRI scans, probabilistic tumour progression prediction with reliable uncertainty quantification, and enhanced clinical interpretation. Work plan and methods used There will be three work packages (WP) designed in this project: WP1 will address the challenges of tumour segmentation in heterogeneous clinical datasets, where transferable deep learning models are lacking due to the significant distribution shift of MRI acquired from different scanners. WP2 will focus on the prediction for tumour progression based on the segmented tumour, where the robustness of the prediction model relies on both the dynamic modelling of progression and reliable uncertainty quantification. Explainable AI methodology will also be developed in WP2 to improve clinical interpretation of the proposed AI models. WP3 will work with the NIHR Brain Injury MIC, Cambridge Enterprise and the Office for Translational Research to start commercialising this product and, with PPI involvement, inform future stages of development. Timeline for delivery: This is a one year project to develop a prototype. Anticipated impact and dissemination: By the end of this stage we will have a working prototype and a commercialisation plan that can proceed to acquiring CE marking.
Status | Finished |
---|---|
Effective start/end date | 1/09/21 → 31/08/22 |
Collaborative partners
- University of Bath
- University of Cambridge (lead)
- University of Liverpool
Funding
- National Institute for Health Research
Fingerprint
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.