Several studies over the past few years have demonstrated that optical coherence tomography (OCT) can be used for the assessment of the severity of early caries lesions. OCT is a noninvasive technique for creating cross sectional images of internal biological structure.1, 2, 3 It is nondestructive and has great potential for imaging in vivo. Polarization-sensitive OCT (PS-OCT) is a form of OCT that is sensitive to changes in the polarization of the reflected light.4 PS-OCT, has been successfully used to acquire images of both artificial and natural caries lesions, assess their severity in depth, assess the remineralization of such lesions on enamel,3, 4, 5 and determine the efficacy of chemical agents in inhibiting demineralization.5 Recent studies in our laboratory on dentin and cementum surfaces demonstrate that PS-OCT is also well suited to assess the severity of demineralization on those surfaces, as well as determine the efficacy of anticaries agents such as fluoride and lasers.6, 7
One of the most promising applications of PS-OCT is the nondestructive measurement of the remineralization of existing caries lesions. In-vitro measurements of the remineralization of artificial lesions produced on enamel surfaces demonstrated that the thickness of the thin layer of higher mineral content that is typically formed near the lesion surface during the remineralization process can be measured with PS-OCT.8, 9, 10 This layer is of considerable importance, since the formation of this layer of fluoroapatite limits diffusion into the lesion leading to the arrest of lesion progression, and the lesion becomes inactive and there is no further need for intervention.
Polarization-sensitive depth-resolved reflectivity measurements can provide a measure of the severity of natural and artificial caries lesions on smooth surfaces and in occlusal pits and fissures.11, 12, 13, 14 The high reflectivity of dentin and enamel surfaces produces a very strong reflection that can interfere with the measurement of early demineralization on the surface. This strong reflection can be reduced using polarized light. Moreover, demineralized areas on the tooth depolarize the incident linearly polarized light, and the reflectivity in the orthogonal or perpendicular polarization state to the incident polarization can be directly integrated to provide a measure of the lesion severity.
Although the penetration depth of near-IR light is considerably less in dentin than for enamel due to the markedly higher scattering coefficient of dentin, PS-OCT and OCT images of dentin with root caries show that penetration depths of can be achieved, demineralized dentin can be discriminated from sound dentin and cementum,15 and root canal walls and fractures can be imaged.16, 17 In contrast to sound enamel that is highly transparent in the near-IR, sound dentin strongly scatters light in the near-IR and is highly birefringent.18, 19 Even with this limitation, we have successfully demonstrated that a similar approach to that used to assess the severity of demineralization on enamel surfaces can be applied to the dentin surfaces as well.6, 7 Amaechi 20 measured the percent reflectivity loss due to demineralization on dentin surfaces, and showed that this correlated well with mineral loss from microradiography. However, the validity of an approach that relies on the measurement of reflectivity loss is questionable, since light scattering increases with demineralization for both enamel and dentin, producing an overall increase in reflectivity from the area of the lesion, not a decrease in reflectivity. The approach that we have chosen is to measure the reflectivity from each layer of the lesion and subsequently integrate over a specific depth to yield the integrated reflectivity in units of , which is compared with the “gold standard”—the integrated mineral loss that is calculated in a similar fashion by integrating the mineral loss over a specific depth.11, 12, 13, 14, 21
It has been well established that topical fluoride agents can inhibit the demineralization of enamel, dentin, and cementum.22 Current methods that are used to determine the efficacy of anticaries agents require teeth scheduled for extraction and partial destruction of the tooth through sectioning. These methods include microradiography, microhardness, and polarized light microscopy.
In summary, the principal objective of this study was to demonstrate that a significant reduction in the reflectivity from the area of dentin lesions can be measured using PS-OCT after exposure of those lesions to a remineralization solution. PS-OCT measurements were also compared with the well established methods of transverse microradiography and polarized light microscopy that require tooth extraction and thin sectioning.
