Background: Non-small cell lung cancer (NSCLC) represents nearly 85% of lung cancer cases globally and remains a major cause of cancer mortality. Although immune checkpoint inhibitors improve survival, durable responses occur in a limited number of patients, highlighting the urgent need for reliable, minimally invasive predictive biomarkers. Objective: To evaluate whether specific serum protein signatures can predict clinical response to immune checkpoint inhibitors in patients with NSCLC. Methods: This prospective cohort study was conducted at National Institute of Cancer Research and Hospital, Dhaka, Bangladesh from January 2024 to December 2024. A total of 67 histologically confirmed NSCLC patients receiving ICI therapy were enrolled using purposive sampling. Baseline serum samples were obtained before treatment and analyzed for selected protein biomarkers related to immune regulation and inflammation. Clinical response was evaluated at 12 weeks using the RECIST criteria. Data were processed in SPSS 23.0, applying logistic regression and ROC curve analyses. Results: Of the 67 patients, 29 (43.3%) achieved partial response or stable disease, whereas 38 (56.7%) experienced disease progression. Higher baseline pro-inflammatory protein levels were significantly associated with poor response (p<0.05), while elevated immune-activating proteins predicted favorable outcomes (p<0.01). The combined serum protein signature showed strong predictive accuracy (AUC=0.81; 95% CI: 0.70–0.91). Conclusion: Distinct serum protein signatures may serve as non-invasive predictive biomarkers for clinical response to ICIs in NSCLC patients. These findings support further validation in larger, multicenter studies to optimize personalized immunotherapy strategies.