Machine Learning-Based Exosomal microRNA Signature for Preoperative Staging and Chemotherapy Eligibility in Colon Cancer

Colon cancer remains one of the leading causes of cancer-related mortality worldwide, and despite advances in screening and surgical techniques, a substantial proportion of patients continue to experience disease recurrence after curative resection. For patients with pathological stage IIB or higher disease, adjuvant chemotherapy is routinely recommended due to their elevated recurrence risk. However, accurately identifying these biologically aggressive cases before surgery remains a major clinical challenge, as current imaging-based staging often underestimates tumor burden and fails to capture underlying malignant potential. This diagnostic gap has hindered the optimal implementation of neoadjuvant chemotherapy (NAC) in colon cancer.

To address this issue, the investigators established the EXPOSE study (Exosomal microRNA Signature for Pre-Operative Stage and Eligibility Evaluation), a multicenter translational research initiative aiming to develop and validate a noninvasive, biologically informed diagnostic assay capable of identifying patients with high-risk colon cancer-equivalent to pathological stage IIB or higher-who may benefit from NAC.

The EXPOSE study will proceed through three structured phases.

1. In the discovery phase, exosomal microRNAs will be profiled using comprehensive small RNA sequencing to identify key biomarkers reflecting tumor aggressiveness.
2. In the training phase, the investigators will quantify candidate microRNAs using RT-qPCR and integrate their expression patterns via machine-learning algorithms to construct a predictive model for high-risk disease.
3. Finally, in the validation phase, the model's reproducibility, diagnostic accuracy, and generalizability will be tested in an independent clinical cohort.

The final EXPOSE assay is expected to serve as a liquid biopsy-based tool for preoperative staging, enabling more precise identification of biologically advanced colon cancer cases. Upon completion, this study will deliver a rigorously validated diagnostic model that combines molecular and clinical data to guide neoadjuvant treatment decisions and enhance personalized care in colon cancer management.