Artificial Intelligence and Imaging Technologies in Orthodontic Assessment, Referral and Care Pathways: A Scoping Review

Abstract* Background Artificial intelligence (AI), integrated with imaging technologies, is increasingly applied across orthodontic assessment and care pathways. However, the evidence remains scattered across imaging modalities, clinical applications, AI methodologies, and stages of clinical translation. Methods A scoping review was conducted in accordance with the Joanna Briggs Institute methodology and reported in line with the PRISMA-ScR guidelines. PubMed, Scopus, and IEEE Xplore were searched for studies published between January 2016 and June 2026, with searches supplemented by Google Scholar, OpenAlex, grey literature sources, and citation searching. Eligible studies were mapped descriptively and synthesised thematically using Braun and Clarke’s framework. Results Fifty-four studies were included. AI applications used intraoral and facial photographs, cephalometric radiographs, CBCT, intraoral scans, three-dimensional dental models, and multimodal imaging. Applications ranged from landmark localisation, segmentation, and malocclusion classification to treatment-need assessment, screening, referral support, treatment planning, and remote monitoring. Evidence was strongest for bounded, anatomically verifiable image-analysis tasks, whereas applications involving referral, triage, treatment-need decisions, longitudinal prediction, and workflow integration were less mature. Common limitations included retrospective, single-centre designs; heterogeneous reference standards; limited external validation; inconsistent reporting of calibration and uncertainty; and insufficient evaluation of patient-centred, equity, and service-level outcomes. Conclusions Imaging-based AI in orthodontics has evolved from automated image interpretation to clinician-assisted decision support across the care pathway. Current evidence most strongly supports AI for structured image-analysis tasks, but does not yet support autonomous referral, triage, or treatment-planning decisions. Future research should prioritise prospective multicentre validation, transparent reporting, explainability, fairness, uncertainty handling, and workflow-integrated evaluation to enable safe and clinically useful implementation.