ASCO Annual Meeting location:Chicago, Illinois, USA date:4-8 June 2010
Background: To assess the value of early metabolic response with 18-FDG-PET-CT after one cycle of chemotherapy as a tool to predict pathological response and outcome in patients with locally advanced cancer of the oesophagus and GE-junction (GEJ) undergoing subsequent neoadjuvant chemoradiotherapy (nCRT). Methods: In a prospective, blinded study from 2004 - 2008 a total of 55 patients with cT3-4 N0/+ cancer of the oesophagus (n = 42) and GEJ (n = 13), underwent a PET-CT at baseline, 2 weeks after a single cycle of 5FU/cisplatinum (interval PET) and after completion of nCRT (5FU/cisplatinum first and last week of RT: 45 Gy in 25 fractions) (late PET). Patients showing signs of progressive disease (n = 8) were excluded from oesophagectomy. Early metabolic response (mR), late mR and complete mR were defined by a decrease in mean SUV of ≥ 35%, ≥ 50% and 100% respectively. The Mandard tumor regression (TRG) scoring system was used to score pathological response (pR). Cancer- specific survival (CSS) and time-to-recurrence (TTR) rates were estimated by the Kaplan-Meier method. Results: Mean decrease in SUVmean of the tumor was 36.4 ± 29.3% on interval PET and 70.3 ± 27.0% after completion of nCRT. Early mR was identified in 26/55 (47.3%) patients. Early mR and late mR were not associated with longer CSS (p = 0.77 and 0.61 respectively) and TTR (p = 0.49 and 0.57 respectively). A significant difference in CSS (p= 0.03) but not in TTR (p 0.74) was found in patients who had complete metabolic regression on late PET (n = 14). Good pR (TRG1 and TRG2) was shown in 54.3% of resected tumors. PPV and NPV for early mR in relation to good pR were 60% and 52% respectively. Conclusions: A decrease in SUVmean of ≥ 35% on PET-CT after one cycle of chemotherapy is not a good predictor of pR and outcome in patients with locally advanced oesophageal cancer subsequently treated by nCRT. Only complete mR after completion of CRT seems to have a significant prognostic value. Refinement and validation of trial-specific cut-off criteria for metabolic response evaluation is necessary.