Image analysis for oesophageal cancer: examining benefits for patients and doctors

Abstract

Oesophageal cancer is a common and severe oncologic disease. The incidence of oesophageal cancer has increased over the past decades, posing a growing health problem worldwide. The majority of patients diagnosed with oesophageal cancer do not qualify for curative treatment, often due to metastatic disease. For the minority of patients eligible for curative therapy, surgical resection of the oesophagus, including lymph node dissection, is the standard treatment, often preceded by neoadjuvant chemoradiotherapy. Diagnostic imaging plays an essential role in both diagnosis and formulation of treatment plans for oesophageal cancer. Diagnosis of oesophageal cancer typically involves endoscopy with biopsies and fluorodeoxyglucose positron emission tomography computed tomography (18FDG-PETCT) scans for clinical staging. Subsequent to the assessment of tumour stage, physical condition, and the patient’s preferences, the optimal treatment plan is determined. The aim of this thesis is to explore possible improvements of outcome of the treatment of oesophageal cancer through medical image analysis. Part I of this thesis focuses on the staging of oesophageal cancer using endoscopic ultrasound and addresses the assessment of response to chemoradiotherapy through 18FDG-PET-CT and functional magnetic resonance imaging (MRI). Part II involves frailty assessment using imaging and its predictive value for postoperative outcomes. Part III covers innovations in surgical training and the role of image analysis in this area, namely exploration of automatic recognition of anatomy in surgical recordings using artificial intelligence (AI) and the possibilities of remote training of surgeons through telementoring.