About the Article
Article Files
[This article belongs to Volume - 70, Issue - 6]
Published on : 2025-06-17 14:47:10
Article Code: AMJ-17-06-2025-12322
Title : Circulating Cell-Free DNA Methylation Signatures as Multi-Cancer Early Detection Biomarkers: A Machine Learning-Integrated Liquid Biopsy Platform
Author(s) : Dr. Robert Kim, Dr. Fatima Al-Rashid, Dr. David Okafor
Abstract :
Early detection of malignancies remains the cornerstone of improving oncologic outcomes. We developed a
comprehensive cell-free DNA (cfDNA) methylation profiling assay coupled with deep neural network architecture
(CancerGuard-AI) capable of simultaneous detection and tissue-of-origin localization across 12 solid tumor types.
Analyzing 2,500 plasma samples from stage I-III cancer patients and healthy controls, our platform achieved 94.3%
sensitivity at 98.1% specificity, outperforming existing multi-cancer early detection (MCED) assays. Differential
methylation patterns at CpG islands associated with SEPT9, SHOX2, and PCDHGB6 demonstrated high discriminatory
power for pancreatic and ovarian cancers—traditionally late-diagnosed pathologies. This non-invasive, cost-effective
screening paradigm represents a transformative approach to population-scale cancer surveillance.