Agreement between intraoral photographs and clinical examination for assessing caries lesion activity using ICDAS in preschool children

Objectives: Accurate detection of caries lesion activity is essential for early intervention in preschool children. However, clinical visual-tactile examinations may be logistically challenging in large-scale or remote settings. This study aimed to evaluate the validity and reliability of intraoral photographic assessments in detecting caries lesion activity, compared to conventional clinical examinations, using the International Caries Detection and Assessment System (ICDAS) criteria. Methods: A cross-sectional study was conducted among 180 children aged 3–4 years attending public kindergartens in Vietnam. Each child underwent a full-mouth visual-tactile examination, followed by standardized intraoral photography. Caries lesions were scored using ICDAS codes 1–6 and classified as active (A) or inactive (I). Lesions were further categorized as non-cavitated (ICDAS 1–2) or cavitated (ICDAS 3–6). Diagnostic accuracy of photographic assessments was evaluated against clinical examination results using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), Cohen’s Kappa, and intraclass correlation coefficient (ICC). Results: A total of 1,910 teeth and 6,608 tooth surfaces were analyzed. At the tooth level, intraoral photographic assessments demonstrated high diagnostic performance: Non-cavitated lesions (ICDAS 1–2): Sensitivity 87.0%, Specificity 95.7%, Kappa = 0.789, ICC = 0.789 Cavitated lesions (ICDAS 3–6): Sensitivity 85.9%, Specificity 97.5%, Kappa = 0.706, ICC = 0.702. At the surface level, validity was moderate: Non-cavitated lesions: Sensitivity 25.5%, Specificity 81.2% Cavitated lesions: Sensitivity 73.6%, Specificity 83.8% Most discrepancies arose from early-stage (ICDAS 1–2) lesions, where subtle enamel changes were difficult to detect in two-dimensional images. Conclusions: Intraoral photographic assessments demonstrate strong agreement with clinical examination for detecting cavitated caries lesions at the tooth level in preschool children. Although less sensitive for early, non-cavitated lesions, this approach offers a practical and reliable tool for caries surveillance and teledentistry programs, especially in community or resource-limited settings.