Accuracy of dental radiographs for caries detection.

Data sourcesMedline, Embase, Cochrane Central and grey literature, complemented by cross-referencing from bibliographies. Diagnostic reviews were searched using the Medion database.Study selectionStudies reporting on the accuracy (sensitivity/specificity) of radiographic detection of primary carious lesions under clinical (in vivo) or in vitro conditions were included. The outcome of interest was caries detection using radiographs. The study also assessed the effect of the histologic lesion stage and included articles to assess the differences between primary or permanent teeth, if there had been improvements recently due to technical advances or radiographic methods, or if there are variations within studies (between examiners or applied radiographic techniques).Data extraction and synthesisData extraction was done by one reviewer first, using a piloted electronic spreadsheet and repeated independently by a second reviewer. Consensus was achieved by discussion. Data extraction followed guidelines from the Cochrane Collaboration. Risk of bias was assessed using QUADAS-2. Pooled sensitivity, specificity and diagnostic odds ratios (DORs) were calculated using random effects meta-analysis. Analyses were performed separately for occlusal and proximal lesions. Dentine lesions and cavitated lesions were analysed separately.Results947 articles were identified with the searches and 442 were analysed full text. 117 studies (13,375 teeth, 19,108 surfaces) were included. All studies were published in English. 24 studies were in vivo and 93 studies were in vitro. Risk of bias was found to be low in 23 and high in 94 studies. The pooled sensitivity for detecting any kind of occlusal carious lesions was 0.35 (95% CI : 0.31/40) and 0.41 (0.39/0.44) in clinical and in vitro studies respectively while the pooled specificity was 0.78 (0.73/0.83) and 0.70 (0.76/0.84). For the detection of any kind of proximal lesion the sensitivity in the clinical studies was 0.24 (CI 0.21/0/26) and 0.43 (0.41/0.45) and the specificity was 0.97 (0.95/0.98) and 0.89 (0.88/0.90).With regard to the dentine lesions the sensitivities were 0.36 (0.24-0.49) for proximal to 0.56 (0.53-0.59) for occlusal lesions and specificities ranged between 0.87 (0.85-0.89) and 0.95 (0.94-0.96). No reports were found for cavitated occlusal lesions. For proximal lesions sensitivities were above 0.60 and sensitivities above 0.90. Diagnostic Odds Ratios (DOR) were >1 (indicates a useful test) and were higher in proximal than in occlusal lesions. The DOR calculated for proximal lesions in vitro studies was 16.0 (11.5/22.4) and DOR 7.5 (3.4/16.5) for clinical studies). Heterogeneity calculated using I(2) test was moderate: > 50-67%.ConclusionsCaries detection determined by dental radiographs is highly accurate for proximal lesions and dentine caries lesions. For initial carious lesions the test needs to be used with other more sensitive methods in populations that present with high caries risk.