While survival rates have steadily improved for childhood brain tumours, with 75% of patients now living beyond five years, non-invasive diagnosis of the condition and accurate prediction of how the disease will progress in individual patients has remained a challenge. Childhood cancers are very different to those found in adults, they are often more difficult to treat, and the treatment can be incredibly debilitating.

Researchers at WMG, University of Warwick, led by Professor Theodoros Arvanitis have pioneered advanced imaging and computational analysis approaches to provide children with brain tumours a non-invasive method of disease diagnosis, its management, and prediction of survivability.

Quicker and more accurate diagnosis

Building upon previous research on advanced MRI-based functional imaging, and using machine learning, the team were able to produce improved childhood brain tumour diagnostic classifiers. Although visual assessment can be subjective, the research demonstrates that advanced imaging techniques that incorporate machine learning algorithms can be used for paediatric brain tumours in clinical practice, as a completely non-invasive method of providing a quick and accurate diagnosis, while helping predict survivability.

Working in partnership with the Children's Cancer and Leukaemia Group, as well as external academic institutions and an associated network of NHS trusts, WMG’s research into functional imaging technology has led to success in achieving results to differentiate between various types and grades of tumours. Currently most studies only demonstrate the use of imaging techniques to differentiate between grades or types of one kind of tumour.

Professor Theodoros Arvanitis, WMG commented: “This imaging is vital to ensuring children receive the best treatment of brain tumours. The suite of MRI-based functional imaging methods and associated computational analysis were first pioneered in Birmingham Children’s Hospital, in collaboration with Warwick and University of Birmingham – and now around half of all major centres specialising in paediatric oncology are using it.”

Improving treatment

Research at WMG has helped create clinical improvements in diagnosis and management of childhood brain tumours, with the additional benefit of predicting survivability for children with these brain tumours, which deliver benefits directly to patients, the NHS and healthcare providers, as well as national charities.

Dr Karen Manias, Birmingham Children’s Hospital, Birmingham Women’s and Children’s NHS Foundation Trust commented: “The collaboration with Professor Arvanitis has allowed us to bring these emerging functional imaging technologies into the clinical care of children with cancer. Every day we are seeing the children benefit from improvements in diagnosis, treatment planning and survival prediction.”

Professor Andrew Peet, Chair of Clinical Paediatric Oncology at the University of Birmingham and Paediatric Oncologist at Birmingham Children’s Hospital, Birmingham Women’s and Children’s NHS Foundation Trust concluded: “Through our advanced neuro-imaging approach and AI technologies we have been able to advance the diagnosis and treatment of childhood cancer significantly. The imaging technology helps us to make more accurate and earlier diagnosis of brain tumours in children.

Thanks to our advanced imaging and analysis, we are able to start planning a treatment regime such as a course of radiotherapy straight away, without significant waits for biopsy results from the lab. Treatment can be tailored more accurately, ensuring patients are given the most effective management while minimising side effects. It also means that we have faster and more informed conversations with the families of our patients, many of whom say that the most difficult time is the period spent “in limbo” waiting for diagnosis. This is undoubtedly a contributing factor to an increase in survival rates for children with cancer, with more than 75% of patients now living beyond five years.”

To learn more about this research visit:

Characterisation of Paediatric Brain Tumours (warwick.ac.uk)

Diffusion MRI and machine learning models classify childhood brain tumours – Physics World

Child brain tumours can be classified by advanced imaging and AI - WMG: News (warwick.ac.uk)

Radiomics: using the power of informatics to more accurately diagnose and treat childhood brain tumours - WMG explains (warwick.ac.uk)