Comparison of conventional and TACT (Tuned Aperture Computed Tomography) digital subtraction radiography in detection of pericrestal bone-gain

Abstract

Tuned Aperture Computed Tomography or TACT is a relatively new technique that transforms correlated two-dimensional images into a three-dimensional display. Associating TACT with digital subtraction radiography (TACT-DSR) may enhance its use in the detection of diseases. This study aims to assess observer performance in the task of detecting simulated bone-gain in periodontal defects, comparing conventional and TACT DSR. Buccal, lingual and proximal pericrestal bony defects in a human dry skull were imaged with and without increments of crushed-bone and wax mixtures using standardized vertical bitewing projections, acquired with an intraoral digital receptor in a tomographic X-ray machine. One hundred and twenty pairs of baseline and follow-up images were obtained for each imaging modality. Post-acquisition processing including image-registration, density-correction and subtraction operation were applied to both transmission images and TACT-slices. Eight calibrated observers evaluated the presence or absence of pericrestal bone-gain using a 5-point confidence scale. Details about the location of periodontal defects and the amount of bone-gain for each subtracted image were recorded and used in the statistical analysis. Observer performance and study design factors were analyzed by means of receiver operating characteristic (ROC) curve analysis and ANOVA. Mean values of areas under ROC curves for bone-gain detection performance were 0.82 for conventional DSR and 0.90 for TACT-DSR. TACT-DSR provided statistically superior detection performance of simulated periodontal bone-gain over conventional DSR for tooth-obscured defects (p = 0.001) and for the smallest bone mass (p = 0.024). The findings support TACT-DSR's potential to provide greater sensitivity and technique flexibility in detecting periodontal bone-gain than standard DSR.