Brain metastases (BMs) are a common complication in patients with EGFR-mutant lung adenocarcinoma, with a high incidence and a close association with poor prognosis. Identifying high-risk populations for brain metastasis and carrying out effective intervention is of great significance for reducing or delaying the occurrence of BM. Effective tools for assessment are currently lacking, and given the high incidence of BM, there is an urgent need to develop new biomarkers.

A retrospective analysis was performed on 56 subjects enrolled at the First Affiliated Hospital of Guangzhou Medical University between 2012 and 2018; all patients had EGFR-mutant lung adenocarcinoma (LUAD) that recurred after radical resection. Patients were divided into a brain metastasis group (BM, n=28) and a non-brain metastasis group (NBM, n=28), and tissue proteomics testing was performed using LC-MS/MS technology. To identify potential markers predicting LUAD brain metastasis, comparative analyses were conducted between the different groups to evaluate proteins associated with brain metastasis.

The study showed that there was no significant correlation between clinical factors of recurrent patients and the occurrence of brain metastasis. At the proteomic level, differentially expressed proteins between the brain metastasis group and the non-brain metastasis group were identified, and eight proteins with the best performance in distinguishing the two groups were selected (RRS1, CPT1A, DNM1, SRCAP, MLYCD, PCID2, IMPAD1 and FILIP1). Based on the random forest algorithm, an optimal model was constructed using these eight proteins, with an ROC value reaching 0.9401, indicating that the model performs excellently in distinguishing the BM group from the NBM group.

To further analyze the potential mechanisms underlying the occurrence of brain metastasis, GO enrichment analysis was performed on the differentially expressed proteins of the two groups. The results showed that the BM group mainly exhibited upregulation of proteins involved in lipid metabolism, while the expression of proteins related to cell cycle pathways was downregulated. Subgroup analysis showed that the protein characteristics of patients receiving different adjuvant treatment regimens after surgery differed significantly. In brain metastasis patients receiving chemotherapy, proteins involved in chromosome assembly and nuclear-related pathways were significantly downregulated, while proteins involved in lipid metabolism and immune-related pathways were upregulated. In brain metastasis patients receiving TKI treatment, upregulated proteins were mainly enriched in pathways related to membrane alterations such as cell adhesion, membrane organization and inflammatory response, while downregulated proteins were enriched in pathways related to metabolites.

This study developed a model based on protein biomarkers that can accurately predict the risk of postoperative brain metastasis in patients with EGFR-mutant lung adenocarcinoma, providing important reference value for postoperative patient management.

At present, the vast majority of LC-MS/MS clinical testing methods are laboratory-developed methods. Because laboratories use different instruments, reagents and standards, different detection parameters need to be set, resulting in great differences among laboratory-developed testing methods and bringing substantial difficulties to the standardization of LC-MS/MS testing.

The approval of several of ProteinT’s sample pretreatment kits has resolved the drawbacks of the lack of standards and quality control in the LC-MS/MS clinical testing process, greatly improving the above obstacles and comparability issues encountered by different laboratories when using mass spectrometry to detect tissue proteins. This helps to further improve the clinical application status of LC-MS/MS methods and accelerate the clinical translation of proteomics.