AI for radiotherapy planning
Research type
Research Study
Full title
ASPIRE: Artificial Intelligence Solutions for Personalised Radiotherapy
IRAS ID
258562
Contact name
Tom Crosby
Contact email
Sponsor organisation
Velindre University NHS Trust
Duration of Study in the UK
1 years, 11 months, 30 days
Research summary
ASPIRE is a study that aims to train and test an artificial intelligence (AI) algorithm to automatically outline oesophageal tumours using computed tomography (CT) and positron emission tomography (PET) scans, using Intel’s advanced computing hardware. We think that the ability to perform this task will significantly reduce the time taken to plan radiotherapy for each patient and therefore reduce the time delay to treatment. Currently, doctors that plan radiotherapy (clinical oncologists) spend considerable time outlining the tumour, making sure that all the visible disease on the scans are included in the treatment field. However, work from our group has shown that there are significant differences in radiotherapy plans created by different doctors. Although planning is an important task, the process is time-consuming, time which arguably could be spent providing more patient care. Automatic outlining of the tumours will eliminate these differences between individuals, ensuring that every patient has the best and most consistent radiotherapy plan possible. We plan to use existing data from CT and PET scans with tumours that are already outlined to teach the AI system to work out where the edge of the tumour is. We will then test the performance of the AI algorithms using prospectively collected data (data collected over the next 2 years). The ability to automatically outline tumours will also help doctors monitor the volume of disease before, during and after therapy, providing both patients and clinicians valuable information to help make shared decisions about future treatments.
REC name
N/A
REC reference
N/A