Prospective serial proteomic analysis uncovers mechanistic pathways of chemotherapy resistance in advanced non-small cell lung cancer.

Predicting chemotherapy response in advanced non-small cell lung cancer (NSCLC) remains a clinical challenge, as baseline profiles often fail to capture dynamic molecular adaptations under treatment. This prospective study employed serial plasma proteomics to identify mechanistic pathways associated with chemotherapy resistance in 44 patients with stage IV NSCLC receiving platinum-based doublet chemotherapy. By analyzing blood samples collected immediately before the first and second cycles using liquid chromatography-tandem mass spectrometry, we demonstrated that a ratio-based proteomic model (early-treatment/pre-treatment) yielded superior separation between controlled and uncontrolled disease (UCD) compared to baseline-only assessment. Among 159 quantified proteins, 13 showed significant differential abundance, with UCD patients exhibiting marked upregulation of tetranectin, coagulation factor XIII A chain, and complement factor H-related protein 2. Ingenuity Pathway Analysis revealed that therapeutic resistance was characterized by three dominant axes: the activation of complement-coagulation-acute-phase signaling, the induction of lipid-nuclear receptor activity (LXR/RXR and DHCR24 signaling), and the relative attenuation of immune-regulatory pathways such as IL-12 signaling. These findings highlight the potential of serial proteomic profiling to uncover treatment-induced molecular adaptations, providing insights for therapeutic monitoring and hypothesis generation in precision oncology. SIGNIFICANCE: This study demonstrates the added value of prospective serial plasma proteomic profiling, compared with baseline-only approaches, for capturing early treatment-associated molecular adaptations in advanced non-small cell lung cancer (NSCLC) receiving chemotherapy. By quantifying proteomic changes between pre-treatment and early-treatment time points, we identified coordinated alterations involving the complement-coagulation-acute-phase axis and lipid-nuclear receptor signaling programs, including LXR/RXR and DHCR24, alongside relative attenuation of immune-regulatory pathways. Rather than reflecting isolated protein effects, these findings highlight interconnected host-tumor response programs that emerge under therapeutic pressure and may contribute to early adaptive resistance. Importantly, this work moves beyond static baseline markers by emphasizing dynamic, pathway-level changes and provides a hypothesis-generating framework for longitudinal therapeutic monitoring. Candidate proteins such as tetranectin and coagulation factor XIII A chain are proposed as molecular features associated with treatment response, warranting further validation in larger, prospective cohorts before translational application.