Early diagnosis of lung cancer, essential for reducing its high mortality rate, is currently challenging, partly due to the lack of specific biomarkers. Here, we attempted to develop a noninvasive and potentially sensitive screening method based on the proteomic analysis of unstimulated and stimulated saliva samples, collected by passive drooling and salivary swabs, respectively, from healthy heavy smokers enrolled in a nonprofit screening project. Protein content analyzed before and after sample cryopreservation for various periods and the associated two-dimensional electrophoresis revealed that protein extraction after short-term cryopreservation prevented the loss of detectable proteins. Mass spectrometric analysis of these electrophoretically resolved proteins revealed the presence of salivary proteins whose levels may be dysregulated in various types of lung cancer. Finally, in pilot experiments conducted on stimulated saliva from a patient with a lung cancer nodule, we detected altered content or selective presence of proteins involved in lung carcinogenesis, such as serpin B3 or the proteins S100A14 and aldoketoreductase-A1, respectively. While acknowledging that these findings require further validation, we believe that the use of saliva and related proteomic analyses may contribute to the identification of potential early lung cancer biomarkers, which could hopefully improve clinical management of the tumor and patient survival.