Machine Learning-Based Exosomal microRNA Signature for Preoperative Staging and Chemotherapy Eligibility in Colon Cancer
Recent studies have highlighted the potential benefits of neoadjuvant chemotherapy (NAC) in colon cancer; however, its indication is generally limited to cases corresponding to pathological stage IIB or higher. Accurately identifying such high-risk cases before surgery remains challenging using conventional clinical diagnostics alone. Therefore, we hypothesized that integrating molecular biomarkers with preoperative clinical assessment could provide a more precise and sensitive evaluation of tumor aggressiveness. In this context, we focused on exosomal microRNAs, which are actively secreted from tumor cells and remain stable in circulation, and aimed to develop a machine learning-based biomarker panel. To achieve this, we initiated a multicenter study utilizing preoperative plasma samples to establish a reliable biomarker model for risk stratification and treatment decision-making in colon cancer.
• Pathologically confirmed colon cancer (Stage I-IV, UICC TNM 8th edition)
• Underwent curative-intent resection (with or without perioperative therapy)
• Preoperative plasma (or serum) samples available
• Clinical and prognostic data available