Ependymoma is the second most common paediatric malignant brain tumour and has a dismal outcome. WHO histological grade provides insight to prognosis and in most series confers a poor survival. Here we present a method to non-invasively predict the grade of paediatric posterior fossa ependymoma using diagnostic MR imaging (T2w and ADC) and machine learning. We found that WHO Grade II and III tumours can both be predicted with a sensitivity/specificity of 0.7±0.23 and 0.67±0.15 respectively. We believe these results provide the basis for a clinically important aid to decision making in the early stages of treatment.
This abstract and the presentation materials are available to members only; a login is required.