AIMS: This prospective study was designed to determine the utility of 18F-labelled deoxyglucose (FDG) in positron emission tomography (PET) (FDG-PET) for assessing the response to neoadjuvant chemoradiation therapy (CRT) in locally advanced oesophageal tumours. PATIENTS AND METHODS: Thirty-six patients with locally advanced oesophageal cancer (clinical T4 stage) without organ metastases, underwent FDG-PET before and 1 month after CRT. Patients were classified as major responders by serial FDG-PET when the post-CRT PET demonstrated a strong reduction of FDG uptake at the primary tumour site (>80% reduction of tumour-to-liver uptake ratio) without any abnormal FDG uptake elsewhere in the body. PET response was compared with histology obtained during post-induction transthoracic oesophagectomy. RESULTS: A strong correlation was found between the extent of lymph node (LN) involvement as shown by the pre-CRT PET and the major response rate (P = 0.001): such response occurred in nine of 11 N0M0 patients (82%), in three of nine N(1-2)M0 patients (33%) and in two of 16 patients (13%) with distant lymphatic spread. Such a correlation was not found for computed tomography or endoscopic ultrasonography. The sensitivity of serial FDG-PET for a major CRT response was 10 of 14 (71%), its specificity 18 of 22 (82%). The concordance between the response assessment by PET and histopathology was 78%. The median survival time after CRT of PET major responders compared with PET non-major responders was 16.3 months and 6.4 months, respectively. The metabolic response as measured by serial FDG-PET is a stronger prognostic factor for overall survival (P = 0.002) than the extent of LN involvement seen on the pretreatment FDG-PET (P = 0.087). CONCLUSIONS: These data indicate that CRT response as assessed by serial FDG-PET is strongly correlated with pathological response and survival.