Diagnostic Accuracy of Teledentistry for Dental Caries Detection: A Systematic Review and Meta-Analysis

Background/Objectives: Image-based teledentistry can support remote caries screening, but diagnostic performance varies. This systematic review evaluated its accuracy against in-person dental examination. Methods: PubMed, Embase and CENTRAL were searched again on 22 September 2026. Fifty-three reports were retained: 43 diagnostic-accuracy and ten agreement/AUC-only reports described separately. Separate logit-scale multilevel random-effects models included 32 outcomes from 20 reports per endpoint. Moderator models excluding age included 26 outcomes from 16 reports and assessed dentition, image-acquisition role, and device. Risk of bias was assessed using QUADAS-2 and a separate adapted framework for agreement reports. Results: Summary sensitivity was 85.50% (95% CI 79.57–89.92%) and specificity was 92.25% (84.77–96.22%), with substantial heterogeneity (both Q tests p < 0.0001). The moderator omnibus test was significant for sensitivity (p = 0.0015), but not specificity (p = 0.9632). Intraoral camera use was associated with higher sensitivity than smartphone use after Holm adjustment (p = 0.0011). Dentition and acquisition-role contrasts were not significant after adjustment. Conclusions: Image-based teledentistry may support screening and triage. Mixed diagnostic units, heterogeneous thresholds, overlapping outcomes, sparse moderator categories and source-data limitations require cautious interpretation and prevent universal device or user rankings.