This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further informationregarding Elseviers archiving and manuscript policies are

encouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Micron 44 (2013) 404409

Contents lists available at SciVerse ScienceDirect

Micron

journa l homepage: www.e lsev ier .com/ locate /micron

In situ demineralisation of human enamel studied by synchrotron-based X-raymicrotomography A descriptive pilot-study

Julia Lautensacka,, Alexander Rackb,c, Claudia Redenbachd, Simon Zablerc,e,Horst Fischer f, Hans-Georg Grberg

a Department of Prosthodontics and Dental Materials in Department of Prosthodontics, Implantology and Biomaterials, Medical Faculty, RWTH Aachen University, Pauwelsstrae 30,52074 Aachen, Germanyb European Synchrotron Radiation Facility, 6 rue Jules Horowitz, 38043 Grenoble, Francec Helmholtz Zentrum Berlin fr Materialien und Energie, Institute of Applied Materials, Hahn-Meitner-Platz 1, 14109 Berlin, Germanyd Department of Mathematics, University of Kaiserslautern, P.O. 3049, 67653 Kaiserslautern, Germanye Lehrstuhl fr Rntgenmikroskopie, Julius-Maximillians Universitt Wrzburg, Am Hubland, 97074 Wrzburg, Germanyf Department of Conservative and Preventive Dentistry, Medical Faculty, RWTH Aachen University, Pauwelsstrae 30, 52074 Aachen, Germanyg Department of Dental Materials and Biomaterials Research, RWTH Aachen University, Pauwelsstrae 30, 52074 Aachen, Germany

a r t i c l e i n f o

Article history:Received 19 May 2011Received in revised form 8 September 2012Accepted 9 September 2012

Keywords:CariesMineral concentrationIntraoral applianceIntra-oral cariogenicity test systemSynchrotron radiationQuantitative image analysis

a b s t r a c t

An in situ study was designed to investigate naturally developed demineralisation in human enamel in awidely non-destructivemanner in combinationwith X-raymicrotomography. Samples of human enamelwere carried in theoral cavity of participants for 24hdaily for either 21or29daysusing so-called intraoralmandibular appliances (ICTs). Demineralisation was thereby generated in a natural way without caus-ing caries in the subjects dentition. By employing synchrotron-based X-ray microtomography (XMT) incombination with volume image analysis, a quantication and three-dimensional visualisation of dif-ferent stages of mineral density loss was possible. Basic features of the demineralised samples weresimilar to those reported in earlier in vitro studies. However, the analysed samples showed signicantdifferences in the morphology of surface attack and the degree of mineral density loss depending onthe carrier, the exposure time and the position within the ICT. In particular, the varying local condi-tions within a carriers oral cavity seem to be different than in an in vitro study. Our results show thatthe combination of ICTs and quantitative image analysis applied to XMT data provides an analyticaltool which is highly suited for the fundamental investigation of naturally developed demineralisationprocesses.

2012 Elsevier Ltd. All rights reserved.

1. Introduction

Several studies attempting to understand the kinetics of earlydemineralisation which starts underneath the enamel surface havebeen performed in the past. Imaging techniques such as opticalmicroscopy (Darling, 1958) and electron microscopy (Haikel et al.,1983; Hicks and Silverstone, 1985; Jones and Boyde, 1987; Pearceand Nelson, 1989; Shellis and Hallsworth, 1987; Boyde et al., 1995,1999) were frequently used. Those techniques have the disadvan-tage of penetrating the material under study a few nanometres to1m only. Therefore, the resulting two-dimensional images pro-vided a limited insight into the lesion development and requiredinvasive tissue preparation for the investigation of structural

Corresponding author at: Department of Prosthodontics and Dental Materials,Medical Faculty, RWTH Aachen University, Pauwelsstrasse 30, D-52074 Aachen,Germany.

E-mail address: julia.lautensack@gmx.de (J. Lautensack).

changes below the surface of the specimens (Baumgartner et al.,2000).

