The Role of Real-World Data in Advancing Trustworthy and Unbiased AI-Powered Precision Oncology in Breast Cancer

Breast cancer patient management relies heavily on identifying molecular subtypes and tracking dynamic tumor evolution. Liquid biopsy (including circulating tumor cells, extracellular vesicles, and cell-free DNA) offers a minimally invasive window into this heterogeneity, providing real-time data on treatment-induced changes and residual disease. However, extracting clinically actionable insights from these high-dimensional, patient-derived data sources is difficult, particularly when analyzing routine, unselected clinical cohorts.While machine learning can identify patterns across these modalities, translating these computational models into routine practice remains a major hurdle. Models trained on highly curated clinical trial populations often fail to generalize to the clinic, and they can perpetuate diagnostic disparities if algorithmic bias is not addressed. This collection focuses on clinical translation: validating AI tools on real-world patient cohorts (RWD), mitigating algorithmic bias, and integrating blood-based biomarkers with standard clinical and pathological records.This Research Topic brings together clinicians, translational researchers, and computer scientists to accelerate the safe clinical deployment of AI in breast cancer. We welcome original research, reviews, and clinical studies on:- AI-driven analysis of circulating tumor cells (CTCs), including transcriptomic and phenotypic profiling- Machine learning for extracellular vesicle (EV) and cargo characterizationCirculating tumor DNA (ctDNA) and minimal residual disease (MRD) monitoring using AI in clinical cohorts- Multi-modal integration of liquid biopsy, digital pathology, and electronic health records- Validation of precision oncology AI models on unselected, real-world patient databases- Explainable AI (XAI) frameworks tailored for oncological decision-making- Algorithmic fairness and bias-mitigation strategies in diverse patient populations- Clinical utility and cost-effectiveness of AI-integrated biomarker platformsPlease note: Manuscripts consisting solely of bioinformatics, computational analysis, or predictions of public databases which are not accompanied by validation (independent clinical or patient cohort, or biological validation in vitro or in vivo, which are not based on public databases) are not suitable for publication in this journal. Topic editors: Michele Zanoni, Tania Rossi Source: https://www.frontiersin.org/research-topics/80081/the-role-of-real-world-data-in-advancing-trustworthy-and-unbiased-ai-powered-precision-oncology-in-breast-cancer