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

Home

Official URL: https://annotatemd.com/fhir/ig/ImplementationGuide/annotatemd.dermatology-annotation				Version: 0.1.0
Draft as of 2026-04-04				Computable Name: DermatologyImageAnnotationIG

Dermatology Image Annotation FHIR Implementation Guide

This Implementation Guide (IG) defines FHIR R4 profiles for representing structured dermatology image annotation results. It addresses a fundamental gap in clinical dermatology: the absence of a standard format for recording diagnostic observations that originate from clinical photography annotation workflows, whether performed by expert annotators, treating clinicians, or AI-assisted pipelines.

The IG provides a common data model that annotation platforms, EHR systems, and research tools can use to exchange dermatology image classification results as FHIR Observations. By encoding annotations with SNOMED CT findings, LOINC observation codes, and explicit method tracking, the IG enables interoperability between clinical care, quality assurance, and machine learning research.

Who this IG is for

• Clinicians and dermatologists who annotate clinical photographs or review AI-generated predictions
• Annotation platform developers who need a standard output format for structured annotation results
• EHR vendors who want to receive and display dermatology image classification data
• Machine learning researchers who require standardized training data with traceable provenance

Profiles

This IG defines three Observation profiles:

• DermatologyImageAnnotation captures a single annotation result from clinical photography, including the skin finding, annotation method, QA status, and confidence metrics.
• AIPrediction represents an AI model's classification output before clinician review, tagged with HL7 AI Transparency metadata and linked to the originating Device.
• AnnotationSeries groups annotations for the same lesion over time, enabling longitudinal tracking and change detection.

Status

This IG is a draft at version 0.1.0. It has not yet undergone HL7 ballot or connectathon testing.

Publisher: Annotate MD (fhir@annotatemd.com)

All canonical URLs under https://annotatemd.com/fhir/ig are stable and owned by Annotate MD. The IG package identifier is annotatemd.dermatology-annotation.

Annotate MD is pursuing HL7 endorsement for this IG through the appropriate work group process. Until endorsement is granted, this IG represents the publisher's specification and has no official HL7 status.

FHIR version and R5 migration

This IG targets FHIR R4 (4.0.1) because R4 remains the most widely deployed version across EHR systems. Organizations planning R5 migration should note that Observation.derivedFrom gains a broader set of reference targets in R5, and the new Observation.triggeredBy element may offer a cleaner pattern for linking AI predictions to the annotations they inform. The IG will publish R5 guidance when EHR adoption warrants it.