Exosome-based microRNA Biomarkers for Non-invasive and Early Detection of Colorectal Cancer and Advanced Adenomas

Principal Investigator

Name
Ajay Goel

Degrees
Ph.D.

Institution
City of Hope

Position Title
Professor and Chair, Department of Molecular Diagnostics and Experimental Therapeutics

Email
ajgoel@coh.org

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
2025-0016

Initial CDAS Request Approval
Jun 16, 2026

Title
Exosome-based microRNA Biomarkers for Non-invasive and Early Detection of Colorectal Cancer and Advanced Adenomas

Summary
Colorectal cancer (CRC) despite being, potentially preventable, still ranks third in diagnosis and second in mortality worldwide. CRC screening trials based on fecal immunochemical testing (FIT) showed survival benefits through a stage-shifting effect, reducing stage IV diagnoses but not preventing CRC incidence. In fact, FIT specificity and sensitivity decline for early-stage CRC and precursor lesions, emphasizing its impact on mortality rather than incidence. Endoscopy-first approaches have a higher sensitivity, but the contentious results of the recent NordICC trial have yet again highlighted the challenges of such an approach, including invasiveness, high costs, and poor patient compliance.

Non-invasive CRC screening tests promote compliance and participation in screening but still suffer from adequate sensitivity to detect the earliest stages of CRC (stage I) and advanced adenomas (AAs). Notably, these screening assays have been offered to the public under the assumption that CRC-derived analyte(s) would detect AA, too. However, these approaches had two limitations: 1) if both CRC and AA release the same analyte, AAs do so in considerably smaller quantities, hence the lower sensitivity values, and 2) AA and CRC represent two opposite ends of the adenoma-carcinoma sequence continuum, therefore they may not yet release the same analytes. Therefore, there is an imperative need to develop innovative liquid biopsy assays that can robustly detect both AAs and early-stage CRCs to optimize patient compliance, resource allocation, and timely identification and prevention of colorectal neoplasia.

MicroRNAs (miRNAs) are small non-coding RNAs that regulate every human cancer, including CRC, and may thus be ideal biomarkers. Indeed, circulating cell-free miRNAs (cf-miRNAs) have been shown to have diagnostic potential. Furthermore, the recent discovery that cancer cells actively excrete miRNAs in small extracellular vesicles called exosomes (exo miRNAs) has revolutionized the field, as tumor-derived exosomal cargo enables the identification of cancer-specific molecular markers. To address the limitations of cf-miRNA biomarkers and to interrogate the potential clinical significance of exo-mRNAs, during the previous cycle of funding (1R01CA227602-01A1), we completed a systematic and comprehensive biomarker discovery and validation effort in patients with CRC, AAs, and non-disease control participants.

Utilizing the power of unbiased and genome-wide sequencing-based miRNA profiling approaches, together with rigorous bioinformatics and machine-learning algorithms, we: 1) identified panels of 19 cf-miRNAs and 20 exo-miRNAs that could robustly identify patients with early-stage CRC (area under the curve [AUC] = 0.92 and 0.95 for cf- and exo-miRNAs); 2) combined these into a “transcriptomic signature” for CRC (AUC = 0.98); 3) identified panels of 19 cf-miRNAs and 20 exo-miRNAs that could robustly identify patients with AAs (area under the curve [AUC] = 0.95 and 0.90, respectively) and 4) combined them into a highly accurate transcriptomic signature (AUC = 0.97); and 5) begun validation in several cohorts, including a cohort of early-onset CRC vs. age- and sex-matched controls (AUC = 0.98) demonstrating the robustness of our approaches in several different age cohorts, and 6) produced preliminary data in matched pre- post- surgical samples demonstrating that the our assay turns negative at the 4th-5th day after surgery.

Aims

Of note, we hold patent-pending intellectual property for our transcriptomic signature with the USPTO. These successes collectively demonstrate our ability to establish a transcriptomic signature for the early detection of CRC/AA.
In this renewed cycle of this funding, we are building upon our previous success, and our large-scale validation demonstrates the clinical significance of our transcriptomic signature in real-life clinical settings. We are evaluating its performance as a non-invasive assay in large, racially/ethnically diverse, prospective cohorts of CRC, AA, and NDC patients with various risk profiles; we are currently deciphering its diagnostic potential for detecting AAs with high-grade dysplasia (HGD) or invasive disease and, important to this proposal, we intend to use it to determine lead-time for disease development in pre-diagnosis specimens from individuals who later developed CRC. The current four milestone-based specific aims of this proposal were:

Aim #1: Expand our biorepository via continued prospective enrollment of patients with CRC, AA, Low-Risk Adenomas, and Non-Disease controls, including those with a familial risk, with an additional focus on enrollment of and specimen collection from patients of racial/ethnic minority populations, to support our project.

Aim #2: Further validate the transcriptomic signature and establish its performance in prospective cohorts of patients with early-onset CRC (ie, younger than 50), in light of the recent increasing trends for CRC incidence.

Aim #3: Evaluate the ability of our transcriptomic signature to detect CRC and AA at its earliest stages in pre-diagnosis serum specimens and to determine lead time before disease presentation. This specific aim pertains to the use of PLCO samples; we plan to evaluate the performance of our signature alone or in combination with other markers in longitudinally collected samples from patients with CRC/AA that include pre-, at-, and post-diagnosis time points and use it to determine the lead time before disease presentation.