Challenges in Integrating Heterogeneous Biomedical Data for Precision Medicine: AI Models and Applications

This Research Topic focuses on using artificial intelligence (AI) to combine and analyze diverse biomedical data types to improve the performance and reliability of AI models in precision medicine. Precision medicine aims to tailor diagnosis, treatment, and prevention strategies to the unique characteristics of individual patients. Biomedical data comes from many sources and in many formats, such as: structured clinical data (e.g., lab test results, diagnostic codes); unstructured text (e.g., clinician notes), omics data (e.g Genomics / transcriptomics / proteomics); medical imaging (e.g., CT, MRI); patient-generated data (e.g., wearables, mobile health apps). These heterogeneous data sets can provide complementary insights, if they are effectively brought together and analyzed. In this context, computational methods, such as machine and deep learning, statistics and mathematical approaches, are critical because of their ability to learn patterns across large, complex datasets, handle high dimensionality, integrate data of different types, predict clinical outcomes or identify disease subtypes.The proposed topic presents several significant challenges. As said before, biomedical data are highly diverse, spanning structured clinical records, unstructured text, medical images, multi-omics data, and real-time signals from wearable devices. These data sources differ in scale, format, quality, and semantics, making data harmonization and interoperability difficult. Clinical datasets often contain missing values, noise, and biases, which can negatively affect AI model performance and generalizability. Another major challenge is the high dimensionality of omics and imaging data, which increases computational complexity and the risk of overfitting. Additionally, many AI models function as “black boxes,” raising concerns about interpretability, trust, and clinical adoption. Privacy, data security, and ethical issues—such as fairness and informed consent—also remain critical barriers.Despite these challenges, the primary goal of this research topic is to present AI methods capable of effectively integrating multi-modal biomedical data to generate accurate, interpretable, and clinically actionable insights. By leveraging heterogeneous data, AI aims to improve disease risk prediction, enable earlier and more precise diagnoses, identify patient subgroups, and optimize personalized treatment strategies. Ultimately, this research seeks to enhance patient outcomes, reduce trial-and-error medicine, and support data-driven clinical decision-making in real-world healthcare settings.Scope and themes:- Trustworthy models: explainability and interpretability- Machine learning and deep learning models- Generative AI models - Multiscale, Multimodal, Multi-institutional data- Multi-omics data analysis- EHR data analysis- Medical imaging and signal processing - Type of papers: methods, original research, review, systematic review, technology and code Topic editors: Massimo La Rosa, Antonino Fiannaca, Maurizio Giordano, Laura La Paglia, Mara Sangiovanni Source: https://www.frontiersin.org/research-topics/78300/challenges-in-integrating-heterogeneous-biomedical-data-for-precision-medicine-ai-models-and-applications