MALIMETS-B Feasibility Study
Research type
Research Study
Full title
A feasibility study for the development of a machine learning support for reading whole body diffusion weighted magnetic resonance imaging (WB-MRI) in metastatic breast cancer for the detection and quantification of the extent of disease before and after treatment.
IRAS ID
259078
Contact name
Andrea Rockall
Contact email
Sponsor organisation
Imperial College London
Duration of Study in the UK
0 years, 11 months, 30 days
Research summary
The most common site of metastatic disease in breast cancer is the bone. Bone metastases can range in appearance and change in response to treatment. However, on both computed tomography (CT) and bone scintigraphy (BS) measuring these changes is difficult and assessing response of bone metastases to treatment is challenging. WB-MRI relies on utilising the differences in cellular density and is therefore superior to CT and BS. Better imaging techniques are required to accurately identify patients who are not responding to treatment at an early stage, so that timely treatment changes can be implemented. This may improve patient outcomes and overall survival.
Although DW WB-MRI has its advantages, it is challenging and time consuming to interpret. This is because each whole-body scan comprises several thousand images. Machine Learning (ML) is a computer technique where computers can be trained to automatically identify sites of disease and assess how these have changed in response to treatment.
A large data centred study of this nature with multiple research partners can be challenging to set up and open due to multiple hurdles. We have been advised by the potential funder of the study (a large national cancer charity) to undertake a feasibility study to demonstrate that data sharing agreements between institutions are possible and that ethical approval is granted, in addition to demonstrating feasibility of the pseudonymisation and data transfer at scale. In addition, we propose to test the method needed for natural language processing (NLP) for labelling of the MRI scans for the purposes of training an algorithm to recognise different categories of WB-MRI scans.
REC name
North West - Greater Manchester South Research Ethics Committee
REC reference
19/NW/0684
Date of REC Opinion
5 Nov 2019
REC opinion
Favourable Opinion