2.1.

Artificial Smooth Surface Caries

40 sound human posterior teeth were mounted on $12\text{-}{\mathrm{mm}}^3$ acrylic blocks after root resection. Teeth were sterilized using gamma irradiation and stored in a moist environment to preserve tissue hydration with 0.1% thymol added to prevent bacterial growth. The outer enamel surface was polished to remove any previous remineralized enamel and expose normal sound enamel. An acid resistant varnish was applied on the teeth on all areas outside the $3\times 2\text{-}\mathrm{mm}$ smooth surface test region, which was scanned using the PS-OCT system. Artificial, or in vitro, lesions on the teeth were formed using a well-characterized pH cycling model, replicating the cycle of demineralization and remineralization that takes place naturally in the oral environment.^{25, 26} Teeth were exposed to either a 3 or $10\text{day}$ $(n=20)$ daily regimen of $6\text{-}\mathrm{h}$ demineralization and $17\text{-}\mathrm{h}$ remineralization.²⁷ Teeth exposed to the pH cycling for $10\text{days}$ were intended to represent severe early caries, and the teeth exposed for $3\text{days}$ represent a less advanced lesion in both depth and overall mineral loss. During demineralization, each tooth was exposed for $6\mathrm{h}$ a day to a $40\text{-}\mathrm{mL}$ aliquot buffer solution containing $2.0\text{-}\mathrm{mmol}∕\mathrm{L}$ calcium, $2.0\text{-}\mathrm{mmol}∕\mathrm{L}$ phosphate, and $0.075\text{-}\mathrm{mol}∕\mathrm{L}$ acetate maintained at pH 4.3 and a temperature of $37°\mathrm{C}$ . After demineralization, each tooth was immersed for $17\mathrm{h}$ in a $20\text{-}\mathrm{mL}$ mineralizing solution of $1.5\text{-}\mathrm{mmol}∕\mathrm{L}$ calcium, $0.9\text{-}\mathrm{mmol}∕\mathrm{L}$ phosphate, $150\text{-}\mathrm{mmol}∕\mathrm{L}$ KCl, and $20\text{-}\mathrm{mmol}∕\mathrm{L}$ cacodylate buffer maintained at pH 7.0 and $37°\mathrm{C}$ .

Sample teeth from both the $3\text{day}$ $(n=10)$ and $10\text{day}$ $(n=10)$ regimens were subsequently exposed to a fluoride containing remineralization solution. This solution was identical in composition to the mineralizing solution used in the pH cycling during the formation of the artificial lesions, except that $2\text{-}\mathrm{ppm}$ ${\mathrm{F}}^{-}$ in the form of NaF was added to the solution to enhance the remineralization effect. Teeth were then continuously exposed for $20\text{days}$ to the fluoride remineralization solution. For histological evaluation, $200\text{-}\mu \mathrm{m}$ -thick transverse sections from the demineralized and remineralized samples were obtained with a microtome and a digital micrometer, after embedding the samples in a resin polymer to preserve the lesion surface layer. Other tooth sections were also embedded in a resin polymer, polished to a $3\text{-}\mu \mathrm{m}$ finish, so that the entire transverse lesion and underlying enamel was exposed for high resolution IR spectromicroscopic analysis.

2.2.

Polarization-Sensitive Optical Coherence Tomography

The fundamentals of PS-OCT are detailed in numerous publications,^{28, 29} and the system used in this study has been described previously.¹⁸ PS-OCT measures the intensity and polarization state of backscattered light as a function of depth at a specific lateral position. In this study, the PS-OCT system integrated an optical coherence domain reflectometer (OCDR), which measured the backscattered signal defined as an $a$ -scan, with a computer-controlled high speed $XY$ -scanning stage (Newport Corporation, Irving, California) to produce a 2-D optical tomographic image or $b$ -scan.

