Dermatology Image Annotation FHIR Implementation Guide - Local Development build (v0.1.0) built by the FHIR (HL7® FHIR® Standard) Build Tools. See the Directory of published versions

SMART Imaging Access

SMART Imaging Access

This IG complements the SMART Imaging Access specification from the Argonaut Project. Together, the two specifications address the full clinical photography annotation pipeline for dermatology.

Complementary scope

SMART Imaging Access solves the image retrieval problem. It defines how applications access medical images from PACS, VNA, and other imaging archives through FHIR-based authorization and WADO-RS endpoints. A dermatology annotation application uses SMART Imaging Access to retrieve clinical photographs from institutional imaging systems.

This IG solves the annotation output problem. It defines how to represent structured annotation results as FHIR Observations and return them to the patient record. Once a clinician or AI system annotates a retrieved image, this IG provides the format for storing that result.

The annotation pipeline

The two specifications form a complete pipeline:

1. Retrieve image — The annotation application uses SMART Imaging Access to obtain the clinical photograph from the institution's imaging infrastructure.
2. Annotate — A clinician, AI model, or AI-assisted workflow classifies the image and produces a structured annotation.
3. Store result — The application writes a FHIR Observation conforming to this IG back to the EHR, linking the annotation to the source image through derivedFrom.

Linking annotations to source images

The derivedFrom element in DermatologyImageAnnotation and AIPrediction accepts references to ImagingStudy, DocumentReference, or Media resources. These are the same resource types that SMART Imaging Access uses to represent available images. This alignment means that the reference obtained during image retrieval can be used directly as the derivedFrom target in the annotation result, with no translation step required.

Regulatory context

The ONC Diagnostic Imaging Interoperability Request for Information (January 2026) signals regulatory interest in extending imaging interoperability beyond radiology. Dermatology is explicitly mentioned as a specialty where imaging workflows lack standardization. The combination of SMART Imaging Access for retrieval and this IG for annotation output addresses the full data lifecycle that ONC's RFI describes.

Implementation note

This IG does not depend on SMART Imaging Access, and neither specification requires the other. Implementations can adopt either independently. However, systems that implement both gain a standards-based pipeline from image retrieval through annotation storage.