BackgroundNon-small cell lung cancer (NSCLC) has a complex etiology, making early diagnosis difficult and leading to high mortality rates, thus necessitating personalized treatment strategies. While protein level ratios have shown potential as biomarkers or therapeutic targets, their causal relationship with NSCLC remains unclear. This study aimed to investigate these causal links using Mendelian randomization (MR), providing insights into potential biomarkers and therapeutic avenues.MethodsWe executed an intricate two-sample MR study to explore the stochastic causal links between 2,821 protein level ratios and NSCLC. The genome-wide association study (GWAS) statistics for NSCLC and protein level ratios were sourced from the Finnish Database (version 10) and the UK Biobank, respectively. For the instrumental variables (IVs) related to protein level ratios, we selected IVs with a P value <1.0×10-5. Throughout this analysis, we applied five established MR techniques.ResultsOur study identified causal relationships between 142 protein level ratios and NSCLC. Notably, the AKR1B1/SUGT1 protein level ratio and the PLPBP/STIP1 protein level ratio demonstrated the most significant negative correlations with NSCLC risk. On the other hand, the ARHGEF12/IRAK4 protein level ratio and the BANK1/LBR protein level ratio exhibited the most significant positive correlations. Furthermore, sensitivity analyses did not reveal any significant heterogeneity or horizontal pleiotropy.ConclusionsStudying specific protein level ratios in patients can reveal the molecular mechanisms and pathological processes of NSCLC, which has certain clinical significance for early diagnosis of NSCLC, understanding drug resistance mechanisms and developing personalized treatment strategies. However, these findings necessitate further validation through extensive clinical research.