Overview
EN ISO 18374:2025 - "Dentistry - Artificial intelligence (AI) and augmented intelligence (AuI) based 2D radiograph analysis - Data generation, data annotation and data processing" - defines requirements for developing and documenting goals, limitations, target end users and target patient populations for AI/AuI software that analyses 2D dental radiographs. The standard focuses on static (non‑dynamic) AI/AuI and sets out expectations for appropriate training, validation and test datasets and for consistent data annotation and processing. It does not prescribe implementation-level algorithms or software architecture.
Key topics and requirements
- Scope and purpose: Clear definition of intended use, clinical goals, target users and patient populations.
- Dataset requirements: Guidance on quantity, scope and handling of training, validation and test data to support model claims.
- Data quality control: Procedures and criteria to ensure image quality, consistency and suitability for model development.
- Bias analysis factors: Identification and documentation of factors that may introduce bias (e.g., demographic or acquisition differences).
- Annotation (labelling):
- Annotation strategy and procedures
- Competency requirements for annotators and reviewers
- Reproducibility and inter‑annotator agreement
- Data pre‑ and post‑processing: Recommended controls and documentation for image preprocessing steps and output handling, plus assessment of associated risks.
- Data protection and privacy: Requirements for protecting patient data and demonstrating appropriate controls.
- Test method and reporting: Defined test data criteria, inclusion/exclusion rules, validation processes and required contents of test reports and electronic instructions for use.
Applications and who should use it
EN ISO 18374:2025 is directly relevant to:
- Dental AI/AuI developers building diagnostic or decision‑support tools for 2D radiograph analysis.
- Medical device manufacturers producing imaging software for dentistry.
- Clinical researchers and data scientists preparing datasets or performing validation studies.
- Regulators and conformity assessment bodies assessing claims and evidence for dental AI products.
- Dental clinics and health IT teams procuring or integrating AI tools and evaluating vendor documentation.
Practical use cases include clinical decision support, automated image triage, measurement/annotation tools, and quality control of radiographic imaging workflows.
Related standards (applicable)
This standard complements broader medical device, data protection and AI governance frameworks - for example, standards and regulations addressing medical device quality management, clinical evaluation and data privacy - and should be used alongside applicable national/regional regulatory requirements.
Keywords: EN ISO 18374:2025, dentistry, AI, AuI, 2D radiograph analysis, data annotation, data processing, training data, validation data, bias analysis, data protection.
