Definitive chemoradiotherapy is a curative treatment option for oesophageal carcinoma, especially in patients unsuitable for surgery. The PRODIGE5/ACCORD17 trial aimed to assess the efficacy and safety of the FOLFOX treatment regimen (fluorouracil plus leucovorin and oxaliplatin) versus fluorouracil and cisplatin as part of chemoradiotherapy in patients with localised oesophageal cancer.

We did a multicentre, randomised, open-label, parallel-group, phase 2/3 trial of patients aged 18 years or older enrolled from 24 centres in France between Oct 15, 2004, and Aug 25, 2011. Eligible participants had confirmed stage I-IVA oesophageal carcinoma (adenocarcinoma, squamous-cell, or adenosquamous), Eastern Cooperative Oncology Group (ECOG) status 0-2, sufficient caloric intake, adequate haematological, renal, and hepatic function, and had been selected to receive definitive chemoradiotherapy. Patients were randomly assigned (1:1) to receive either six cycles (three concomitant to radiotherapy) of oxaliplatin 85 mg/m(2), leucovorin 200 mg/m(2), bolus fluorouracil 400 mg/m(2), and infusional fluorouracil 1600 mg/m(2) (FOLFOX) over 46 h, or four cycles (two concomitant to radiotherapy) of fluorouracil 1000 mg/m(2) per day for 4 days and cisplatin 75 mg/m(2) on day 1. Both groups also received 50 Gy radiotherapy in 25 fractions (five fractions per week). Random allocation to treatment groups was done by a central computerised randomisation procedure by minimisation, stratified by centre, histology, weight loss, and ECOG status, and was achieved independently from the study investigators. The primary endpoint was progression-free survival. Data analysis was primarily done by intention to treat. This study is registered with ClinicalTrials.gov, number NCT00861094.

134 participants were randomly allocated to the FOLFOX group and 133 to the fluorouracil and cisplatin group (intention-to-treat population), and 131 patients in the FOLFOX group and 128 in the fluorouracil and cisplatin group actually received the study drugs (safety population). Median follow-up was 25·3 months (IQR 15·9-36·4). Median progression-free survival was 9·7 months (95% CI 8·1-14·5) in the FOLFOX group and 9·4 months (8·1-10·6) in the fluorouracil and cisplatin group (HR 0·93, 95% CI 0·70-1·24; p=0·64). One toxic death occurred in the FOLFOX group and six in the fluorouracil-cisplatin group (p=0·066). No significant differences were recorded in the rates of most frequent grade 3 or 4 adverse events between the treatment groups. Of all-grade adverse events that occurred in 5% or more of patients, paraesthesia (61 [47%] events in 131 patients in the FOLFOX group vs three [2%] in 128 patients in the cisplatin-fluorouracil group, p<0·0001), sensory neuropathy (24 [18%] vs one [1%], p<0·0001), increases in aspartate aminotransferase concentrations (14 [11%] vs two [2%], p=0·002), and increases in alanine aminotransferase concentrations (11 [8%] vs two [2%], p=0·012) were more common in the FOLFOX group, whereas serum creatinine increases (four [3%] vs 15 [12%], p=0·007), mucositis (35 [27%] vs 41 [32%], p=0·011), and alopecia (two [2%] vs 12 [9%], p=0·005) were more common in the fluorouracil and cisplatin group.

Although chemoradiotherapy with FOLFOX did not increase progression-free survival compared with chemoradiotherapy with fluorouracil and cisplatin, FOLFOX might be a more convenient option for patients with localised oesophageal cancer unsuitable for surgery.

UNICANCER, French Health Ministry, Sanofi-Aventis, and National League Against Cancer.