Oxaliplatin-based adjuvant chemotherapy is standard of care for radically resected stage III colon cancer and an accepted option for high-risk stage II. Two recent retrospective studies strongly suggested that low-dose aspirin used (100 mg/d) after surgical resection of colorectal cancer with a PIK3CA mutation could act as a targeted therapy with a major protective effect on the risk of recurrence. We propose a double-blind randomized phase III study to evaluate aspirin (100 mg/d during 3 years or until recurrence) versus placebo. Main inclusion criteria are patients aged 18 or 20, stage III or high risk stage II. The primary endpoint of the study is 3-year disease-free survival (DFS). Hypotheses are to improve 3-years DFS from placebo: 72% to aspirin: 83% (HR = 0.56). 94 events and 264 patients with PIK3CA mutation are required. The secondary endpoints are DFS at 5 years, the overall survival rate at 5 years, grade 3-4 severe bleeding.