Plasma proteomic profiles associated with low-dose computed tomography findings and incident lung cancer: a...

Plasma proteins may reflect tumor-host biology, but their relationships with low-dose computed tomography (LDCT) findings and incident lung cancer remain uncertain. Although LDCT screening has improved survival, challenges remain in risk stratification and interpretation of screening findings. The aim of the present study was to investigate associations between plasma proteomic profiles, LDCT findings and incident lung cancer, and to explore their potential for lung cancer risk stratification.

In this prospective study, baseline plasma proteomic profiles and LDCT radiographic findings (baseline and follow-up) were analyzed in 193 participants with significant smoking history. A targeted panel of 92 proteins was used to evaluate associations with LDCT classifications and incident lung cancer. Five participants with non-lung malignancies were excluded, leaving 188 participants, including six incident lung cancer cases. Exploratory prediction models for lung cancer were developed using machine learning approaches.

Individuals with positive LDCT findings showed overexpression of multiple plasma proteins compared with those with indeterminate or negative findings; participants who developed lung cancer also showed differential protein expression compared with participants without cancer. While individual protein odds ratio (OR) estimates were imprecise, a seven-protein machine learning model, consisting of TNFRSF9, FGF2, CCL20, CD4, MMP12, CXCL11, and CD40 demonstrated an area under the curve (AUC) of 0.915 [95% confidence interval (CI): 0.829-1.000] in predicting lung cancer. Proteomic profiling revealed associations of lung cancer with pathways, such as TNFα and IL6-JAK-STAT3 signaling, commonly linked to early carcinogenic processes.

Plasma proteomic profiling may provide complementary, biologically informative signals alongside LDCT in lung cancer screening. By integrating circulating biomarkers with imaging, this approach enables investigation of early disease biology in a pre-diagnostic setting, providing insights into systemic alterations that may precede clinically apparent cancer. While these findings are hypothesis-generating and do not establish the clinical utility of plasma proteomics in LDCT screening, these findings justify the need to evaluate this approach for additional risk stratification in LDCT lung cancer screening. Larger independent studies are required to determine whether plasma proteomics can improve risk stratification beyond LDCT and thereby reduce unnecessary follow-up examinations.