Can detailed analysis of PET/CT images support artificial intelligence approaches to predict the overall survival for upper GI cancer patients?

Abstract

This project proposes to exploit the emerging field of “Radiomics” to evaluate the 18F-FDG-PET scans from a regionally held database of upper GI cancer patients. We propose investigating radiomics statistical textural analysis and deep learning methods to predict the overall survival and progression-free survival (estimated by the number of patients which recur at 1 year post-treatment) of upper GI cancer patients following treatment with surgery and / or chemo-radiotherapy. The aims of this study are to replicate and contribute to the limited published data describing radiomics analysis of upper GI cancer and investigate deep learning algorithms to achieve the same results.