An alternative approach is the application of X-ray microcomputed tomographic (XMT) techniques which provide three-dimensional images of the scanned specimen (Hsieh, 2009; Stock,2008; Buzug, 2008). Numerous articles reporting on enamel dem-ineralisation using high resolution computed tomography havebeen published which (e.g., Bing et al., 2011; Hamba et al., 2011;Schwass et al., 2009; Tanaka et al., 2010). Frequently synchrotronradiation is employed for XMT in order to achieve higher spatialresolution and contrast (Bonse and Busch, 1996). At a synchrotronlight source the available photon ux density (Photons/mm2/s) isorders of magnitude higher and the beam propagation is quasi par-allel. The latter allows for extending the distance between sourceand sample up to several 100m. Consequently, unlike in conven-tional radiography which is based on X-ray tubes, the inuenceof the nite X-ray source on the image resolution is surpressed(cf., e.g., Rack et al., 2008b). Furthermore, the high ux allows theuse of monochromators. Monochromatic radiation increases the

0968-4328/$ see front matter 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.micron.2012.09.006

Author's personal copy

J. Lautensack et al. / Micron 44 (2013) 404409 405

contrast while reducing beam-hardening artefacts (Brooks and DiChiro, 1976).

Today, X-ray microtomography (XMT) is a mature techniquewith high potential for dental research (Delbem et al., 2009).Working in absorption contrast, the images acquired with (quasi-)monochromatic radiation allow for a quantitative measure ofchanges in the degree of demineralisation inside, e.g., enameltissue. Several studies attempting to understand the three dimen-sional kinetics of demineralisation have been performed usingXMT (Kernen et al., 2008; Kinney et al., 1994, 1995; Anderson andElliott, 2000; Stock, 2008; Farah et al., 2009; Huang et al., 2007;Delbem et al., 2009; Kielbassa et al., 2000; Dowker et al., 2003,2004; Davis and Wong, 1996; Elliott et al., 1981), even thoughextrapolating the three-dimensional de- and remineralisation pro-cesses from such ameasurement is a non-trivial task (Dowker et al.,2003).

Some studies found in the literature are based on in vitro inves-tigations, i.e. the demineralisation is articially produced outsidethe oral cavity. Obviously, the clinical relevance of in vivo stud-ies is signicantly higher, in particular due to the multi-factorialnature of dental caries (ten Cate, 1996; Zero, 1995). Unfortunately,XMT is inapplicable for human dental in vivo studies for mul-tiple reasons. One reason is the high mutagen potential of theemployed hard X-ray radiation. Hence, demineralised teeth haveto be extracted prior to imaging (Dowker et al., 2004; Huang et al.,2007, 2010). This makes time-resolved investigations of deminer-alisation processes impossible. Furthermore, information on theduration of and the conditions during demineralisation is lim-ited.

As an alternative, we propose an in situ study using intraoralappliances (intra-oral cariogenity test systems, ICT)which allow foranexposureof enamel samples to thehumanoral cavityover longerperiods of time. Consequently, they represent promising tools tostudy naturally developed demineralisation in human oral cavitieswithout causing any decay of the natural dentition (Koulouridesand Volker, 1964; Koulourides and Chien, 1992). Thus, in situ stud-ies combining ICTswithXMT imagingof thedemineralised samplesmay bridge the gap between the highly controlled laboratory con-dition in vitro and the uncontrolled clinical situation in vivo. It is theaim of our pilot study to test the applicability of these techniquesfor the investigation of natural demineralisation processes. To thebest of our knowledge such studies have not been performed sofar. Of particular interest is whether the enamel demineralisationdiffers in its morphology from articially generated lesions.

Furthermore, earlier studies focused mainly on the evaluationthe three-dimensional (3D) imagedata bymeansof volume render-ings and density proles along line sections. In our work methodsfrom quantitative volume image analysis are introduced. Theirapplication allows one to fully exploit the information providedby the 3D image data.

2. Materials and methods

2.1. Preparation of specimens and intraoral appliance

Eight caries free third molars were extracted and stored for 24hin Ringers solution (DAB 7; Delta-Pharma, Pfullingen, Germany).After extraction the remaining tissue was removed with a scalpelunder running water. With a diamond-coated wire saw (wire:300m; saw: Type 4240, Well, Mannheim, Germany) the crownswere separated from the roots. During the sawing process theteeth were wetted constantly with phosphate enriched Na-Cl solu-tion. 33 cubes of roughly 1mm3 volume consisting of enamel ofthe occlusal surfaces and parts of contiguous dentine were pre-pared.

