Metabolic tumour volume segmentation for oesophageal cancer on hybrid PET/CT using convolutional network architecture

Oesophageal cancer (OC) has a particularly poor prognosis with an overall 5-year survival rate of only 15%. OC is rising in incidence and is a cancer with unmet clinical need. The segmentation of metabolic tumour volume (MTV) is time consuming and subject to intra and inter-observer variability. This study aims to increase the efficiency of MTV segmentation in OC by developing a hybrid PET/CT deep-learned model based on convolutional network architecture.