Baseline IL-15 and dynamic MUC-16 predict PD-1 immunochemotherapy efficacy in advanced gastric cancer

Despite the emergence of immune checkpoint inhibitors (ICIs) combined with chemotherapy as the standard first-line treatment for advanced gastric cancer and gastroesophageal junction cancer (GC/GEJC), reliable biomarkers to predict treatment efficacy remain a critical unmet need. The profound heterogeneity of gastric cancer often leads to pervasive resistance, limiting the clinical effectiveness of PD-1-based immunochemotherapy. Identifying patients likely to benefit, and dynamically monitoring their response, is crucial for optimizing therapeutic strategies and avoiding unnecessary toxicity from ineffective treatments. This study addresses this gap by exploring plasma proteomic markers.

This prospective observational cohort study enrolled 31 patients with advanced gastric cancer receiving PD-1-based immunochemotherapy. Serial plasma samples were collected at baseline and after every two treatment cycles. Researchers performed longitudinal plasma proteomic profiling to identify dynamic protein changes. The primary endpoints included correlation with therapeutic response and overall survival. The findings for MUC-16 were further validated using conventional serum CA-125 testing to confirm the clinical relevance of the proteomic discoveries. A composite risk model was constructed by integrating multiple identified biomarkers.

Responders to PD-1-based immunochemotherapy exhibited significantly lower baseline levels of interleukin-15 (IL-15), with a P-value of 0.01. A substantial 161.9% reduction in plasma Mucin-16 (MUC-16), also known as Cancer Antigen-125 (CA-125), after just two cycles of treatment, was identified as an independent factor for prolonged overall survival. This finding showed a Hazard Ratio (HR) of 7.40 (95% CI 2.22-24.60, P = 0.001), and was independently validated by conventional serum CA-125 testing. Integrating baseline IL-15 with dynamic changes in MUC-16 and Matrix Metalloproteinase 12 (MMP12), a composite risk model was developed. This model achieved an Area Under the Curve (AUC) of 0.799 for predicting treatment response. Furthermore, this composite model was also an independent factor for progression-free survival, demonstrating an HR of 3.35 (95% CI 1.56-7.20, P = 0.002).