Screening and identification of novel protein markers of early-stage lung cancer and construction and application of screening models.

Objective: Molecular biomarkers have the potential to improve the current state of early screening of lung cancer. This investigation aimed to identify novel protein markers for early-stage lung cancer and combine them with traditional tumor markers to develop machine learning models for lung cancer screening. Materials and methods: The protein alters of peripheral blood (5 patients with early-stage lung adenocarcinoma, 5 patients with early-stage lung squamous cell carcinoma, and 8 healthy controls) were detected by label-free quantitative proteomics. The novel candidate protein markers were preferentially selected by multi-omics technology. Then, the malignant transformation of BEAS-2B cells and lung carcinogenesis in C57BL/6 mice were induced by coal tar pitch extracts (CTPE) so that the expressions of these markers at different stages of lung carcinogenesis could be dynamically tracked and validated. These markers in human plasma were detected and further confirmed by ELISA. Machine learning models were established to screen high-risk individuals of lung cancer. Results: The C-type lectin domain family 3 member B (CLEC3B), membrane primary amine oxidase (AOC3), hemoglobin subunit beta (HBB), catalase (CAT), and selenoprotein P (SEPP1) were screened as candidate protein markers for early-stage lung cancer. The expressions of CLEC3B, AOC3, CAT, and SEPP1 were statistically significant in various passages of cells cultured with exposure to CTPE compared to the saline group ( Conclusion: The differential changes of AOC3, CAT, CLEC3B, SEPP1, and HBB protein were proven as early molecular events in lung tumorigenesis. The screening models of lung cancer based on the novel protein markers and traditional tumor markers might be applied for the screening of high-risk individuals.