Analytical strategies in early breast cancer diagnostic biomarker discovery by machine learning methods: Promises, advances and outlooks

Breast Cancer (BC) remains one of the leading causes of cancer-related mortality worldwide, with early detection playing a pivotal role in improving patient survival and treatment outcomes. Biomarkers serve as critical molecular indicators that facilitate the early diagnosis of breast cancer, allowing for timely intervention before the disease progresses to more advanced stages. Traditional methods for biomarker discovery, including immunohistochemistry, polymerase chain reaction (PCR), and enzyme-linked immunosorbent assays (ELISA), have been instrumental in identifying breast cancer markers. However, these approaches often require extensive validation, are time-consuming, and may lack the ability to effectively analyze high-dimensional datasets. The rapid advancements in machine learning (ML) have transformed biomarker discovery by enabling the analysis of complex multi-omics data, integrating genomic, proteomic, and imaging datasets to identify novel biomarkers with enhanced accuracy. This study focuses on the application of ML in detecting key biomarkers COL11A1, TOP2A, MMP1, and EZH2 which are associated with tumor invasiveness, proliferation, and metastatic potential in early-stage breast cancer. These biomarkers were identified through ML-based predictive models such as Random Forest (RF), Support Vector Machines (SVMs), XGBoost, and Deep Neural Networks, which have demonstrated superior performance in distinguishing malignant from benign cases. Our findings highlight the potential of ML-driven biomarker discovery in revolutionizing breast cancer diagnostics by improving risk stratification, enhancing predictive accuracy, and facilitating personalized treatment approaches. By leveraging AI-powered methodologies, clinicians can move toward a data-driven, precision medicine approach, ultimately reducing the burden of late-stage breast cancer diagnoses and mortality rates. However, integrating ML models into routine clinical practice requires addressing key challenges, such as data standardization, model interpretability, and validation through large-scale prospective studies. Future advancements in deep learning (DL), federated learning, and explainable AI (XAI) are expected to further refine these models, ensuring their reliability and applicability in clinical settings.