Background: The aim of this study was to compare the diagnostic performance obtained from direct digital radiographic images (in their original form and after applying a grey-scale inversion) and conventional film in the detection of artificial periimplant cancellous bone lesions.
Methods: Four titanium implants were placed into the cancellous bone of a dry mandible and increasingly larger bone defects were created in their approximal sites. Radiographs were taken using conventional film and a digital charge-coupled device sensor. Twelve observers estimated three series of images (conventional, digital original, digital inverse) on a 5-point confidence scale. Data were evaluated statistically by analysis of variance and the sensitivity, specificity and accuracy of the three imaging modalities were calculated.
Results: Total mean observer confidence scores increased as the size of the defect also increased. No statistically significant differences were found among the three images for the absence of defect and the defect that corresponds to the smallest bur size. Significant differences were found for larger bur sizes between the conventional image and the two digital images and for the largest bur size between the digital inverse and the other two images. Specificity was high and sensitivity relatively low.
Conclusions: Peri-implant bone lesions must exceed a certain size to be confidently detected but the lesion absence is detected equally well with all three imaging modalities.