The all-fiber-based OCDR system (Optiphase Incorporated, Van Nuys, California) shown in Fig. 1 is based on a polarization-sensitive Michelson interferometer. Linearly polarized light at $1310\mathrm{nm}$ emitted from a diode source was coupled into the slow or parallel axis of a polarization-maintaining (PM) fiber, and the focused beam illuminated the tooth sample. The reflected/backscattered light from the sample was collected in two orthogonal polarization states of the PM fiber: the slow or parallel axis to the original incident plane and the fast or perpendicular axis. The OCDR used two high speed piezoelectric fiber stretchers, in a push-pull configuration, to alter the path length of the reference beam relative to the sample beam, and a $50∕50$ fiber coupler combined the two beams. A polarized beamsplitter (PBS) then split the two orthogonal components of the combined beam into the parallel and perpendicular axis signals. The two orthogonal signals were then measured by two balanced InGaAs receiver modules. The signals were then electronically demodulated and stored on a Labview™ controlled computer (National Instruments, Austin, Texas).

Fig. 1

PS-OCT system: 1310-nm light from the SLD is linearly polarized and coupled into the slow axis of a polarization-maintaining fiber and equally split between the sample and reference arms of a polarization-sensitive optical coherence domain reflectometer (OCDR). The path length difference between the sample and reference arms is varied using piezoelectric fiber stretchers. The quarter-wave plate (QWP) in the reference arm splits the linearly polarized light equally between both polarization states. A polarizing beamsplitter (PBS) in the detection arm splits the fast (perpendicular) and slow (parallel) axis components of the light onto the two detectors. The OCDR is integrated with a computer-controlled scanning stage to produce tomographic PS-OCT images.

The PS-OCT system used a $20\text{-}\mathrm{mW}$ broadband $1310\text{-}\mathrm{nm}$ superluminescent diode (SLD), (Covega, Jessup, Maryland). The incident beam was focused onto the sample surface using a $20\text{-}\mathrm{mm}$ focal length antireflective (AR)-coated plano-convex lens. This optical configuration produced a $30\text{-}\mu \mathrm{m}$ lateral resolution that was measured by translating a knife edge across the beam waist and then calculating the $1∕{\mathrm{e}}^2$ diameter.³⁰ The SLD source possessed a spectral bandwidth [full width at half maximum (FWHM)] of $50\mathrm{nm}$ that produced within the system an axial resolution of $20\mu \mathrm{m}$ in air. The system axial resolution is increased to as much as $11\mu \mathrm{m}$ when imaging the higher refractive index enamel tissue $(n=1.63)$ . Using a single reflective signal from a mirror, the signal-to-noise ratio (SNR) of the PS-OCT system was measured to be $40\mathrm{dB}$ .

2-D OCT $b$ -scans were obtained by collecting a series of depth-resolved signals by laterally scanning the beam across $12\mathrm{mm}$ of the wet tooth. The $5.74\text{-}\mathrm{mm}$ axial depth of each image was controlled by the maximum path length difference between the sample and reference beams at an $a$ -scan rate of $150\mathrm{Hz}$ . Each full $b$ -scan was done in less than one minute. The acrylic blocks attached to each tooth was fitted into a custom metal jig that allowed orientation and coordinates to be precisely replicated, within $50\mu \mathrm{m}$ , for every scan.³¹ The fixed coordinate position and a reference focus point of sound enamel were necessary to ensure that both the demineralization and remineralization samples were placed into a fixed position.

The PS-OCT $b$ -scan images were acquired at day 0 for the 40 teeth in the study. After either 3 or $10\text{days}$ of pH cycling to create the artificial lesions, PS-OCT images were again collected. Demineralized caries lesion in each group $(n=10)$ were saved for histological evaluation, and the remaining teeth were again scanned with PS-OCT following $20\text{days}$ of remineralization. A total of eight serial $b$ -scan images were acquired for each tooth at $400\text{-}\mu \mathrm{m}$ intervals, which encompassed a $2.8\times 12\text{-}\mathrm{mm}$ area of the exposed tooth surface. A series of line profile integrations over the depth of each lesion that encompassed the entire demineralized area were averaged and used to assess the overall lesion severity. This integration of the line profiles, a summation of the logarithmic reflectivity (dB), is presented in arbitrary units (a.u.). Using a single line profile of the lesion, the real lesion depth was determined by dividing the measured optical depth by the index of refraction of the enamel $(n=1.63)$ .

