Artificial intelligence (AI) to predict early progression in follicular lymphoma
Why did we do this study?
Follicular lymphoma is a chronic blood cancer. Patients often do not need treatment at diagnosis, and many continue without treatment for long periods of time before their disease becomes worse (progresses) and they need chemotherapy. This approach is known as 'watch & wait'. However, some patients experience early progression or their disease changes to a more aggressive cancer (transforms). Early progression is likely to be predictable from information about the patient and the cancer. Predicting which patients are at risk of progression and transformation would reassure patients and clinicians about their treatment decisions, and provide reassurance that ‘watch & wait’ (rather than treatment) is suitable. It might also direct earlier or different treatment in high risk patients.
What was our aim?
To produce an artificial intelligence (AI) method to predict the risk of early progression or transformation using data gathered from patients at diagnosis and later in the treatment pathway. The YHHN data is a unique resource that can be used for this type of study.
What did we do?
The project developed a data set with information on patients and outcomes (survival), the individual genetic causes and molecular drivers of disease and information from pathology images on the nature of the tumour. AI was then trained to predict the risk of progression.
What did we find out?