The specimens were covered with a varnish coat on ve faces,leaving only the natural outer (occlusal) enamel surface free (Nailvarnish: Nivea, Beiersdorf, Hamburg, Germany). The cubes weresterilised for 25min at 1.8 bar and 122 C (Technoclav 50, Techno-mara, Fernwald, Germany).

Afterwards 32 samples were randomly integrated into fourremovable ICTs, as shown in Fig. 1. The assembled samples werecovered with a gold-micro mesh that allowed for accumulation ofplaque on the specimen surface.

2.2. Participants and experimental setup

Four participants volunteered for the pilot-study (3 female, 1male, age: 2327 years). The participants signed appropriate infor-mation consent forms. They were in good general health, had noinfectious diseases and did neither take antibiotics normedicationsthat affect salivary ow for at least twelve months before and dur-ing the study. All participantswere non-smokers andhadmoderatepast caries experience with restored decays. If necessary, radio-graphs were taken to exclude caries and all carious lesions wererestoredbyadentist of theDepartmentofConservativeandPreven-tiveDentistry, RWTHAachenUniversity,Germany. Theparticipantshad a minimum of 28 natural teeth and generally good periodon-tal health. They wore the ICT day and night and removed it onlyfor individual oral hygiene with uoride toothpaste and to cleanthe ICT twice a day with uoride-free toothpaste. To allow plaqueto grow under the micro mesh the participants were instructed toclean the area of the specimens only by using running water. Wechose different time periods of carrying the samples, presumingdemineralisation to emerge after a period of at least three and atmost four weeks. Since the time period for the XMT measurementwas constricted to several hours during one day, all samples had tobe removed from the ICTs on the same day. Therefore, two researchsubjects (participants A and B) startedwith the in situ study 21 daysand two subjects (participants C and D) 29 days before the day ofmeasuring. After extraction and during XMT measurement eachsamplewas stored in a labelled closed container lledwith Ringerssolution. A total of seven samples from participant A, eight fromparticipant B, seven from participant C and one from participant Dwere scanned by XMT.

2.3. Synchrotron-based microtomography using hard X-rays

The XMT measurements were carried out at the BAMline, therst hard X-ray beamline of the BESSYII light source (HelmholtzZentrum fr Materialien und Energie, Berlin, Germany, cf. Racket al., 2008b). Quasi-monochromatic radiation of 25keV X-rayphoton energy was attained by setting a double-multilayermonochromator (approximately 1% energy bandwidth). Attenua-tors were inserted to block photons passing the monochromatorby means of total reection. To convert the synchrotron light aftertransmitting the sample into visible light, a 7-m-thick YAG:Ce(Ce-doped Y3Al5O12) single-crystal scintillator was used. The scin-tillator was coupled via diffraction-limited visible light optics toa CCD camera (VersArray: 2048B, 20482048 pixel, PrincetonInstruments). The indirect detector samples the visible light imagewith 1.6m effective pixel size. The spatial resolution achievedis approximately 78m. The images were taken in absorptionmode, i.e. the detector was placed close downstream of the sam-ples, which allows for neglecting coherence effects on our results,i.e. X-ray inline phase contrast (Zabler et al., 2007). Each spec-imen was scanned with 900 projections over a 180 rotation.Three-dimensional data sets were reconstructed using the lteredbackprojection algorithm (Kak and Slaney, 1988) implemented in

Author's personal copy

406 J. Lautensack et al. / Micron 44 (2013) 404409

Fig. 1. Design of the intraoral mandibular appliances (ICT). The removable appliance is designed for the mandibular jaw. On the lingual side of the 3rd and 4th quadrand foursamples of human enamel are polymerised below the micromesh (photograph).

the software package PyHST of the SciSoft group of the EuropeanSynchrotron ESRF (Banhart, 2008; Mirone et al., 2012).