2.3.

Histological Analysis (Digital Microradiography)

Digital microradiography (DM) of $200\text{-}\mu \mathrm{m}$ -thick transverse tooth sections was used to measure the mineral loss or gain from each of the four groups $(n=6)$ : the 3 and $10\text{day}$ demineralized caries lesions, and the 3 and $10\text{day}$ fluoride remineralized lesions. High resolution microradiographs were taken using Cu $\mathrm{K}\alpha$ radiation from a Philips 3100 x-ray generator. A Photonics Science FDI x-ray digital imager (Microphotonics, Allentown, Pennsylvania), consisting of a $1392\times 1040\text{-}\text{pixel}$ interline charge-coupled device (CCD) directly bonded to a coherent microfiber optic coupler that transfers the light from an optimized gadolinium oxysulphide scintillator to the CCD sensor, was used to acquire the microradiographic images. Images were acquired in real time at a frame rate of $10\mathrm{fps}$ . The image size was $2.99\times 2.24\mathrm{mm}$ with a pixel resolution of $2.15\mu \mathrm{m}$ . The imaging system consisted of a high speed motion control system with Newport UTM150 and 850G stages and an ESP300 controller coupled to a video microscopy and laser targeting system for precise positioning of the tooth sections in the field of view of the imaging system. The volume percent (%) mineral content of each section was determined by comparison with a calibration curve of x-ray intensity versus sample thickness created using sound enamel sections of $86.3\pm 1.9\%$ volume percent mineral varying from $50\text{to}300\mu \mathrm{m}$ in thickness. The calibration curve was validated via comparison with cross sectional microhardness measurements. The volume percent mineral determined using microradiography for section thickness ranging from $50\text{to}300\mu \mathrm{m}$ highly correlated with the volume percent mineral determined using microhardness $(R^2=0.99)$ .

From the DM technique, we obtained the quantitative mineral loss profiles³² taken normal to the outer enamel surface. The relative mineral loss $(\mathrm{vol}\%\times \mu \mathrm{m})$ $\Delta Z$ was calculated as the difference between the sound and lesion profiles on the same sample. Lesion depth was measured from the outer surface, where the mineral volume equaled 20%, to a position where the volume reached a level equal to 95% of the sound enamel.³³

2.4.

Synchrotron Radiation Fourier Transform Infrared Spectroscopy

High resolution IR spectra of laser-irradiated surfaces were obtained using a Nicolet Magna 760 Nic-Plan IR microscope (Nicolet, Madison, Wisconsin) interfaced to Beamline 1.4 at the Advanced Light Source at Lawrence Berkeley National Laboratory.³⁴ The high brightness provided by the synchrotron radiation source allows IR spectra to be acquired with a spatial resolution of $10\mu \mathrm{m}$ . The specular reflectance spectra were acquired by scanning the $10\text{-}\mu \mathrm{m}$ spot imaged by the Fourier transform infrared (FTIR) microscope across the cross sectional surfaces of selected remineralized lesions from each group.

3.1.

