Machine-Learning-Driven Phenotyping in Heart Failure with Preserved Ejection Fraction: Current Approaches and Future Directions

Heart failure with preserved ejection fraction (HFpEF) remains a major clinical challenge due to its heterogeneous presentation and limited therapeutic options. Accurate patient phenotyping is essential to improve diagnosis, prognostication, and treatment personalization. Machine learning (ML) has emerged as a powerful tool to identify clinically meaningful HFpEF subgroups by integrating diverse data sources, including clinical, imaging, biomarker, and physiological parameters. ML-based models can uncover subtle patterns not captured by traditional methods, offering improved risk stratification, earlier intervention, and guidance toward individualized therapy. Future progress will rely on standardized data collection, validation across populations, and incorporation into clinical decision support systems. Advancements in explainable artificial intelligence, federated learning, and multi-omics integration are expected to further refine phenotyping strategies and translate into improved patient outcomes. Continued interdisciplinary collaboration is essential to unlock the full potential of ML in transforming HFpEF management.