2.4. Image analysis

Tomographic reconstruction yields volume image data repre-sented as a three-dimensional matrix of grey values. In our case,each point in this matrix (a three-dimensional pixel which is calledvoxel) corresponds to (1.6m)3 of material. Our image analysisis based on the assumption that the value associated with eachvoxel is approximately proportional to the local attenuation coef-cient (LAC) of the material which is directly connected to its localphysical density. Dowker et al. (2003) introduced a method for thereconstruction of the mineral concentration from the LAC mod-elling enamel as a two-phase system. Since the parameters of thismodel are unknown an uncertainty is introduced into the analysisresults.

In this article, an alternative approach is proposed. Making surethat sound (non-demineralised) enamel is always included in theimagewe can normalise and thereforemeasure the relative densityloss due to demineralisation in a quantitative manner (Rack et al.,2008a).

In a rst step, relevant regions were cropped from the volumeimages. After denoising the resulting images by application of a333 mean-lter, they were binarised using a simple threshold.The exposed enamel surface was identied by a morphological clo-sure (Ohser and Schladitz, 2009). The distance of each enamel voxelto the sample surface was computed using the Euclidean distancetransform.This transformdetermines for each foreground (enamel)voxel the distance to the closest background (liquid) voxel (Ohserand Schladitz, 2009). Finally, for all voxels with the same distancevalue the grey values in the original image were counted. Hence,the distribution of grey values, i.e. the degree of demineralisation,

for points with a given distance to the surface was investigated. Forprocessing and analysis of the volume images we used the MAVIsoftware package (MAVI, 2012). Sectional images of an originalimage and the reconstructed surface are shown in Fig. 2.

3. Results

Representative cross sections and renderings of the recon-structed grey value images of three samples showing differentdegrees of demineralisation are shown in Fig. 3. Samples wornby participants A and B, which were extracted after 21 days,showed an intact surface and no or very mild stages of sub-surfacedemineralisation (Fig. 3a). The samples from participants C and Dwere extracted after 29 days. All seven samples from participantC (e.g., Fig. 3b) showed different gradual changes in the enamelstructure. Seven samples worn by participant D were disintegratedto a degree which made structural investigation impossible. Onlyone sample remained stable enough to be scanned (sample D4,Fig. 3c). In this sample, demineralisation occurred in form of a deepdemineralisation zone which partially fractured thus forming largesubcutaneous cavities. The original sample surface was completelydestroyed.

The images were analysed using the image processing tech-niques described above. Since all samples worn by participants Aand B showed a similar structure, only one sample for each of theseparticipants was included in the analysis. For participant C, onesample was excluded from the analysis due to inappropriate sam-ple geometry. The remaining six samples were analysed. For theremaining sample of participant D, a morphological reconstructionof the sample surface was not possible.

Fig. 4 shows the mean relative grey value (the mean grey valuein a given depth divided by grey value for sound enamel) as afunction of the distance to the reconstructed sample surface. Since

Fig. 2. Sectional images of the original grey value image (a) and the original image masked with reconstructed surface (b) for sample C1.

Author's personal copy

J. Lautensack et al. / Micron 44 (2013) 404409 407

Fig. 3. Representative cross section images and volume renderings showing mild sub-surface demineralisation (a, d: sample B4), intermediate (b, e: sample C3) and severe(c, f: sample D4) stages of demineralisation.

mean grey values for consecutive distance values varied to a certaindegree a smoothing spline was t to the observations to obtain thecurves shown in Fig. 4. Bright areas close to the sample surface (seee.g. Fig. 2 or Fig. 3a) were partly levelled out by this procedure.

In the plots, a value of one corresponds to sound enamel. Thedifference from one can be used as a measure for the degree ofdemineralisation of the sample. Ring artefacts, which are a com-mon type of artefacts in tomographic image data (see Fig. 3e orFig. 5b below), only appeared at relatively large distances to thedemineralised zone such that they did not affect the analysis.

Further results of the analysis are summarised in Table 1. Inthis table the depth of demineralisation is dened as the minimal

0.300.250.200.150.100.050.00

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Re

lative

me

an

gra

y v

alu

e [a

rb. u

nits]

Distance to surface [mm]

A4

B4

C1

C3

C4

C5

C7

C8

Fig. 4. Relative mean grey value (mean grey value divided by grey value for soundenamel) as a function of the distance to the sample surface. High grey values cor-respond to high enamel density; lower grey values indicate demineralisation of thematerial. Subjects A and B exposure time 21 days, C 29 days.

distance to the sample surface such that the relative mean greyvalue shown in Fig. 4 reaches the value of 1.