Remineralization of Severe Early Artificial Caries

PS-OCT $b$ -scan images, in both the parallel and perpendicular axes, of the $10\text{day}$ pH cycled demineralized caries lesions are presented in Fig. 2 . The reflectivity in the parallel axis shows the intense scattering of the demineralized lesion compared to the bordering sound enamel. In the perpendicular axis, the caries lesion rapidly depolarizes the incident polarized light, and the signal is markedly greater than that of the signal in the sound enamel caused by the native birefringence. After exposure to the fluoride remineralization solution for $20\text{days}$ , the lesion demonstrated an increase in both scattering and depolarization. Integration of the perpendicular-axis line profiles of the lesion for both the demineralized samples and the samples exposed to the remineralization solution are presented in Table 1 . The paired data, from identical samples, showed that the caries lesions exposed to $20\text{days}$ of remineralization had a higher integrated reflectivity compared to the demineralized lesions (repeated-measures analysis of variance, $\mathrm{p}<0.05$ ). The integrated reflectivity of the sound enamel in both groups was equal. The estimated real lesion depths, using the index of refraction of sound enamel, showed no difference in depth between the two groups.

Fig. 2

PS-OCT $b$ -scan images of a $10\text{day}$ pH cycled artificial caries lesion before and after a $20\text{day}$ regimen of fluoride remineralization. Images are displayed in a false-color scale (bottom). (a) The blue scale bar is $1\mathrm{mm}$ of optical depth, and the dentin-enamel junction (DEJ) is labeled. Parallel-axis images of (a) the demineralized caries lesion and (c) the remineralized lesion show an increase in backscattering of the lesion area after remineralization. (b) Perpendicular-axis images of the demineralized and (d) remineralized lesion show both the increase of backscattering and depolarization after remineralization. The reflectivity increase in the perpendicular axis is quite evident near the surface of the lesion.

Table 1

Lesion assessment with PS-OCT of the 10day lesions.

The high resolution digital micrographs of the $10\text{day}$ lesions, obtained after cross sectioning, indicated that these teeth were severely demineralized with a large reduction in the volume percent mineral versus the sound samples [Figs. 3a and 3b ]. The quantitative mineral content profiles from the lesion and sound enamel were acquired [ Figs. 3c and 3d]. From these profiles, it is apparent that the artificial process of creating the $10\text{day}$ lesions produced an intact $15\text{-}\mu \mathrm{m}$ -thick “surface zone” of higher mineral volume near the enamel surface. After the $20\text{day}$ fluoride remineralization, samples showed higher volume percent mineral within the entire lesion [Figs. 3b and 3d]. The quantitative mineral content profile illustrated the increase in the mineral volume, with the most dramatic increase near the surface of the lesion. Within a $40\text{-}\mu \mathrm{m}$ -deep surface zone region, the volume percent mineral of the remineralized lesion was restored to close to that of sound enamel. The relative mineral loss $\Delta Z$ was calculated from the mineral volume profiles to quantify lesion severity for each sample. The $\Delta Z$ and lesion depths from the $10\text{day}$ lesions and the fluoride remineralized lesions are summarized in Table 2 . The remineralized samples showed a statistically significant (student’s $t$ test, $n=6$ , $\mathrm{p}<0.02$ ) reduction in the $\Delta Z$ value compared with the $10\text{day}$ lesions. This indicated that the samples remineralized with new mineral and the lesion surface zone showed significant restoration of mineral volume. The lesion depth slightly decreased with remineralization (student’s $t$ test, $n=6$ , $\mathrm{p}<0.05$ ) but only to a marginal degree.

Fig. 3

High resolution digital micrographs of (a) the $10\text{day}$ pH cycled lesions and (b) the lesions after $20\text{day}$ fluoride remineralization treatment were obtained after tooth sectioning. The quantitative mineral content profiles from (c) the demineralized lesion and (d) remineralized lesion were acquired, in addition to the sound enamel profiles. The relative mineral loss $(\mathrm{vol}\%\times \mu \mathrm{m})$ $\Delta Z$ was calculated as the difference between the sound and lesion profiles on the same sample.

3.2.

