Metabolic response at repeat PET/CT predicts pathological response to neoadjuvant chemotherapy in oesophageal cancer

Objectives Reports have suggested that a reduction in tumour 18F-fluorodeoxyglucose (FDG) uptake on positron emission tomography (PET) examination during or after neoadjuvant chemotherapy may predict pathological response in oesophageal cancer. Our aim was to determine whether metabolic response predicts pathological response to a standardised neoadjuvant chemotherapy regimen within a prospective clinical trial. Methods Consecutive patients staged with potentially curable oesophageal cancer who underwent treatment within a non-randomised clinical trial were included. A standardised chemotherapy regimen (two cycles of oxaliplatin and 5- fluorouracil) was used. PET/CT was performed before chemotherapy and repeated 24-28 days after the start of cycle 2. Results Forty-eight subjects were included: Mean age 65 years; 37 male. Using the median percentage reduction in SUVmax (42%) to define metabolic response, pathological response was seen in 71% of metabolic responders (17/24) compared with 33% of non-responders (8/24; P=0.009, sensitivity 68%, specificity 70%). Pathological response was seen in 81% of subjects with a complete metabolic response (13/16) compared with 38% of those with a less than complete response (12/32; P=0.0042, sensitivity 52%, specificity 87%). There was no significant histology-based effect. Conclusions There was a significant association between metabolic response and pathological response; however, accuracy in predicting pathological response was relatively low. Key Points ̇ PET/CT may predict tumour response to chemotherapy in oesophageal cancer. ̇ This was a prospective study using a standardised chemotherapy regimen. ̇ A significant association between PET/CT findings and disease response was found. ̇ However accuracy in predicting pathological response was relatively low. © European Society of Radiology 2012