Primary surgical resection for locally advanced oesophageal cancer is associated with systemic failure and poor survival due to presence of micrometastatic disease at the time of diagnosis. Neoadjuvant chemotherapy prior to surgical resection aims to downstage these locally advanced tumours. A review of reported randomised controlled trials has shown only one sufficiently powered trial with a survival advantage for cisplatin-based chemotherapy. Published meta-analyses of neoadjuvant chemotherapy trials have shown little or no overall survival benefit. A subgroup of patients with biologically favourable tumours who respond to this treatment have been consistently shown to have a survival advantage. These patients need to be differentiated from non-responders preferably at an early stage of this potentially toxic treatment. Current clinical, endoscopic and radiological methods of response evaluation are all unreliable. Response evaluation with 18FDG-PET has been shown to accurately assess the pathological response and also to predict the risk of local recurrence and overall survival. The development of integrated PET/CT imaging may enhance the accuracy of this response evaluation. In the future, molecular markers of response prediction prior to initiation of treatment may allow the development of individualised treatment strategies. New emerging chemotherapeutic agents may prove to be more effective in eradicating micrometastatic disease.