Disease Modelling
Key imaging analysis tasks (segmentation, detection, synthesis) meet their full potential when integrated in the wider consideration of their clinical importance either as imaging biomarker or as essential key for improved disease understanding.
As a covariate in statistical analyses, corrective confounder or variable to explain, imaging derived measurements have taken an increasingly larger place in disease modelling. For instance analysis of the relationships between imaging markers of neurovascular disease and cardiovascular risk factors can help propose new hypotheses of disease evolution that can then be tested using longitudinal information.
Further, using imaging-derived characteristics in association with clinical information as features for machine learning systems can help in predicting disease evolution, patient outcome or in designing appropriate treatment strategy.
Ultimately, linking together information extracted from different organs/systems along a disease pathway can contribute to the creation and/or validation of potential pathological mechanistic relationships.
Publications
Decision fusion of 3D convolutional neural networks to triage patients with suspected prostate cancer using volumetric biparametric MRI
Mehta, P., Antonelli, M., Ahmed, H., Emberton, M., Punwani, S. and Ourselin, S., 2020, In Medical Imaging 2020: Computer-Aided Diagnosis (Vol. 11314, p. 1131433). International Society for Optics and Photonics.
Accurate multimodal probabilistic prediction of conversion to Alzheimer’s disease in patients with mild cognitive impairment
Young, J., Modat, M., Cardoso, M.J., Mendelson, A., Cash, D., Ourselin, S. and Alzheimer’s Disease Neuroimaging Initiative, 2013. NeuroImage: Clinical, 2, pp.735-745.