Lung cancer is the leading cause of cancer-related deaths in western countries. The prognosis for patients with lung cancer depends primarily on the stage of the tumor at the time of clinical diagnosis. New understanding of tumor biology has turned attention away from detection of clinical lung cancer, usually metastatic at presentation, toward recognition of genetic and protein markers which precede malignancy. Mutations of four types of genes contribute to the process of epithelial carcinogenesis by modifying control of cell growth. Examples of three of these changes have been detected in pre-malignant sputum, and validated in subsequent tumor. We have identified gene products (tumor associated and differentiation protein antigens), mutations of k-ras and p53, and microsatellite alterations as potential markers of subsequent malignancy. We consider the morphologic progression seen in archived sputum cells as the paradigm of neoplastic development in the lung. Although the NCl collaborative trials had shown that this progression is not recognized sufficiently often (sensitive) to be useful for lung cancer screening, this progression may be used to assess the timing of gene and peptide markers of carcinogenesis. Previous work has shown that at the time Johns Hopkins Lung Project sputum cells express moderately atypical metaplasia, 53% (8/15) of sputum specimens expressed common (codon 12) k-ras or (codons 273 or 281) p53 mutations. Other investigators have reported that earlier morphologic changes (metaplasia) accompany 3p and 9p losses of heterozygosity. These observations suggest that 3p and 9p loss likely precede k-ras or p53 mutations. Our preliminary data demonstrate that over-expression of a 31 kD tumor associated antigen recently purified, sequenced, and identified as heterogeneous nuclear ribonucleoprotein (hnRNP) A2 (with cross reactivity to splice variant B1), is expressed in most lung cancer cases before any morphologic abnormality. Comparison of the accuracy of this marker with sputum cytology will determine its value for early lung cancer detection. Preliminary evidence confirms this marker greatly improves the accuracy of standard sputum cytology for detection of lung carcinogenesis. Clinical intervention trials must be undertaken to determine whether modulation of hnRNP overexpression is useful as an intermediate endpoint for chemoprevention.