Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
April 18, 2026ACS OmegaOpen Access

Machine Learning-Assisted Plasma Metabolomics Identifies a Five-Metabolite Panel for Colorectal Cancer Detection

View Full Paper
Ask AI
Bookmark
Share

Authors

JWJun-Kai WongCLChung‐Hsien LinHWHsin-Yi Wu

Discussion

Loading...

Member takes

Overview

Integrated metabolomics identifies a five-metabolite panel for CRC detection, suggesting a new screening tool.

Key Points

  • The aim is to identify a metabolite panel for noninvasive colorectal cancer detection.
  • Applied an integrated plasma metabolomics workflow with untargeted and targeted analyses.
  • Analyzed plasma samples from CRC patients and healthy controls using liquid chromatography-mass spectrometry.
  • Used machine learning, specifically random forest, to prioritize features from 146,880 spectral outputs.
  • Evaluated diagnostic performance with a logistic regression model.
  • Five specific metabolites were consistently reduced in CRC patient plasma across two cohorts.
  • The five-metabolite panel achieved an ROC AUC of 0.968 with 97.9% sensitivity and 89.4% specificity.
  • The panel showed robustness across disease stages and age groups.
  • In vitro assays indicated modulation of CRC cell migration and invasion by the selected metabolites.

Cite This Study

Wong et al. (2026) studied this question.

synapsesocial.com/papers/69e3203440886becb653f484https://doi.org/10.1021/acsomega.6c00857
View Full Paper
Ask AI
Bookmark
Share