Radiomics in Action: Multimodal Synergies for Imaging Biomarkers

Abstract

Radiomics has recently begun as a transformative approach in medical imaging, shifting radiology from qualitative description to quantitative analysis. By extracting high-throughput features from CT (Computed Tomography), MRI (Magnetic Resonance Imaging), PET/CT (Positron Emission Tomography/Computed Tomography), and CBCT (Cone Beam Computed Tomography), radiomics enables the characterization of tissue heterogeneity and the development of imaging biomarkers with diagnostic, prognostic, and predictive values. This narrative review explores the historical evolution of radiomics and its methodological foundations, including acquisition, segmentation, feature extraction and modeling, and platforms supporting these workflows. Clinical applications are highlighted in oncology, cardiology, neurology, and musculoskeletal and dentomaxillofacial imaging. Despite being promising, radiomics faces challenges related to standardization, reproducibility, PACS/RIS (Picture Archiving and Communication System/Radiology Information System) integration and interpretability. Professional initiatives, such as the Image Biomarker Standardization Initiative (IBSI) and guidelines from radiological societies, are addressing these barriers by promoting harmonization and clinical translation. The ultimate vision is a radiomics-augmented radiology report in which validated biomarkers and predictive signatures complement conventional findings, thus enhancing objectivity, reproducibility, and advancing precision medicine.