RadiotherapAIsER: Radiotherapy AI to see Early Recurrence

Lead Research Organisation: Imperial College London
Department Name: Dept of Computing

Abstract

Lung cancer is the leading cause of cancer deaths worldwide. After curative treatment, the cancer recurs in up to 28% of patients. Earlier detection of recurrence may improve overall survival and quality of life. Surveillance needs may differ between patients and stratification may therefore allow for better resource allocation. Our goal is to develop novel machine-learning architectures, built on routinely collected clinical imaging data to facilitate earlier detection of recurrence. Our novel spatio-temporal architecture will explore radiomic and deep-learning approaches, combining CT imaging data from multiple locations: primary tumour, afflicted nodes and surrounding lung, and will incorporate baseline and post-treatment surveillance scans to predict probability of and time to recurrence. The model will be suitable for clinical validation, ready for an early-phase clinical trial investigating surveillance stratification. We will explore the potential for building the model into a dissemination pipeline in a healthcare setting to progress lung cancer management and drive improve in patient outcomes.

Publications

10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/S023283/1 01/04/2019 30/09/2027
2366544 Studentship EP/S023283/1 06/11/2019 05/11/2023 Sumeet Hindocha