Enhancing pancreatic cancer diagnostics: Ensemble-based model for automated urine biomarker classification.

This research addresses the critical challenge of early detection in pancreatic ductal adenocarcinoma (PDAC) by exploring urinary biomarkers and integrating artificial intelligence (AI) models. The study emphasizes the significance of liquid biopsy, particularly in urine, as a noninvasive approach for real-time monitoring of PDAC. Traditional diagnostic methods face limitations, necessitating the exploration of novel biomarkers and AI applications. The research employs two distinct methodologies: Method 1, utilizing a Bagged Ensemble of decision trees, and Method 2, employing an optimized decision tree model. Comprehensive evaluations, including accuracy, confusion matrices, and ROC curves, reveal the effectiveness of both methods in PDAC diagnostics. The comparison highlights Method 1's marginally superior performance in training, while Method 2 excels in distinguishing between benign and PDAC cases during testing. The research underscores the potential of AI-driven urinary biomarker classification as a robust diagnostic tool for improved outcomes in pancreatic cancer The evaluation results against state-of-the-art models, further emphasize the superiority of the proposed ensemble-based methods.