Abstract Background Lung cancer is the second most commonly diagnosed lung cancer in the world, and Non-Small Cell Lung Cancer is the most prevalent subtype. Tissue biopsy is the gold standard for detecting lung cancer, but is highly invasive as it necessitates the extraction of a sample of tissue for histologic analysis. It also carries risks of bleeding and/or infection, and is inconvenient from a patient perspective. The development of a minimally invasive test, such as one utilizing a blood or urine sample, capable of providing accurate results for lung cancer detection and/or subtyping, would significantly enhance the clinical landscape and streamline patient care. Methods In this study we utilize A549 and H1299 human lung cancer cell lines, differing in cell type, location within the lung, and genetic composition (Kras & p53 status), along with two additional cancer cell lines (HeLa/cervical cancer and SCBO/bladder cancer), as controls. Cell lysate and exosome proteomics data were obtained using data independent acquisition- liquid chromatography mass spectrometry (DIA-LCMS) and analyzed using ProteoDisco 2.0, a custom-designed software application in R. Raw mass spectra data was subject to quantile normalization to mitigate external, non-biological differences between the samples. Differential protein expression was performed and Student's t-test was used to identify significantly different proteins across the sample types. A correlation coefficient matrix and Principal Component Analysis were used to assess global differences, and hierarchical clustering and heat maps allowed for visualizing of specific groups of proteins that separated all sample types. Results Over 1000 proteins were identified in each of the exosome samples, and >7500 proteins were detected in the lysates. Correlation coefficient matrix and PCA demonstrate that exosomal protein profiles, relative to cell lysates, are better at differentiating between the two lung cancer cell lines and controls. Surprisingly, exosomal proteomic patterns are different for the two lung cancers, implicating activation of different signaling pathways. Specifically, hierarchical clustering identified protein clusters showing enrichment for multiple biological processes: H1299 cells manifest enrichment of apical junction proteins and in epithelial-mesenchymal transition (EMT) ontology, while A549 cell lines show enrichment in proteins involved in oxidative phosphorylation. This combination of functional pathways can be used as a biomarker signature to distinguish between these two lung cancer subtypes, and from controls. Conclusions A short list of exosomal proteins, comprising apical junction, EMT, and oxidative phosphorylation constituents can detect lung cancer and further distinguish between subtypes represented by A549 and H1299 lung cancer cells. This biomarker signature has potential for detecting and subtyping lung cancer, to allow for early detection of disease, development of personalized treatment strategies, and result in improved clinical outcomes for this deadly disease.