Mutations in the p53 gene are detected in greater than 50% of squamous cell carcinomas of the lung and to a lesser extent in adenocarcinomas. The p53 protein is also overexpressed in a relatively high percentage of preinvasive lesions of the bronchial epithelium. However, unlike tumor tissue, immunoreactivity does not necessarily imply that cells in preinvasive lesions carry a mutant p53 allele. In some cases, overexpression may result from a cellular checkpoint reaction to a toxic or mutagenic substance such as exposure to tobacco smoke. In any case, p53 overexpression in preinvasive lesions may serve as a biomarker for high risk assessment of lung cancer and other tumors in the aerodigestive tract. A study was designed to retrospectively analyze p53 overexpression in cells from sputum samples collected prior to histological tumor diagnosis. The rationale was based on the observation that both preinvasive and tumor cells from the bronchial epithelium are exfoliated into the airways and can be detected based on morphology in sputa. Two sets of cases were chosen: 1) patients whose first primary tumor was a squamous cell carcinoma containing a mutant p53 allele with overexpression observed in most of the tumor cells; and 2) patients whose squamous cell tumor did not contain a mutant p53 allele. Cells which stained positive for p53 expression were observed in sputum samples collected from all six patients whose tumors were positive for a mutant p53 allele. Also p53 positive cells were detected on sputum slides for two of the five cases where the tumor DNA did not contain a mutation and/or tumor cells which overexpress p53 were not detected in tissue sections. Although cells which stained positive for p53 were present in sputum from patients whose tumors contained a missense mutation, the presence of p53 overexpression was not specific for tumors which contain an altered p53 allele since overexpression was detected in sputum cells from patients whose tumor DNA did not contain a p53 mutation and/or tumor cells which stained positive for p53 were not observed in tissue sections. However, the p53 positive cells in sputa collected from the latter group of patients could have been exfoliated from other lesions which contained a mutant p53 allele. The accumulation of p53 in some sputum cells was concomitant with expression of simple epithelial type cytokeratins (CK) 8 and 18 or at least one of the other cytokeratins detected by a broad spectrum (PAN) CK antibody mixture. These data imply that most of the sputum cells which overexpress p53 are epithelial cells. Moreover, our results are consistent, at least in part, with other observations that cells which overexpress p53 in dyplasias and hyperplasias express CK 8, 18. We will continue to explore the possibility that expression of cytokeratins 8, 18 and/or other cytokeratins in conjunction with p53 overexpression and/or morphological criteria could define a new class of atypical cells which are predisposed to cancer development.