Several studies indicate that low thymidylate synthase (TS) protein levels in tumor and normal tissues of colorectal cancer patients are associated with better clinical response to fluorouracil-based chemotherapy and higher risk of toxicity. However, no correlation or even reverse correlation has also been reported. These conflicting results may be partly due to the methodological limitations of the immunohistochemical techniques generally used to quantify thymidylate synthase expression. In this sense, a genetic approach aiming at determining the influence of the TS gene polymorphisms on clinical outcome seems more appealing. So far three polymorphisms have been identified and studied in the TYMS gene: the variable number of 28-bp tandem repeats (2R or 3R) in the 5 UTR; the G>C substitution at the 12th nucleotide in the second repeat of the 3R allele (3RG>3RC) and the 6-bp deletion in the 3 UTR (+6bp/-6bp 3 UTR). In vitro studies indicate that each of these polymorphisms can influence thymidylate synthase expression. In particular, the G>C SNP, which alters the E-box sequence binding an upstream stimulatory factor (USF-1), seems more important than the variable number of tandem repeats in determining TS gene expression in that the 3RC allele has a reduced translational activity compared with the 3RG allele, while showing the same activity as the 2R allele. In contrast with the in vitro findings, the clinical studies in colorectal patients failed to find a consistent relationship between the G>C polymorphism and clinical outcome measures (response, survival or toxicity). This discrepancy may be due to methodological heterogeneities amongst the studies, including genotyping in normal or tumor tissues, loss of heterozygosity in tumor cells not evaluated, variable doses and schedules of fluorouracil-based therapy, and variable tumor stage. The complexity of TYMS gene regulation, and the possibility that other polymorphisms may contribute to fluorouracil response, call for further studies before TYMS genotyping can be used in clinical practice to select colorectal cancer patients who are most likely to benefit from chemotherapy.