The results show signicant differences in tissue loss betweendifferent exposure times and carriers. However, also in samplesworn by the same person for the same period differences in degreeand morphology of the demineralisation were observed. To fur-ther illustrate these differences, Fig. 5 shows volume renderings ofsamples C1 and C4. Sample C1 shows a well-organised progress ofdemineralisation which can be divided into two zones at differentdepth separatedbyabright intermittent layer of almost fullmineraldensity which is located approximately 100m below the enamelsurface. In addition to the obvious loss of mineral density subcu-taneous porosities were observed in the form of elongated poreswhich extend perpendicular to and approximately 100200munderneath the enamel surface.

Contrary, the acid-attack on sample C4 is more localised. Areaswithdeeplypenetratingdemineralisationalternatewithareaswithonly mild stages of demineralisation. Note that this local attackresults inacomparably lowdegreeofdemineralisationasmeasuredby the volume image analysis and shown in Fig. 4 for this sample.Highly demineralised areas show a similar layer structure as thehomogeneous sample C1. However, the amount of subcutaneouspores in this sample is much lower than in sample C1.

Table 1Degree and depth of demineralisation for the analysed samples.

Sample Analysed volume[voxel]

Maximal reductionof mean grey value[%]

Depth ofdemineralisation[m]

A4 1d 7387831053 12.4 201B4 2d 7479551040 9.5 158C1 3a 453330666 42.1 267C3 3c 571579772 43.4 243C4 3d 721552490 32.4 252C5 3e 580227453 25.6 135C7 3g 693533542 40.0 269C8 3h 542736748 51.5 210

Author's personal copy

408 J. Lautensack et al. / Micron 44 (2013) 404409

Fig. 5. Visualisation of samples, C1 (a) (0.72mm0.53mm1.07mm) and C4 (b) (1.15mm0.88mm0.78mm). The differentmorphology of surface penetration is clearlyvisible. Imaging artefacts (half-circles) due to detector inhomogeneities are present in (b).

4. Discussion and conclusion

The aimof this pilot-study is to introduce a protocol for studyingdemineralisation in human enamel in situ. For the latter, intraoralmandibular appliances have been successfully applied.

In our study, zones of the incipient enamel lesions includingdened surface layerswereobserved. The zones exhibited the char-acteristics of natural caries and might be similar to the classicalzones described by Darling (1956, 1958) and Gustafson (1957).Therefore, the morphology of in situ demineralised enamel wouldbe equivalent to in vitro developed lesions (Zero, 1995).

The observed demineralisation differed signicantly in grade,depth and morphology due to different carrying periods as well asnon-equal conditions in the participants oral habitat. In sampleswith progressed demineralisation, two types of corrosive attack planar surface attack or pitting were observed. Both types showedsigns of an intermittent layer approximately 100m below thedeepest demineralisation front. This layer has a high concentrationin mineral almost matching the density of the intact enamel. Yet,the X-ray images contain no information about whether this layeris a solid barrier or the accumulation of corrosion products (mainlydissolved Ca2+ ions) in the liquid. Tubular porositywas found in thedemineralisation zones of many samples in this study whereby theradial orientation of the pores most likely follows the patterning ofthe enamel crystalline rods in the corresponding zone.

From our observation on the samples exposed for 29 dayswe postulate a four-layer structure of demineralisation: close tothe enamel surface, demineralisation creates a layer of stronglyreduced density marking the beginning decay. At 100m deptha higher density layer is observed followed by another porous lowdensity layer. The nal fourth layer is formed by sound enamel.Due to the averaging properties of the employed volume analysisalgorithm, the intermittent zone between the two demineralisa-tion layers which is visible in the images did not show up as clearlyin the density proles in Fig. 4.

