Methods: Ten extracted carious molars were scanned using a high definition XMT scanner. Caries, as judged by conventional tactile feedback, was removed with steel burs in a slow hand-piece. The teeth were then precisely repositioned and rescanned. Assuming the mineral phase of dentine to be hydroxyapatite (density = 3.15 g cm-3) and the organic phase to be collagen (density = 0.54 g cm-3), the linear attenuation coefficients were converted to mineral concentration. From the distribution curve, dentine that had a mineral concentration below 1.2 g cm-3 was considered to be carious. Over/under preparation was assessed by comparing images from pre- and post-operative scans.
Results: The mean mineral concentration for non-carious dentine was 1.42 (sd=0.042) g cm-3. For the removed carious dentine, the mean mineral concentration was ~0.5 g cm-3. The proportion of sound dentine contributed 8.5 – 44.3% of the total volume of removed dentine. There were also regions of residual carious dentine in all teeth. Line profiles through the carious lesions showed a distinctive body and leading edge. The carious lesion was shown, for the first time, to be bowl shaped as visualised in the 3-D rendered image. In some lesions, the pulp receded in three dimensions, matching the shape of the carious advancing front.
Conclusion: (1) The repeatability feature of the XMT scan allows measurement of changes of the specimen before and after experimental procedures; (2) Conventional tactile methods of caries detection could lead to excessive removal of sound dentine; (3) Alternative clinical caries discrimination techniques are needed to improve consistency of cavity preparation.
Development of XMT is supported by EPSRC GR/R28911/01