Νovel methylation biomarkers in liquid biopsy and classifying biosignatures for the clinical management of Breast Cancer

Background Despite significant advancements, Breast Cancer (BrCa) remains a devastating disease presenting emerging needs for effective management. Recently, increasing attention has been given to epigenetic biomarkers assessed in liquid biopsy for diagnostic and prognostic applications. This study introduces a 3-step data-driven biomarker discovery pipeline, to identify robust methylation biomarkers and generate high-performance biosignatures specific for clinically significant BrCa end-points, validated in patient cell-free DNA (cfDNA). Methods Publicly available genome-wide methylomes from 520 BrCa tissues and 185 benign breast were analyzed via an Automated Machine Learning (AutoML) platform to identify BrCa-specific methylation promoters, which were further assessed by bioinformatic tools to reveal BrCa biological relevance. Next, their methylation status was evaluated in plasma cfDNA from 195 BrCa patients and 135 healthy individuals by Methylation Specific qPCR (qMSP). Finally, AutoML was applied to experimental and clinical data to develop optimized classifying biosignatures for diagnosis, prognosis and prediction. Results AutoML identified 3 BrCa-specific methylated promoters in CLDN15 , MRGPRD and ZNF430 . There is yet limited knowledge for the implication of these genes in cancer biology, as shown by bioinformatic literature mining. Pathway analysis revealed implication with biological processes potentially related to carcinogenesis. Laboratory validation confirmed elevated methylation levels in BrCa patients for all 3 promoters, which were correlated with poor prognostic and predictive parameters. Classification analysis on experimental laboratory methylation measurements and patients’ clinical data built 4 specific models: a diagnostic biosignature distinguishing BrCa from healthy individuals (AUC 0.798), a classification biosignature differentiating BrCa subtypes (AUC 0.678), a prognostic biosignature predicting relapse (AUC 0.785), and a predictive biosignature for treatment response in metastatic patients (AUC 0.861). Conclusion Our data-driven pipeline successfully identified 3 BrCa specific methylation promoters, in genes not previously known to be related to BrCa. Their role in pathology needs further attention as they could also represent novel targets. Moreover, we built 4 biosignatures demonstrating strong predictive performance. The low number of features and the minimally invasive nature of liquid biopsy highlight the potential for clinical implementation of great value. This 3 step hybrid discovery and validation approach can be broadly applied across various cancer settings to provide readily available diagnostic solutions.