Materials and Methods
15 dentin blocks, approximately , were prepared from extracted human teeth for this study. The dentin surfaces were serial polished to one micron using embedded diamond polishing disks. Two windows on each sample were exposed to the demineralization solution, and one window was exposed to the remineralization solution (Fig. 1 ). The middle partition of each sample and both ends was covered with a thin layer of acid-resistant varnish red nail polish (Revlon, New York) to protect the sound dentin control area before exposure to the demineralization solution. Samples were placed in a demineralization solution for six days with both windows exposed consisting of a aliquot of calcium, phosphate, and acetate at pH 4.9, kept at a constant temperature of . One window was covered with acid-resistant varnish, leaving the other window exposed. The samples were placed in a solution for remineralization consisting of calcium, phosphate, KCl, and cacodylate buffer maintained at pH 7.0 and for . Fluoride, as NaF, was added at the level of F to enhance remineralization.23
Polarization Sensitive Optical Coherence Tomography System
A single-mode fiber autocorrelator-based optical coherence domain reflectometry (OCDR) system with polarization switching probe, high efficiency piezoelectric fiber stretchers, and two balanced InGaAs receivers that was designed and fabricated by Optiphase, Incorporated (Van Nuys, California) was integrated with a broadband high power superluminescent diode (SLD) (Denselight, Jessup, Maryland), with an output power of , a bandwidth of , and a high-speed -scanning system, ESP 300 controller, and 850-HS stages (Newport, Irvine, California), and used for in-vitro optical tomography. The system was configured to provide axial resolution at in air and in dentin, and a lateral resolution of approximately over the depth of focus of . The all-fiber OCDR system is described in more detail in Ref. 24 The PS-OCT system was completely controlled using LabVIEW™ software (National Instruments, Austin Texas).
Polarized Light Microscopy and Transverse Microradiography
After the dentin samples were imaged with PS-OCT, they were cut into sections approximately thick using an Isomet 5000 saw (Buehler, Lake Bluff, Illinois) for polarized light microscopy (PLM) and digital transverse microradiography (TMR). PLM was carried out using a Meiji Techno RZT microscope (Meiji Techno Company, Limited, Saitama, Japan) with an integrated digital camera EOS Digital Rebel XT (Canon Incorporated, Tokyo, Japan). The sample sections were imbibed in water and examined in the bright-field mode with crossed polarizers and a red I plate with retardation. A custom-built digital transverse microradiography (TMR) system was used to measure mineral loss in the different partitions of the sample.25
Integrated Reflectivity, Quantitative Mineral Loss Profiles, and Lesion Depth Measurements
The integrated reflectivity and integrated mineral loss were calculated in each of the six areas on the sample. Line profiles were taken from the orthogonal polarization (⊥-axis) PS-OCT images or b-scans in each of the six regions. The reflectivity was integrated from the dentin surface to a real depth of , yielding the integrated reflectivity of the regions in units of . The real depth was calculated by dividing the optical depth by the refractive index of dentin, 1.5.26 The lesion depth was estimated by taking the depth at which the intensity dropped to a level equivalent to , the peak intensity. Previous studies of demineralization on enamel surfaces have shown that correlates directly with the integrated mineral loss (vol. % ) called .6 Similarly, line profiles were taken from the digital TMR of the thin sections cut along the same position scanned by OCT. X-ray attenuation was converted to volume percent mineral using a sound enamel calibration curve. The line profiles in the four treatment (lesion) areas were subtracted from the sound mineral profiles to yield the integrated mineral loss. Demineralized lesions will shrink in dry conditions, as is visible in PS-OCT and TMR scans; hence, the line profiles were integrated from an extrapolated edge of the sample to a depth of to give units of . The integrations were performed using Igor Pro software (Wavemetrics, Lake Oswego, Oregon). Prior to taking the line profiles for the OCT images, a convolution filter was applied to reduce speckle in the images. Sample groups were compared using analysis of variance (ANOVA) with a Tukey-Kramer post-hoc multiple comparisons test. A paired t-test was used when only two groups were under consideration. Instat™ from (GraphPad Software, San Diego, California) was used for the statistical calculations.
Reflected light images of one of the dentin samples after exposure to the demineralization and remineralization solutions is shown in Fig. 1. The demineralization window (D) appears whiter or more opaque than the remineralization window , as would be expected if additional mineral has been deposited in the area exposed to remineralization. PS-OCT scans were acquired along the long axis of each sample near the center and sectioned for PLM and TMR in the same orientation.
Figures 2 and 3 each contain a (⊥-axis) PS-OCT b-scan image [Figs. 2a and 3a], the corresponding TMR [Figs. 2b and 3b], and PLM images for thin sections cut from the same sample with a similar orientation [Figs. 2c and 3c]. A red-white-blue false color scheme is used for the PS-OCT images with the intensity scale in decibel units (dB), where red and white areas have the highest reflectivity. A similar red-white-blue false color scheme was also used for the volume percent mineral content in the TMR images. The reflectivity of sound dentin is relatively high due to its high scattering coefficient. The lower range of the intensity scale for the PS-OCT images was intentionally set to maximize the contrast of the demineralized dentin in Figs. 2 and 3. Therefore, the sound areas appear blue. The demineralization was most severe in the region marked (D) as anticipated, while the areas exposed to the remineralization solution (R) typically had a layer of intact mineral of enhanced mineral content.
