Toward a standardized and interoperable imaging biomarker catalog

Imaging biomarkers are quantifiable features extracted from medical images that indicate health status, disease characteristics, or treatment response. Their value depends on rigorous standardization and validation-efforts advanced by QIBA and EIBALL. Variability in existing inventories highlights the need for a FAIR (Findable, Accessible, Interoperable, Reusable)-compliant catalog to enable systematic discovery, comparison, and adoption.

We focused on defining the essential variables to describe imaging biomarkers across research, regulatory, and clinical settings. A three-phase approach was undertaken: (1) key resources-including FDA-NIH BEST, QIBA, EMA, ESR-EIBALL, relevant regulations, and scientific literature-were reviewed; (2) attributes were extracted and compared, with redundancies resolved by expert consensus; and (3) consolidated variables were organized into domains aligned with FAIR principles.

A unified biomarker descriptor set was established across five domains: core identification (imaging biomarker name, surrogation, clinical relevance), clinical context (main target, organ(s), disease/substrate, range(s), actionability), imaging and technical information (image modality, acquisition technique, technical parameters, extraction, association type, dimensionality, units), validation (robustness and use endorsed by publications, endorsed by professional societies, regulatory qualifications), and administrative data (repository, version/author). The catalog was tested across representative imaging biomarkers in inflammatory diseases, including ADC, FDG-PET, CT-based radiomics signatures, and Doppler ultrasound indices, demonstrating coherent descriptions across diseases and organs. Diagnostic and prognostic roles were clarified, transparency and reproducibility were promoted, and the need for context-adapted entries was shown.

This work proposes a harmonized, scalable approach for cataloging imaging biomarkers. Consistent descriptors across contexts facilitate integration into research, regulatory, and clinical workflows.

This work proposes a unified catalog structure for imaging biomarkers that improves standardization, comparability, and reusability across clinical, research, and regulatory domains.

Lack of standardization limits the integration of imaging biomarkers into research, regulatory, and clinical workflows. A harmonized catalog was developed using unified descriptors across the main identified domains. This structure enhances traceability, cross-disease comparability, and regulatory readiness of imaging biomarkers.