Prognostic Clinical Predictive Models for Dental Caries Using Artificial Intelligence: Methodological Considerations

Artificial intelligence (AI) is currently used to develop clinical predictive models for dental caries. However, most prognostic models lack key methodological components. Few have been implemented in clinical practice, and even fewer have demonstrated clinical benefits. This narrative methodological review examines considerations for developing, validating, and implementing AI-based caries prognostic models. Based on established prediction model frameworks (TRIPOD+AI, PROBAST+AI, PROGRESS), we contextualize eight critical phases for caries-specific application: (1) problem selection and clinical purpose, (2) data quality and preparation, (3) study design, (4) model development, (5) validation, (6) performance assessment, (7) transparent reporting, and (8) deployment and maintenance. Addressing these steps is essential to reduce bias, improve reproducibility, and support meaningful evaluations of the clinical impact. AI-based prognostic models should be used as decision-support tools that inform clinician and patient choices rather than substitutes for clinical judgment.