Dentin contains a high percentage of collagen, and demineralized dentin tends to shrink considerably when exposed to air, i.e., if it is not immersed in water. This can be seen in Figs. 2 and 3. PS-OCT and TMR were carried out with wet samples in air, while for PLM the samples were immersed in water. There is considerable shrinkage in the demineralized regions for the PS-OCT and TMR images, while the PLM images show minimal shrinkage in those regions. It is interesting to note that shrinkage typically did not occur in the remineralized window (R), since the highly mineralized intact surface zone prevented loss of water from the lesion area. The intact surface area was highly transparent when viewed under polarized light microscopy and had a high mineral content, mineral. This is 10% higher than sound dentin.
Individual a-scans were chosen from the center of each of the two areas on each sample. The mean integrated reflectivity in units of was calculated by integrating the reflectivity in the orthogonal polarization (⊥) image in units of decibels (dB) to a depth of . The volume percent mineral loss was measured in the same location for TMR and integrated over the same depth, , to yield the integrated mineral loss in . The values for the demineralization (D), sound , and the remineralization groups are displayed in Fig. 4 . The lesion depth was also measured with PLM and PS-OCT, and these measurements are tabulated in Table 1 for the remineralization and demineralization groups, respectively.
PS-OCT, PLM, and TMR measurements for all six groups’ mean±SD . Groups in the same row with a different letter are significantly different (p<0.05) n=15 .
The dentin shrinkage introduces an error in , since the calibration curve used to determine mineral content is not accurate for mineral content below 10%.25 Hence, shrinkage of the lesions (see Figs. 2 and 3) leads to an overestimation of the mineral loss by TMR. The values for the demineralization (D), sound (S), and remineralization (R) groups are displayed in Fig. 5 .
The lesion depth was also calculated from OCT images and polarized light microscopy, and those values are tabulated in Table 1. The lesion depths in the demineralization (D) and remineralization (R) zones, as well as the estimated thickness of the surface layer of highly increased mineral content in the remineralization zone , agreed quite well between the three methods of PS-OCT, TMR, and PLM, as can be seen in Fig. 6 . The of the thickness of the layer of increased mineral content was measured to be using TMR, using PS-OCT, and using PLM. Correlation plots of the lesion depth measured using PS-OCT versus TMR and PLM are shown in Figs. 7 and 8 . These indicate that the three methods manifest a reasonably high correlation. Depths from areas of both demineralization and remineralization are combined in the correlation plots.
The reflectivity from the artificial lesion areas exposed only to the demineralization solution were significantly higher than the lesion areas exposed to the same demineralization solution followed by subsequent exposure to a remineralization solution for . This demonstrates that PS-OCT can successfully be used to measure the deposition of new mineral in lesion areas, particularly near the surface. In many of the lesion areas exposed to remineralization, there was a highly transparent zone (PLM) of high mineral content that was deposited on the surface of the lesion. The mineral content of that outer zone was more than 60% mineral by volume, which is greater than sound dentin. The high transparency and mineral density of this zone suggests the growth of a highly oriented crystalline phase of apatite. A simple precipitation of calcium phosphate or hydroxyapatite on the surface would produce a rough and opaque white polycrystalline deposit. We have observed both transparent and opaque surface layers formed on the surface of enamel after exposure to a similar regimen of remineralization.9 PLM images show that the layer of higher mineral content extends above the original dentin surface, which may suggest that the layer may be a precipitate on the lesion surface. However, the surface layer had sufficient mechanical integrity to remain intact during the sectioning process, which indicates that the newly deposited mineral layer is well integrated into the underlying dentin. It is interesting to note that the zone of remineralization is thicker and more pronounced on dentin compared to enamel based on previous imaging studies.8 Moreover, a definitive zone of remineralization was present on virtually all the dentin samples, while the formation of the outer remineralization layer was less consistent on enamel.
One can postulate that the presence of the outer zone of high mineral content indicates that the active lesion has become arrested, and that further chemical intervention is not necessary. If this is validated clinically, then PS-OCT imaging would present a valuable nondestructive method for assessing nonsurgical intervention.
The excellent agreement among the respective lesion depths and the thickness of the zone of remineralization measured using PLM, TMR, and PS-OCT further demonstrates the diagnostic potential of PS-OCT for root caries lesions.
In conclusion, this study demonstrates that PS-OCT can be used to measure remineralization on dentin surfaces and detect the formation of a zone of highly mineralized dentin on the lesion surface after exposure to a remineralization solution. Further work is needed to demonstrate that PS-OCT can be used to measure the remineralization of natural caries lesions both in vitro and in vivo.
The authors acknowledge the support of NIH grants R01-DE017869 and R01-DE14698 and the help of Dennis Hsu and Chulsung Lee.