Imaging is central to diagnosing and monitoring solid cancers and scans are a key part of the patient and family journey. Different types of scans such as CT, MRI and PET, have all become commonplace in hospitals throughout the UK and children benefit from the accurate high-quality images they produce.
Patients and families want to know what the tumour is, how it will be treated and what the outcome will be and want this information as quickly as possible. Imaging is increasingly meeting these challenges and the use of computers with artificial intelligence (AI) promises dramatic improvements in speed and reliability in the future.
Leading the way in imaging research
The UK has a long history of excellence in imaging research, with the discovery and development of both CT and MRI being perhaps the biggest contributions this country has made to cancer care worldwide. Ensuring that children with cancer are at the forefront of this innovation is essential.
Imaging is often thought of as purely producing pictures of the inside of the body. However, there has been a shift towards using imaging to provide information on the properties of tumours, such as the blood flow in them or their chemical make-up. MRI scans can be readily adapted to acquire a whole range of information on tumour properties, giving us a better idea of what type of tumour the child has, and how aggressive it is going to be. These methods together are sometimes called ‘functional imaging’, since they aim to investigate how the tumour is functioning.
In 2004, we set up the Functional Imaging of Tumours study with CCLG and our network of hospitals. We built our own computerised database to store the images and made it available to researchers on the internet.
From the outset, we used computers and early artificial intelligence to analyse the images. Success came quickly and we achieved impressive accuracy in diagnosing the main childhood brain tumour types.
Professor Andrew Peet, of the University of Birmingham
Seventeen years later, the CCLG functional imaging database has more than 1,500 cases with scans collected from 10 centres and is at the forefront of the international effort to develop imaging and AI. The Functional Imaging of Tumours study is still open and recruiting patients, and underpins a number of important research studies through its database of scans.
Here are two of these projects, funded by the Little Princess Trust (LPT) in partnership with CCLG, which demonstrate recent successes.
Improving the diagnosis of children’s brain tumours by functional radiomics
Institution: University of Birmingham
Amount Awarded: approx. £100,000
Undertaken by early career researcher Dr James Grist, who won the Sir Peter Mansfield Prize for Innovation in MRI for the work, this project focuses on predicting survival for an individual child. Some tumours don’t behave as doctors would expect, and usual treatments don’t work. Unfortunately, it can take a long time to know whether a treatment is working, sometimes too long when tumours are not responding to treatment.
We’re using AI to help analyse the images coming off the MRI scanner before they have started treatment to predict whether a child will be okay with normal treatments, or if they would need to be considered for new therapies to treat their tumour.
We’re going to be taking this discovery forward to test it at sites around the UK, bringing hope to children with a brain tumour around the country.
What‘s exciting about this discovery is that we’re using information gathered from the moment the child comes to hospital, and not having to wait a long time to see if our treatments are working.
Dr James Grist
Moving artificial intelligence of functional imaging for children’s tumours into a multi-centre environment through a clinical decision support system
Institution: University of Birmingham
Amount Awarded: approx. £100,000
While functional imaging is becoming widely available, interpreting new scans can be challenging for doctors and this is being addressed in this research study. Undertaken by early career researcher Dr Heather Rose, who won the prize for best research poster at the 2021 CCLG Winter Meeting, the team have designed an app that allows hospitals to use these advanced scans to help diagnose their patients.
This app presents results in comparison to scans in our database which allows doctors to assess a patient’s scan against scans where we already know the type of tumour present. AI uses mathematics and the scan results to predict the type of tumour and feed this information back to doctors.
This app will be made available for testing by staff in hospitals to gather feedback and make further improvements. It’s fantastic to think that these amazing methods are going to be available to doctors within the hospital.
Dr Heather Rose
From Contact magazine issue 93 - December 2021
