Molecular biology · Bioinfo & AI
Integrative multi-omics profiling and machine learning reveal enhancer RNA signatures for early detection and prognostic stratification in breast cancer
Summary
This study investigated the potential of enhancer RNAs (eRNAs) as biomarkers for breast cancer, complementing traditional protein-coding transcripts. By profiling eRNA expression across over 1000 TCGA-BRCA samples and healthy tissues, researchers developed a 10-eRNA signature that effectively stratified patients into high- and low-risk groups, demonstrating independent prognostic value. Furthermore, a 19-eRNA panel was designed for early detection, showing excellent performance on an independent cohort. These tools have been integrated into eRNACare, a publicly accessible web application, highlighting the promising role of eRNAs for breast cancer risk stratification, therapeutic guidance, and early diagnosis.
Critical appraisal
Clinical. This research presents promising eRNA biomarkers for prognostic stratification and early detection in breast cancer. The 10-eRNA signature demonstrated independent prognostic value (HR = 3.176, 95% CI: 2.075-4.864, P < .001) with a 3-year AUC of 0.743, suggesting potential for identifying high-risk patients who may benefit from more intensive monitoring or therapies. The 19-eRNA panel for early detection achieved an AUC of 0.959 on an independent cohort, which could justify clinical trials to validate its use as a non-invasive screening tool. These findings could, in the long term (3-5 years for validation, 5+ years for clinical integration), refine patient management strategies and potentially alter screening practices.
Molecular biology. The study discovered eRNA signatures as a novel, largely unexploited layer of noncoding biomarkers for breast cancer. By profiling eRNA expression in TCGA-BRCA and GTEx samples, researchers identified prognostic eRNAs linked to key regulatory axes like ESR1/FOXA1/GATA3/AR, providing mechanistic insights into tumor progression. The multi-omics integration strengthens the biological validity of these eRNAs. Validation in independent cohorts and their connection to known molecular pathways pave the way for RNA-based clinical tests for diagnosis and prognosis, potentially via liquid biopsies or RNA sequencing panels.
Bioinfo & AI. The bioinformatic approach is robust, combining multi-omics profiling with advanced machine learning techniques. For prognostic stratification, LASSO-Cox regression was employed, while for early detection, bootstrap LASSO feature selection was followed by benchmarking ten classifiers, with a Multilayer Perceptron achieving the best generalization (AUC = 0.959). The use of large datasets (1073 TCGA-BRCA samples, 179 GTEx tissues, independent GSE225846 cohort) ensures good statistical power. The deployment of these modules in eRNACare, a publicly accessible web application, is an excellent example of technology transfer, allowing clinicians and researchers to access individualized survival and tumor probability scores, which could facilitate the integration of these biomarkers into clinical workflows.
The summary and the appraisal were written by a language model from the title and abstract of the original article. A curator decided to publish it. Methodology