Detection of non-cavitated occlusal caries using different diagnostic methods

Objectives: To compare the diagnostic performance of cone-beam computed tomography (CBCT), digital periapical radiography (DPR), visual examination (ICDAS-based), and DIAGNOdent Pen for detecting suspicious noncavitated occlusal caries, using histological validation as the reference standard.
Materials and Methods: In this in vitro study, 110 extracted permanent molars were cleaned, mounted in acrylic blocks, and assessed independently by four observers (two restorative dentists; two oral and maxillofacial radiologists). Visual scoring followed an ICDAS-derived ordinal scale; DPR and CBCT were scored using ordinal depth-based criteria analogous to radiographic interpretation; DIAGNOdent Pen readings were interpreted per manufacturer thresholds. Teeth were then sectioned and examined under 40× magnification to establish histological status. Diagnostic performance (accuracy, sensitivity, specificity) was calculated against histology.
Results: Histology classified 29 teeth as sound (26.4%), 41 as enamel caries (37.3%), and 40 as dentin caries (36.4%). For all modalities, performance was higher for dentin than enamel lesions overall. Against histology, visual examination accuracies ranged 0.71–0.91; DPR accuracies ranged 0.81–0.95; CBCT accuracies ranged 0.80–0.92 across observers, with generally high sensitivities but variable specificities. DIAGNOdent showed high specificity (0.91) but low sensitivity (0.28) and low accuracy (0.41).
Conclusions: No single method was sufficiently reliable for diagnosing initial noncavitated occlusal caries alone. Device-based methods may be helpful when used as adjuncts—particularly to support assessment of dentin involvement—but visual clinical examination remains indispensable. The routine use of CBCT solely for caries detection is not justified given radiation considerations.