Data tables were created based on relevant radiographic features, and only those studies that were felt to be free of bias with regard to the particular feature were included in the estimates.