Identification of root canals in molars by tuned-aperture computed tomography

Aim To compare the tuned-aperture computed tomography system of imaging to conventional D-speed film for their ability to identify root canals in extr acted human molars. Methodology Thirteen maxillary and six mandibular human molars were mounted in acrylic blocks to simulate clinical conditions by surrounding the teeth with a radiodense structure. The teeth were then imaged with conventional D-speed film using a standard paralleling technique, and with a modified or thopantomograph OP100 machine using a Schick no. 2 size CCD sensor as the i mage receptor. The source images were registered and TACT slices were gener ated using TACT Workbench(TM) Software. Three observers were asked to ident ify the number of canals in the conventional film group and the TACT image group using specific criteria. Ground truth was established by cross-sectio ning the teeth at the coronal, middle, and apical thirds of the roots and d irectly visualizing the root canal morphology. Results TACT imaging detected 36% of 4th canals in maxillary molars and 80% of third canals in mandibular molars. Conventional film detected 0% of fou rth canals in maxillary molars and 0% of third canals in mandibular molars. The differences in canal detection between the two techniques were statist ically significant (Wilcoxon matched pair sign rank test, P = 0.001). Conclusions In this study, the TACT system of digital imaging was superior to conventional film in the detection of root canals in human molars and ma y be useful for the detection of root canals that will probably be missed u pon conventional X-ray examination.