Advanced Time-Stepping Interpretation of Fly-Scan Continuous Rotation Synchrotron Tomography of Dental Enamel Demineralization

Abstract

High-resolution spatial and temporal analysis and 3D visualization of time-dependent processes, such as human dental enamel acid demineralization, often present a challenging task. Overcoming this challenge often requires the development of special methods. Dental caries remains one of the most important oral diseases that involves the demineralization of hard dental tissues as a consequence of acid production by oral bacteria. Enamel has a hierarchically organized architecture that extends down to the nanostructural level and requires high resolution to study its evolution in detail. Enamel demineralization is a dynamic process that is best investigated with the help of in situ experiments. In previous studies, synchrotron tomography was applied to study the 3D enamel structure at certain time points (time-lapse tomography). Here, another distinct approach to time-evolving tomography studies is presented, whereby the sample image is reconstructed as it undergoes continuous rotation over a virtually unlimited angular range. The resulting (single) data set contains the data for multiple (potentially overlapping) intermediate tomograms that can be extracted and analyzed as desired using time-stepping selection of data subsets from the continuous fly-scan recording. One of the advantages of this approach is that it reduces the amount of time required to collect an equivalent number of single tomograms. Another advantage is that the nominal time step between successive reconstructions can be significantly reduced. We applied this approach to the study of acidic enamel demineralization and observed the progression of demineralization over time steps significantly smaller than the total acquisition time of a single tomogram, with a voxel size smaller than 0.5 μm. It is expected that the approach presented in this paper can be useful for high-resolution studies of other dynamic processes and for assessing small structural modifications in evolving hierarchical materials.

Figure 1

Schematic illustration of the continuous-tomography analysis of the enamel. (a) Overview of the analysis carried out with the iterations corresponding to a new saved scan file as a function of the index of projections (equivalent to the angle of rotation). Image of one horizontal slice of the sample before and after the scan iterations acquired with the 4× objective lens and a voxel size of 0.8125 μm (tomography scan referred to as “tomo”) in contrast with the 10× objective lens used during the iterations under acid immersion before neutralization. Each scan file covered a range from 0 to 1080° and is further described in (b). The 180° scans were extracted to reduce to the size of data sets for analysis. (b) The plot of the angle and projections as a function of tomogram (over the range of 180°) for one scan iteration that contains projections indexed from 0 to 15 000 (equivalent to six consecutive tomograms every 180°) and took around 2 h 8 min. The 3D reconstructed data sets were generated by extracting 2501 consecutive projections, which were then reconstructed to provide 3D details—this is shown with the 3D rendering of some reconstructed data sets with a voxel size of 0.325 μm. Dark- and flat-field images were acquired at the beginning of Scan 1. From the 3D reconstructed volume, slices were selected to illustrate the evolving structural changes seen in enamel from Scan 1 to Scan 4 after immersion in citric acid. (c) Schematic of the analysis of the tomograms from the Scan 1 with the associated rotation angles. This highlighted the overlap of tomography ranges used for the reconstruction of tomogram, which led to an improvement in temporal resolution.

Figure 2

Optical images of the sample before synchrotron analysis. Images of the tooth sample covered with blue varnish with magnified region in (b) showing the window left for the demineralization of the enamel.

Figure 3

Progress of demineralization of enamel over a time period, captured using synchrotron X-ray tomography. (a) 3D rendering of the enamel before demineralization and highlighting of the position of a slice. (b) Visualization of the progression of the demineralization with the slice described in (a) and sequence of slices (generated from 3D reconstruction) at different events during demineralization of enamel with details of the projections used to reconstruct the data set (see Figure 1 for the details of the file and projections) and the time. The time difference (in the format hh:mm:ss) corresponds to the duration between the starting time of the acquisition of the first projection of two tomograms with the description of the slices. (c) 2D images of regions of interest (1080 × 1080 × 2110 pixels) highlighted with blue boxes in the slices in (b) with the details of the rods and inter-rod substance. The data sets were reconstructed with a voxel size of 0.325 μm.

Figure 4

Evolution of the structure of the enamel through the cavity in a tomogram. (a) 3D rendering of the enamel with the highlight of slices through the thickness of the sample. Two regions of interest are illustrated on a slice located in the demineralized region and described in (b,c). (b) Sequence of slices with separation distances of 32.5 μm along the depth of the tomogram from a region of interest (1080 × 1080 × 2110 pixels) showing the demineralized region within the sample, blue region. (c) Magnified region described in (a,b) with the 3D rendering of a region of interest (in yellow) showing the enamel, rods, and the demineralized region (745 × 552 × 300 pixels). The tomography data sets were reconstructed with a voxel size of 0.325 μm.

Figure 5

SEM images of the sample after X-ray tomography. (a) SEi and BSi analysis at various magnifications of the enamel showing the demineralized region and overall damages on the sample and (b) high-magnification image.