The current standard for treating stage II-III colorectal cancer involves surgical resection, often followed by adjuvant chemotherapy (ACT) for high-risk patients. However, identifying patients who would benefit most from ACT remains challenging. Molecular residual disease (MRD) in these patients is a critical factor in determining the likelihood of recurrence. This study aims to develop an exosome-based liquid biopsy assay using circulating exosomal microRNAs (exo-miRNAs) to predict the presence of MRD in postoperative patients. By analyzing exo-miRNAs from blood samples collected after surgery, the goal is to identify high-risk patients who may benefit from ACT while sparing low-risk patients from unnecessary treatment. A risk-stratification model, termed the exosome-based molecular residual disease prediction for adjuvant chemotherapy induction (EMRATI) score, will be developed to improve ACT decision-making and enhance clinical outcomes for stage II-III CRC patients.