Noninvasive evaluation and clinical value prediction of tumor-infiltrating neutrophil-to-T-cell ratio in pancreatic ductal adenocarcinoma.

In pancreatic ductal adenocarcinoma (PDAC), the tumor immune microenvironment (TIME) is central to prognosis and therapeutic response, yet predictive biomarkers are still lacking. Here, we used PDAC patients from TCGA, CPTAC, and multiple centers to investigate and validate the relationship between the prognosis and tumor-infiltrating neutrophil-to-T-cell ratio (NTR) in the TIME, including the correlation of NTR with chemotherapy response in PDAC. We developed seven artificial intelligence models to evaluate NTR in the TIME, then evaluated their performance. The multi-modal ROI-model, designated PORCELAIN (Pancreatic Cancer Tumor-Infiltrating Neutrophil-to-T-Cell Ratio Evaluation with Artificial Intelligence), achieved superior external validation performance, demonstrating significant stratification of overall survival and recurrence-free survival. The NTR in the TIME is a potential biomarker for PDAC prognosis and treatment stratification. Importantly, PORCELAIN provides a noninvasive approach to assess NTR in the TIME, offering potential value for clinical management. Trial registration number: NCT06760234.