AI-Assisted Panoramic Radiography Enhances Caries Detection by Dentists

Objectives : To evaluate the impact of a U-Net-based artificial intelligence (AI) system on caries detection in panoramic radiographs, with specific focus on enhancing diagnostic performance for junior clinicians. Materials and Methods: In this retrospective multicenter randomized crossover trial, 15 dentists (12 juniors with ≤3 years experience, 3 seniors with >10 years experience) interpreted 402 panoramic images under four conditions: AI-assisted, unassisted, AI-only, and expert reference. Primary endpoints included tooth-level sensitivity/specificity; secondary outcomes comprised interpretation time, case-level metrics, predictive values, and area under curve (AUC). Ground truth was established by three senior experts using pixel-wise annotations. Results: AI assistance significantly improved junior dentists’sensitivity (Δ=15.0%, 82.4% vs 67.4%; 95% CI: 7.3-22.7%, P=0.0004) while maintaining specificity (97.4% vs 97.2%). Interpretation time decreased by 22.7% (50.37s vs 65.12s, P=0.003). Case-level analysis showed improved sensitivity (93.3% vs 84.7%, P=0.0013) and negative predictive value (93.9% vs 88.0%, P=0.008). The standalone AI achieved 79.2% sensitivity/98.4% specificity (AUC=0.938), outperforming unassisted juniors in both metrics (P<0.05). Conclusions: The dual-stage U-Net system demonstrated clinically meaningful improvements in caries detection accuracy and workflow efficiency, particularly benefiting less experienced practitioners. Clinical Relevance: This AI-assisted approach shows promise for standardized caries screening in resource-variable clinical settings, potentially reducing diagnostic disparities between junior and senior practitioners.