Open Access
Add to BibSonomy
Abstract
Polarized light microscopy (PLM) is an established method in dental histology for investigating the ultrastructure and carious process of teeth. This study introduces a novel approach for measuring the degree of polarization (DOP) in a modified PLM setup and uses the DOP to assess the changes of the optical properties of enamel and dentin due to caries. The validation is provided by a comparison with complementary imaging methods, i.e. standard PLM and µCT. The results show that demineralization is reliably displayed by the DOP in accordance with the common imaging methods, and that this quantitative analysis of depolarization allows the characterization of the different pathohistological zones of caries.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Research in the field of the ultrastructure of dental hard tissues, i.e. enamel and dentin, is of high interest in dentistry [1,2] as it paves the way towards future diagnostics and therapies. Enamel is the hardest substance in the human body. It consists of 95% hydroxyapatite, 1% organic matrix and 4% water [3]. The smallest structural unit of enamel is formed by hydroxyapatite (HAP) crystallites, which form enamel prisms. The enamel prisms run undulating from the enamel-dentin interface to the tooth surface. The orientation of the crystals is responsible for tremendous hardness and resistance of the enamel. Enamel is a cell-free, crystalline structure. Accordingly, a defensive reaction against the caries process is not possible [4]. By volume, dentin forms the largest part of the tooth. It consists of 70% hydroxyapatite, 20% organic matrix and 10% water [5]. The main part of the organic matrix is formed by collagen type I [3], collagen type III can also be found [6]. It is permeated by dentinal tubules, which contain the cellular extensions of odontoblasts located in the pulp. These are involved in defensive reactions of the endodont against the carious process. As a living tissue, dentin has a limited regenerative capacity [7].
Polarized light microscopy (PLM) is a form of microscopy that utilizes polarized light for the analysis of a sample [8]. For a propagating electromagnetic wave, the orientation of the oscillating electric field can be described by the polarization state. When the polarized light passes the sample, the polarization state is modified by the optical properties of the examined material [9,10]. Birefringent, diattenuating and depolarizing structures can be visualized by polarized light microscopy [11]. When examining dental hard tissues and the influence of the carious process, polarized light microscopy is an essential component of dental histology as dental hard structures are birefringent [12] and, thus, PLM can visualize changes in its HAP and collagen structures. Due to this fact, it already represented an important diagnostic tool in pathology in the last century [13]. Although the structure of tooth enamel is well understood, the exact orientation of the HAP crystallites remains insufficiently clarified [4].
The importance of polarized light microscopy for the examination of dental hard tissues and the caries process has been discovered almost 100 years ago [14]. Darling [15,16] examined the optical properties of dental hard tissues and discovered that cariously altered enamel leads to the change of birefringence as observed by the polarizing microscope. He interpreted the zones of different birefringence in cariously altered enamel as zones of different pore volume and made a major contribution to the understanding of the histopathology of caries through his studies in the middle of the last century [17]. Back then, carious lesions in enamel were often examined by using different immersion mediums to display different zones of enamel caries [18]. However, early carious lesions are still examined by using immersion mediums like quinoline in order to measure the actual depth of the enamel caries and to obtain conclusions about the pore volume [17,19]. Additionally, different approaches have been made to investigate the ultrastructure of dentin with polarized light microscopy. Advanced carious lesions have been qualitatively examined displaying the different histopathological zones in dentin [1]. It has been shown that changes in the collagen structure of dentin can be detected [20]. A recent study focussed on the investigation of the alignment of collagen fibrils in dentin [6].
Polarized light imaging holds promise for assessing biological tissues [21]. Several studies investigated the potential of linear and circular polarization imaging as well as the relation between depolarization and the examined scatterer sizes and scattering coefficient [22]. In general, it can be observed that a larger scattering leads to increased depolarization or a lower degree of polarization (DOP), respectively [23]. Several approaches have also been demonstrated towards depolarization imaging of demineralized dental hard tissues. The connection between changes of mineral content and depolarization in enamel were shown by measurements of transmitted light [24] and Mueller matrix imaging [25]. Furthermore, the scattering behavior in demineralized enamel was investigated in a recent study, indicating that polarization and photon pathway dynamics can be used to probe surface and subsurface scattering effects [26]. However, the DOP or depolarization index was not calculated in the latter imaging approaches, while previous DOP measurements were limited to single points. Additionally, the precise causes of depolarization in dentin remain relatively unresolved. [27,28]
While PLM is in general not applicable in vivo, other optical methods such as polarized light imaging [26] and optical coherence tomography (OCT) [29] have been demonstrated as promising tools for the in vivo assessment of dental hard tissue. Regarding polarization-related alterations in deeper subsurface enamel, especially polarization-sensitive OCT (PS-OCT) appears to be suitable towards caries monitoring in vivo. PS-OCT is an advanced imaging technology that provides high-resolution, cross-sectional images of dental tissues while also offering information about tissue birefringence and reflectivity [30]. This combination of capabilities makes PS-OCT particularly effective for detecting early-stage carious lesions [31]. We have recently shown that depolarization imaging based on the degree of polarization (DOP) derived from PS-OCT measurements is able to detect early stages of enamel demineralization [32–34] and that it facilitates the differentiation of surface staining and discoloration from carious lesions [35].
