Objectives: To compare the diagnostic quality of conventional F-speed film, digital images in TIFF and JPEG formats for detection of recurrent caries.

Method: MOD cavity preparations were made in eighteen premolars and eighteen molars. Half of the cavities were restored with amalgam and half with radiopaque composite. Artificial lesions were created just below the proximal box floor on either the facial, central or lingual side with #1/2, 2 and 4 round burs. These lesions were created only on one proximal surface. The other side served as the control. The teeth were mounted in plaster blocks that were positioned on an optical bench during the image acquisition. Radiographic images were made using F-speed film at 28 impulses and digital storage phosphor plates at 6 impulses. Digital images were saved as TIFF and JPEG formats. Four blinded and calibrated raters evaluated the images for presence or absence of recurrent caries on the mesial or distal surface using a five-point scale, where 1 indicated lesions definitely present and a value of 5 indicated lesion definitely absent. Raters reread one of the randomly selected image-type with an interval of at least four days between the viewing sessions.

Results: Significant differences (p<. 0001) were found in diagnostic accuracy of the three types of images. TIFF format showed the greatest percentage of accurate rating (57.64%). Conventional F-speed images showed 30.56% accurate rating and JPEG format had 38.86% accurate rating. No significant differences were found between the different lesion locations (p=0.6914). Raters identified 88% of the lesions correctly. Results of Kappa test (? = 0.75) indicated a moderately high degree of agreement between the raters.

Conclusions: TIFF format showed the greatest diagnostic accuracy among the three images types assessed for artificial recurrent caries detection.

Back to the Diagnostic Systems Program
Back to the ADEA/AADR/CADR Meeting & Exhibition (March 8-11, 2006)

Top Level Search