Remineralization of Mild Early Artificial Caries

PS-OCT $b$ -scan images of the $3\text{day}$ demineralized caries lesions are presented [Figs. 4a and 4b ]. The images in both axes show that the $3\text{day}$ lesions are much shallower in depth than the lesions created after $10\text{days}$ of pH cycling. Before these lesion samples were placed in the fluoride remineralization solution, the bordering sound enamel, which is normally protected by a varnish, was deliberately exposed to the solution as well [Fig. 4d, arrows]. After $20\text{days}$ , it is evident [Figs. 4c and 4d] that a highly scattering and depolarizing outer layer of mineral approximately $30\mu \mathrm{m}$ thick is formed on the surface of the lesion and extended to the area of the exposed sound enamel. The boundary between this outer layer and the underlying lesion is clearly resolved in the perpendicular axis, due to the minimal effects of surface reflection in this axis.

Fig. 4

PS-OCT $b$ -scan images of the $3\text{day}$ pH cycled artificial caries lesion are presented in (a) the parallel and (b) perpendicular axes. The lesions exposed to $20\text{days}$ of fluoride remineralization are shown in (c) the parallel and (d) perpendicular axes. Before these lesions were remineralized, the bordering sound enamel, which is normally protected by a varnish, was also deliberately exposed to the solution [d, arrows]. After remineralization, a highly scattering [(c) and (d)] and depolarizing (d) outer growth layer $\left(30\mu \mathrm{m}\right)$ is formed on the surface of the lesion and extended to the area of the exposed sound enamel.

After carefully sectioning the samples, the DM images of the remineralized samples (Fig. 5 top) showed a highly mineralized outer $35\text{-}\mu \mathrm{m}$ -thick layer with a mineral volume content equal to that of the sound enamel. The presence of this outer layer confounds an accurate lesion severity analysis, since $\Delta Z$ is a product of both mineral volume and depth. The remineralized lesions have a larger depth than the demineralized lesions, because of the addition of the outer surface layer. Synchrotron radiation IR spectroscopy was employed to identify the chemical composition of this outer layer. Spectra were acquired from the different regions of the remineralized $3\text{day}$ lesion (Fig. 5 bottom). The first spectra [Fig. 6a ] represents the normal sound enamel with the three characteristic peaks due to the phosphate group in carbonated hydroxyapatite near $1000{\mathrm{cm}}^{-1}$ , and the two smaller peaks near $1400{\mathrm{cm}}^{-1}$ due to the carbonate group. The carbonate group is absent in the second spectra [Fig. 6b] of the $3\text{day}$ lesion body. The apatite phosphate bands are also altered from both the initial preferential dissolution of carbonated enamel crystals and the subsequent remineralization from the pH cycling mineralizing solution. The spectra of the remineralized outer layer [Fig. 6c] displays the characteristic apatite bands, with an absence of the carbonate group, and the narrowness of the phosphate band indicates this layer has high crystallinity.

Fig. 5

(Top) A high resolution digital micrograph image, after transverse sectioning, of the $3\text{day}$ pH cycled artificial lesion after the $20\text{days}$ of fluoride-enhanced remineralization illustrates the high mineral volume of the outer growth layer (arrows). (Bottom) A $500\times \text{microscope}$ image of the various regions of the remineralized lesion: (a) sound enamel, (b) lesion body, (c) outer remineralization layer, and (d) the embedding resin for sample analysis. The bordering sound enamel (arrow) near the lesion illustrates that the remineralization layer is formed above the surface.

Fig. 6

Spectra by FTIR microscopy were acquired from the different regions of the remineralized $3\text{day}$ lesion. (a) Sound enamel spectra with the characteristic peaks due to the phosphate group in carbonated hydroxyapatite near $1000{\mathrm{cm}}^{-1}$ and the two smaller peaks near $1400{\mathrm{cm}}^{-1}$ due to the carbonate group. (b) Spectra of the lesion body shows that the carbonate group is absent, which is likely due to the preferential loss of carbonate during demineralization. (c) The outer remineralization layer displays narrow characteristic hydroxyapatite bands, indicating high crystallinity, with an absence of the higher solubility carbonate substitution group.

Table 2

Lesion assessment with digital microradiography of the 10day lesions.