It is well established that the lesion development depends ondifferent individual factors, e.g., salivary rate, mineral concentra-tion of the salivary, buffering capacity and nutrition behaviour(Silverstone et al., 1988). Under certain in vivo conditions a dem-ineralisation can even stagnate, or furthermore regress in termsof remineralisation (Robinson et al., 2000; Robinson, 2009). Sig-nicant differences in demineralisation, in particular between

participants C and D, were also found in our study. However, dueto the small number of research subjects, a detailed investigationof the inuence of individual factors was not possible. For this pur-pose, it is desirable to perform a study including a much largernumber of research subjects with differing individual conditions.Besides the progress of demineralisation, the remineralising roleof uoride in caries prevention in human subjects can be studiedin a precise way. A detailed analysis of in situ remineralisation ofenamel will be the topic of our future research.

Due to insufcient information on the development of naturaldemineralisation processes, it was found challenging to select theoptimal point in time for sample removal. Two different carryingperiods were selected assuming that demineralisation occurs afterat least two and atmost fourweeks. Additionally, only a single timewindow was available for the XMT measurements such that thein situ part of our study had to be completed up to a hard deadline.Our results show that a period of three to four weeks is a criticaltime window for start and development of demineralisation pro-cesses and seems to be most suited for sample removal. For futurestudies, it is advisable to remove samples at multiple time pointsin order to follow the demineralisation process in different individ-uals. Even a reintegration of scanned samples into the ICT can beconsidered.

5. Summary

We have presented a pilot study aiming at the development of aprotocolwhichallows for the investigationofnatural demineralisa-tion processes in human enamel. For this purpose, enamel sampleswere inserted in theoralhabitat of four research subjectsusing ICTs.After a certain carrying period, the samples were scanned by XMTand the degree of demineralisation was quantied using methodsfrom 3D image analysis.

The ICT was found to be an instrument well-adapted for theinvestigation of in situ development of naturally demineralisedenamel. The chosen dimension of the samples was well suited fortting into the ICTs and subsequently into the XMT detectors eldof view. ICTs provided environmental conditions for natural plaqueformation on the sample surface without causing decay on the par-ticipants dentition. Access of dietary substrate and salivary wasgiven, which resulted in plaque accumulation that led to a diag-nosable demineralisation.

Author's personal copy

J. Lautensack et al. / Micron 44 (2013) 404409 409

Synchrotron XMT provided highly resolved images clearlyshowing the density variations within the enamel structure.Methods from quantitative image analysis do not only allowfor a visualisation but also for a quantication of the degreeof demineralisation in different samples. In contrast to classi-cal two-dimensional imaging techniques, XMT captures the fullythree-dimensional morphology of the demineralised areas. Fur-thermore, the degree of demineralisation in a given distance to thetrue (typically non-planar) sample surface can be investigated in athree-dimensional manner.

Acknowledgements

The authors would like to thank Dr. A. Haibel and Dr. I. Manke(Hahn Meitner Institute, Berlin, Germany) as well as Dr. H. Riese-meier (Federal Institute of Materials Research and Testing (BAM),Berlin, Germany) for their support and suggestions during theexperiments aswell as for discussion. Beamtimeat theBAMlinewasprovided by the BESSY II light source (Helmholtz-Zentrum Berlin,Germany). Special thanksgo toMaximilian Lautensack, AglaiaRack,Elina Redenbach.

References

Anderson, P., Elliott, J.C., 2000. Rates of mineral loss in human enamel duringinvitrodemineralisationperpendicular andparallel to thenatural surface.CariesResearch 34, 3340.

Banhart, J., 2008. Advanced Tomographic Methods in Materials Research and Engi-neering. Oxford University Press.

Baumgartner, A., Dichtl, S., Hitzenberger, C.K., Sattmann, H., Robl, B., Moritz, A.,Fercher, A.F., Sperr, W., 2000. Polarization-sensitive optical coherence tomogra-phy of dental structures. Caries Research 34, 5969.

Bing, H., Shengbin, H., Chaoliang, Z., Junjun, J., Yuqing, H., Liying, X., Xuedong, Z.,2011. Mineral densities and elemental content in different layers of healthyhuman enamel with varying teeth age. Archives of Oral Biology, March 14, 2011[Epub ahead of print].

Bonse, U., Busch, F., 1996. X-ray computed microtomography (CT) using syn-chrotron radiation (SR). Progress in Biophysics and Molecular Biology 65,133169.

Boyde, A., Davy, K.W.M., Jones, S.J., 1995. Standards for mineral quanticationof human bone by analysis of backscattered electron images. Scanning 17,67.

