Gastrointestinal (GI) cancers remain a major global health burden, largely due to the lack of effective and accessible early screening strategies. Current diagnostic approaches-including endoscopy, computed tomography (CT), and magnetic resonance imaging (MRI)-are either invasive, resource-intensive, or insufficiently sensitive for detecting early-stage disease, and are therefore not suitable for population-wide screening or for simultaneously identifying multiple GI tumor types. As a result, many patients are diagnosed at advanced stages, when therapeutic options are limited and prognosis is poor. Circulating microRNAs (miRNAs) offer a promising alternative, as they are stable in peripheral blood and reflect tumor-related molecular alterations. In this study, the investigators aim to develop and validate a robust, noninvasive miRNA-based signature capable of distinguishing GI cancers from non-malignant controls. By integrating multi-cohort datasets and applying machine learning-based feature selection and predictive modeling, the investigators will construct a screening panel optimized for reproducibility, scalability, and early-stage detection. This noninvasive miRNA signature has the potential to support accessible, cost-effective, and clinically practical population-level screening for GI cancers, ultimately facilitating earlier diagnosis and improving outcomes for participants.