The influence of radiographic exposure factors on the diagnosis of occlusal caries.

Abstract

To assess the impact of variations in radiographic density on occlusal caries diagnosis and observers' diagnostic strategy. Three series of radiographs of 60 extracted molars were examined by nine dental students and caries diagnosed using a five point confidence-rating scale. Film density in each series was varied by varying exposure time: 1 s (high density), 0.4 s (medium density) and 0.067 s (low density). The 'true' diagnoses were obtained by stereomicroscopy. Diagnostic quality was measured with ROC analysis as Az. The difference in Az, values between the low and medium density series was statistically significant. The sensitivity and the specificity values for students' diagnoses in outer third of dentine were 85% and 50% for the high density series, 76% and 56% for the medium density series and 42% and 77% for the light series. Occlusal caries was diagnosed best from the darkest radiographs. Specificity was higher with light radiographs but sensitivity increased with density. Therefore underdiagnosis is more frequent with light radiographs, while overdiagnosis occurs more often with dark. The awareness of these effects could be an important guideline in adjusting the density of digital images.