Neuroendocrine neoplasms (NENs) are a group of rare, heterogeneous cancers which are increasing in incidence. Depending on the grade, NENs are classified as slow-growing or fast-growing; different lesions in the same patient may exhibit vastly different behaviours. Because of this heterogeneity, it is challenging for clinicians to accurately predict prognosis and to optimise treatments based on current clinical data.
Radiomics is the study of image-based biomarkers, which can be extracted from routinely collected medical imaging. These metrics describe features of the image from simple (shapes, brightness, contrasts) to more complex features (contrasts between adjacent structures and their arrangement in the image). These radiomic features are difficult for human reporters to appreciate, but increasing research has identified their promise in predicting disease biology and aiding personalised treatment planning. The vast amounts of data generated from radiomics analysis also lends itself to model building using AI methods.
This session will explore the potential of radiomics for oncology, using NENs as an example. The translation of radiomics to clinical practice and the application of AI based on radiomics will also be evaluated.