Among various methods which have been developed for facilitating the screening of point mutations in human genomic DNA, PCR-Primer Introduced Restriction Analysis (PCR-PIRA) is of particular interest due to its practicality and short procedure allowing detection of point mutations by simple restriction enzyme digestion directly after PCR amplification. However, one limitation of PCR-PIRA method is the absence of restriction sites in the region of detection, thus creation of the recognition site in primers has been introduced. Detection of a point mutation at codon 12 in K-ras oncogene by BstNI requires one base change in the primer sequence so that only the normal but not mutant PCR product will be digested by the enzyme. However, false positive results generated from undigested normal DNA sequence are always obtained. This effect is compounded when it is used to analyse mixed cell populations in paraffin embedded section of cancer cells. Assay of a mutant band generated from normal DNA by densitometric quantitation enabled the determination of background values and thereby eliminated false positive results. Samples with higher ratios between mutant and normal bands than the background one after the first PCR-PIRA would be subjected to the second PCR-PIRA in order to confirm the results. Screening of such mutations in cervical carcinomas from paraffin embedded sections using the above criteria should reduce misinterpretation of PCR-PIRA results.