Background Resection remains the best treatment for carcinoma of the oesophagus in terms of local control, but local recurrence and distant metastasis remain an issue after surgery. We aimed to assess whether a short preoperative chemoradiotherapy regimen improves outcomes for patients with resectable oesophageal cancer. Methods 128 patients were randomly assigned to surgery alone and 128 patients to surgery after 80 mg/m(2) cisplatin on day 1, 800 mg/m(2) fluorouracil on days 1-4, with concurrent radiotherapy of 35 Gy given in 15 fractions. The primary endpoint was progression-free survival. Secondary endpoints were overall survival, tumour response, toxic effects, patterns of failure, and quality of life. Analysis was done by intention to treat. Findings Neither progression-free survival nor overall survival differed between groups (hazard ratio [HR] 0.82 [95% CI 0.61-1.101 and 0.89 [0.67-1.19], respectively). The chemoradiotherapy-and-surgery group had more complete resections with clear margins than did the surgery-alone group (103 of 128 [80%] vs 76 of 128 [59%], p=0.0002), and had fewer positive lymph nodes (44 of 103 [43%] vs 69 of 103 [67%], p=0.003). Subgroup analysis showed that patients with squamous-cell tumours had better progression-free survival with chemoradiotherapy than did those with non-squamous tumours (HR 0.47 [0.25-0.86] vs 1.02 [0.72-1.44]). However, the trial was underpowered to determine the real magnitude of benefit in this subgroup. Interpretation Preoperative chemoradiotherapy with cisplatin and fluorouracil does not significantly improve progression-free or overall survival for patients with resectable oesophageal cancer compared with surgery alone. However, further assessment is warranted of the role of chemoradiotherapy in patients with squamouscell tumours.