The aim of the study was the identification of a pharmacogenetic profile predictive of the tumor regression grade (TRG), considered as tumor response parameter, after neo-adjuvant treatment in rectal cancer patients. A total of 238 rectal cancer patients treated in a neo-adjuvant setting by a fluoropyrimidines-based chemo-radiotherapy (RT) were genotyped for 25 genetic polymorphisms in 16 genes relevant for treatment-associated pathways. Two polymorphisms were associated with TRG in a multivariate analysis: hOGG1-1245C4G, which can affect radiosensitivity and MTHFR-677C4T, which is involved in fluoropyrimidines action. Patients bearing at least one variant allele had a lower chance to get TRGp2 (OR ¼ 0.46 95% CI 0.23-0.90, P ¼ 0.024; and OR ¼ 0.48 95% CI 0.24-0.96, P ¼ 0.034; respectively). An association trend was observed for ABCB1-3435C4T, which is responsible for the multi-drug resistance (odds ratio (OR) ¼ 1.96, 95% confidence interval (CI) 0.98-3.95, P ¼ 0.057). Exploratory classification and regression tree (CART) analysis highlighted high-order gene-gene and gene-environment interactions and a genetic signature associated with differential response, with hOGG1-1245C4G as the most predictive factor. Other significant variables were: ABCB1-3435C4T, MTHFR-677C4T, ERCC1-8092C4A, ABCC2-1249G4A, XRCC1-28152G4A, XRCC3-4541A4G and patients gender. On the basis of CART results, patients were categorized into three groups according to tumor response probability: intermediate and high profiles had a higher probability to get TRGp2 as compared with low profiles (OR ¼ 4.12 95% CI 1. Po0.001 and OR ¼ 12.44, Po0.0001, respectively). This study evidences a major role of hOGG1-1245C4G and MTHFR-677C4T polymorphisms in the tumor response of rectal cancer patients treated with chemo-RT in neo-adjuvant setting, and shows the relevance of gene-gene and gene-environment interactions for complex phenotypes as tumor response.