Building the Foundations for Standardized Clinical AI: The SYMPHONY Contributions

SYMPHONY delivers publicly exploitable results that advance standardized clinical AI. It provides a reusable architecture unifying HL7 FHIR, OpenEHR, DICOM and IHE workflows, including a hybrid FHIR–OpenEHR model and standardized AI results via DICOM SR and segmentation. It adds an AI orchestration framework and a Feedback Loop API, validated across four clinical use cases, enabling scalable, compliant and vendor‑neutral AI workflows.

The SYMPHONY project delivers a set of publicly exploitable results that significantly advance standardized clinical AI integration. Its main outcome is the reusable architectural blueprint that unifies HL7 FHIR, OpenEHR, DICOM and IHE workflow profiles into a coherent, standards-based interoperability layer for hospitals. This includes the interface model, a hybrid FHIR–OpenEHR data strategy, and standardized AI result handling using DICOM Structured Reporting and Segmentation Objects. SYMPHONY also developed an AI workflow orchestration framework based on IHE AIW‑I, AIR and IID, and introduced a novel Feedback Loop API enabling clinician corrections to become ground‑truth annotations for continuous AI improvement. These assets were used across four real clinical use cases—Prostate Cancer, Abdominal Aortic Aneurysm, Atrial Fibrillation, and Multiple Sclerosis—demonstrating vendor‑neutral AI integration, patient‑engagement pathways, and chained AI inference. Collectively, these results provide a practical, standards‑aligned foundation that healthcare institutions and vendors can reuse to build scalable, compliant, and future‑ready AI-enabled clinical workflows.

Standardized protocols and profiles like DICOM, FHIR, openEHR, IHE

Different healthcare providers

Public deliverable