The aim of this study was to evaluate the diagnostic accuracy of visual, fibreoptic transillumination (FOTI), and bite-wing radiographic examination performed by 4 observers for the identification of cavitated carious lesions in contacting approximal surfaces, and to assess the inter-observer agreement with these methods and with direct visual examination conducted after tooth separation, the method used as validation for definitive determination of cavitation. A total of 338 unrestored approximal surfaces in 53 students were examined independently by 4 dentists using the diagnostic methods under study. The results from the diagnostic methods were compared with the results from the validation method for each observer. The sensitivities for identification of cavitated lesions using visual examination ranged from 0.12 to 0.50. For FOTI and radiography, the sensitivities ranged from 0.00 to 0.08 and from 0.56 to 0.69, respectively. The specificities exceeded 0.90 for all observers with all methods. Kappa values expressing inter-observer reproducibility were lowest for FOTI, followed by visual and radiographic examination. On the basis of these results, it was concluded that FOTI was the least reliable of the diagnostic methods tested. For the validation method, the inter-observer agreement was only 'substantial'. This implies that the method cannot be used as a validation for other diagnostic methods applied for the identification of cavitated carious lesions in contacting approximal surfaces. However, visual inspection after tooth separation may serve as a supplementary diagnostic tool to conventional visual and radiographic examination for clinical management of aproximal carious lesions.