Boyde, A., Travers, R., Glorieus, F.H., Jones, S.J., 1999. The mineralization densityof iliac crest bone from children with osteogenesis imperfect. Calcied TissueInternational 64, 185190.

Brooks, R.A., Di Chiro, G., 1976. Beamhardening inX-ray reconstructive tomography.Physics in Medicine and Biology 21, 390398.

Buzug, T.M., 2008. Computed Tomography. From Photon Statistics to Modern Cone-Beam CT. Springer Verlag, Heidelberg, Germany, ISBN: 978-3-540-39407-5.

Darling, A.I., 1956. Studies of the early lesion of enamel caries with trans-mitted light, polarised light and radiography. British Dental Journal 101,289299.

Darling, A.I., 1958. Studies of the early lesion of enamel caries. Its nature, mode ofspread and points of entry. British Dental Journal 105, 119135.

Davis, G.R., Wong, F.S., 1996. X-ray microtomography of bones and teeth. Physio-logical Measurement 17, 121146.

Delbem,A.C.B., Sassaki, K.T., Vieira, A.E.M., Rodrigues, E., Bergamaschi,M., Stock, S.R.,Cannon, M.L., Xiao, X., De Carlo, F., Delbem, A.C.B., 2009. Comparison of meth-ods for evaluating mineral loss: hardness versus synchrotron microcomputedtomography. Caries Research 43, 359365.

Dowker, S.E.P., Elliott, J.C., Davis, G.R., Wilson, R.M., Cloetens, P., 2004. Syn-chrotron X-ray microtomographic investigation of mineral concentrationsat micrometre scale in sound and carious enamel. Caries Research 38,514522.

Dowker, S.E.P., Elliott, J.C., Davis, G.R., Wassif, H.S., 2003. Longitudinal study of thethree-dimensional development of subsurface enamel lesions during in vitrodemineralization. Caries Research 37, 237245.

Elliott, J.C., Dowker, S.E.P., Knight, R.D., 1981. Scanning X-ray microradiographyof a section of a carious lesion in dental enamel. Journal of Microscopy 123,8992.

Farah, R.A., Swain, M.V., Drummond, B.K., Cook, R., Atieh, M., 2009. Mineral densityof hypomineralised enamel. Journal of Dentistry 38, 5058.

Gustafson, G., 1957. The histopathology of caries of human dental enamel.With spe-cial reference to the division of the carious lesion into zones. Acta OdontologicaScandinavica 15, 1355.

Hamba, H., Nikaido, T., Inoue, G., Sadr, A., Tagami, J., 2011. Effects of CPP-ACP withsodium uoride on inhibition of bovine enamel demineralization: a quantita-tive assessment using micro-computed tomography. Journal of Dentistry [Epubahead of print].

Haikel, Y., Frank, R.M., Voegel, J.C., 1983. Scanning electronmicroscopy of the humanenamel surface layer of incipient carious lesions. Caries Research 17, 113.

Hicks, M.J., Silverstone, L.M., 1985. Internal morphology of surface zones from acid-etched caries-like lesions: a scanning electron microscopic study. Journal ofDental Research 64, 12961301.

Hsieh, J., 2009. Computed Tomography: Principles, Design, Artifacts and RecentAdvances, 2nd ed. SPIE, Bellingham.

Huang, T.T., Jones, A.S., He, L.H., Darendeliler, M.A., Swain, M.V., 2007. Characteri-sation of enamel white spot lesions using X-ray micro-tomography. Journal ofDentistry 35, 737743.

Huang, T.T., He, L.H., Darendeliler, M.A., Swain, M.V., 2010. Correlation of mineraldensity and elastic modulus of natural enamel white spot lesions using X-raymicrotomography and nanoindentation. Acta Biomaterialia 6, 45534559.

Jones, S.J., Boyde, A., 1987. Scanning microscopic observations on dental caries.Scanning Microscopy 1, 19912002.

Kak, A.C., Slaney, M., 1988. Principles of Computerized Tomographic Imaging. IEEEPress, New York.

Kernen, F., Waltimo, T., Deyhle, H., Beckmann, F., Stark, W., Mller, B., 2008. Syn-chrotron radiation-based micro computed tomography in the assessment ofdentin de- and remineralization. Proceedings of SPIE 7078, 19.

