This study compares single-strand conformation polymorphism (SSCP)/sequencing and immunohistochemistry (IHC) in a population-based colon cancer study to determine the best methods to evaluate p 53 alterations in tumors. Epidemiologic data collected from the Utah portion of a multicenter case-control study of colon cancer (n = 268) was used to compare somatic p53 mutations detected using SSCP/sequencing of exons 5 through 8 with those with p53 protein overexpression detected by IHC. A total of 136 tumors (51%) had p 53 mutations identified using SSCP/sequencing. IHC detected 164 tumors (61%) with protein overexpression (using a cut point of > or =20% positive cells) and 142 tumors (53%) when > or =50% positive cells were used. Sensitivity of IHC (> or =20% level) using SSCP/sequencing as the reference method was 85%. Specificity of IHC (> or =20% level) using SSCP/sequencing as reference was 63%. When > or =50% positive cells were used, specificity increased to 77%. Associations with age, gender, tumor site, stage, and Ki-ras were similar for both methods. An inverse relationship between microsatellite instability and p 53 was detected with the higher threshold for IHC positivity and SSCP/sequencing. SSCP/sequencing was able to discriminate between mutated p 53 and wild-type p 53 when evaluating dietary associations whereas IHC was not able to discriminate between these tumor types. Using a level of 50% or more positive cells increases specificity relative to sensitivity in comparison with lower staining levels, and is comparable with sequencing in its ability to detect an inverse relationship with the MSI. Advantages gained by sequencing are its ability to examine specific mutations and the improved ability to discriminate between cases with p 53 mutation and wild type when evaluating associations.