Multiple studies using primary tumors have reported that alterations in p53 expression and detection of TP53 mutations are associated with clinical aggressiveness and poor response to specific therapies. However, there is no general agreement regarding the optimal technical approach to the analysis of p53. We have studied a series of 100 primary colorectal adenocarcinomas by immunohistochemistry with the monoclonal antibody PAb1801, and single-stranded conformation polymorphism (PCR-SSCP, exons 4-8) followed by direct sequencing of shifted bands. p53 Nuclear staining was undetectable (score 0) in 29 of 100 cases. However, gene mutations were detected in 15 of these cases, with all of these mutations leading to abnormal proteins. p53 Nuclear staining was detectable and scored as less than 10% tumor cells positive in 15 of 100 cases but was still considered to be displaying a p53-negative phenotype because the cut-off value for positivity was 10% positive tumor cells. Nevertheless, TP53 gene mutations were detected in 2 of these cases. p53 Nuclear immunoreactivities were detectable and scored as more than 10% tumor cells positive in 56 cases, considered the p53-positive phenotype. TP53 gene mutations were identified in 51 of these 56 cases. These results reveal that immunohistochemical assessment does not predict TP53 mutation status in colorectal adenocarcinoma, mainly in cases displaying absence of nuclear staining. It is thus concluded that molecular profiling should be conducted in parallel with immunophenotyping when analyzing colorectal tumors for p53 status.