Kielbassa, A.M., Schendera, A., Schulte-Mnting, J., 2000. Microradiographic andmicroscopic studies on in situ induces initial caries in irradiated and nonirradi-ated dental enamel. Caries Research 34, 4147.

Kinney, J.H., Marshall Jr., G.W., Marshall, S.J., 1994. Three dimensional mapping ofmineral densities in carious dentin: Theory and method. Scanning Microscopy8, 197205.

Kinney, J.H., Balooch, M., Haupt Jr., D.L., Marshall, S.J., Marshall Jr., G.W., 1995.Mineral distribution and dimensional changes in human dentine during dem-ineralization. Journal of Dental Research 74, 11791184.

Koulourides, T., Volker, J.F., 1964. Changes of enamel microhardness in the humanmouth. Alabama Journal of Medical Sciences 1, 435-347.

Koulourides, T., Chien, M.C., 1992. The ICT in situ experimental model in dentalresearch. Journal of Dental Research 71, 822827.

Mirone, A., Wilcke, R., Hammersley, A., Ferrero, C., 2012. PyHST High Speed Tomo-graphic Reconstruction. http://www.esrf.eu/UsersAndScience/Experiments/TBS/SciSoft/

MAVI, 2012. Modular Algorithms for Volume Images, Fraunhofer ITWM,www.mavi-3d.de

Ohser, J., Schladitz, K., 2009. 3D Images of Material Structures. Wiley, Weinheim.Pearce, E.I., Nelson, D.G., 1989. Microstructural features of carious human enamel

imaged with backscattered electrons. Journal of Dental Research 68, 113118.Rack, A., Riesemeier, H., Zabler, S., Weitkamp, T., Mller, B.R., Weidemann, G., Mod-

regger, P., Banhart, J., Helfen, L., Danilewsky, A.N., Grber, H.G., Heldele, R.,Mayzel, B., Goebbels, J., Baumbach, T., 2008a. The high resolution synchrotron-based imaging stations at the BAMline (BESSY) and TopoTomo (ANKA).Proceedings of SPIE 7078, 70780X-170780X-9.

Rack, A., Zabler, S., Mller, B.R., Riesemeier, H., Weidemann, G., Lange, A., Goebbels,J., Hentschel, M., Grner, W., 2008b. High resolution synchrotron-based radio-graphy and tomography using hard X-rays at the BAMline (BESSY II). NuclearInstrumental Methods and Physical Research A 586, 327344.

Robinson, C., Shore, R.C., Brookes, S.J., Strafford, S., Wood, S.R., Kirkham, J., 2000. Thechemistry of enamel caries. Critical Reviews in Oral Biology and Medicine 11,481495.

Robinson, C., 2009. Fluoride and the caries lesion: interactions and mechanism ofaction. European Archives of Paediatric Dentistry 10, 136140.

Schwass, D.R., Swain, M.V., Purton, D.G., Leichter, J.W., 2009. A system of calibratingmicrotomography for use in caries research. Caries Research 43, 314321.

Shellis, R.P., Hallsworth, A.S., 1987. The use of electron microscopy in studyingenamel caries. Scanning Microscopy 1, 11091123.

Silverstone, L.M., Hicks, M.J., Featherstone, M.J., 1988. Dynamic factors affectinglesion initiation and progression in human dental enamel. The dynamics andnature of enamel caries. Quintessenz International 19, 683710.

Stock, S.R., 2008.MicroComputed Tomography:Methodology and Applications. CRCPress, London, New York, Boca Raton.

Tanaka, R., Shibata, Y., Manabe, A., Miyazaki, T., 2010. Micro-structural integrityof dental enamel subjected to two tooth whitening regimes. Archives of OralBiology 55, 300308.

ten Cate, J.M., Imfeld, T., 1996. Dental erosion, summary. European Journal of OralSciences 104, 241244.

Zabler, S., Cloetens, P., Zaslansky, P., 2007. Fresnel-propagated submicrometer X-rayimaging of water-immersed tooth dentin. Optics Letters 32, 29872989.

Zero, D.T., 1995. In situ caries models. Advances in Dental Research 9, 214230.