Here, we bridge the gap between PLM for the examination of thin sections and the evaluation of the DOP for depolarization imaging by introducing a method for PLM-based DOP measurements. As our method requires only minor modifications of a PLM setup for determining the depolarizing properties of a sample, it is introduced as depolarized light microscopy (DLM) and its capability of evaluating carious lesions in dental hard tissues is investigated. We therefore visualize occlusal caries lesions with different caries stages according to the ICDAS II classification [36] by using this advanced polarized light microscopy technique. For validation, we used X-ray micro-computed tomography (µCT), an in dental research established radiographic technique for ex vivo demineralization imaging [37], and the mineral content derived thereof. In general, our method aims at depolarization imaging, suitable for everyday application in ex vivo fundamental research, constituting an extension of a microscopy technique. Furthermore, our contribution includes the creation of an image catalog encompassing various stages of dental caries.
2. Materials and methods
2.1 Pathohistology of dentin caries
For the evaluation of the DLM images, we refer to the established pathohistological PLM interpretation as shown in Fig. 1. Before cavitation of the enamel surface occurs, a defensive reaction of the pulp-dentin unit takes place. This is manifested by tubular sclerosis in the dentin, the formation of tertiary dentin at the dentin-pulp interface and inflammation of the pulp [5]. Progression of caries, unless prophylactic measures are taken, leads to loss of tooth surface integrity due to enamel cavitation [38]. After penetration of the dentin by bacteria, the disease progresses much faster than in the enamel [39]. Bacterial penetration leads to demineralization and proteolytic liquefaction of the tissue. Phases of defensive reaction of the pulp-dentin unit alternate with phases of bacterial destruction due to carious progression [7]. The optical properties of enamel and dentin change due to alteration of the ultrastructure by the carious process [27,40,41]. According to the established literature, the following histopathological zones in the advanced dentin lesion can be displayed with transmitted light or bright-field microscopy and PLM including SE: sound enamel, DE: demineralized enamel, DD: demineralized zone in dentin, TD: translucent zone in dentin, SD: sound dentin [42].
Fig. 1. Pathohistology of caries based on the visualization of an occlusal carious lesion with conventional histological methods in dentistry: transmitted light microscopy (a) and polarized light microscopy (b). The following zones are displayed: SE: sound enamel, DE: demineralized enamel, $\rightarrow$: Dentin-Enamel-Junction, DD: demineralized zone in dentin, TD: translucent zone in dentin, SD: sound dentin.
2.2 Tooth samples and preparation
A total of 17 teeth with suspected occlusal caries but without loss of surface integrity was included in the study. Based on the guidelines set forth by the Central Ethics Committee at the German Medical Association, there is no need for ethical approval when utilizing teeth extracted for medical purposes in research [43]. All teeth were kept in thymol solution at ${7}^{\circ }$ Celsius to prevent the teeth from drying out. To compare slices extracted from $\mathrm{\mu}$CT volumes and microscopy images of the thin sections, a study design of subsequent imaging and preparation shown in Fig. 2 was developed. First, photographs of the occlusal surfaces (a) for documentation and $\mathrm{\mu}$CT volumes (SCANCO vivaCT 40, SCANCO Medical AG, Brüttisellen, Switzerland) as a reference procedure (b) were acquired of all teeth. As the $\mathrm{\mu}$CT device is calibrated on a regular basis and following the recommendations of the manufacturer, the mineral content could be derived from the $\mathrm{\mu}$CT images [44]. In addition, the teeth were classified according to ICDAS II (International Caries Detection and Assessment System) criteria, a scoring system that provides detection codes from 0 to 6 for occlusal caries based upon the severity of the lesion suspected during visual inspection [36]. After embedding the teeth in light-curing resin (Technovit 7200 VLC, Kulzer Technik GmbH, Hanau, Germany), thin sections were prepared using thin-section technique by Donath [45] with a thickness of approximately ${80}\;\mathrm{\mu}$ (Fig. 2(c)).
Fig. 2. Sample preparation and measurement procedure. Initially, for all intact teeth (a) photographs of the occlusal surface and (b) 3D volumes using X-ray micro-computed tomography ($\mathrm{\mu}$CT) were acquired. Subsequently, the samples were embedded in light-curing resin with all teeth positioned on their occlusal surface (c). Thin sections (d) were then prepared from each tooth for polarized light microscopy (e).
To be able to assign the correct section planes afterwards, additional OCT volumes (not shown) were acquired from the intact and embedded teeth’s occlusal surfaces. Using a previously described approach for the spatial registration of $\mathrm{\mu}$CT and OCT volumes [46], an alignment of $\mathrm{\mu}$CT and OCT of the intact teeth as well as OCT of intact and embedded teeth was possible. Eventually, identifying the sectioning plane from the embedded teeth was simplified by the positioning on the occlusal surface. Due to this step-wise imaging and embedding process, all imaging techniques including $\mathrm{\mu}$CT are presented in the same sectional plane using a custom MATLAB script (MathWorks, Natick, MA, USA).
2.3 Microscopy
A total of 24 selected thin sections was examined with a modified Leica DMRB polarizing microscope (Leica Microsystems GmbH, Wetzlar, Germany) using an objective with magnification of 2.5 times.
Fig. 3. Setup of the modified polarized light microscope and exemplary results. Broadband light from a halogen light source is linearly polarized by a polarizer prior to the sample (a, photograph of the thin section), which is followed by an analyzer section with different settings: (b) transmitted light microscopy using neither analyzer nor fullwave plate, (c) PLM in crossed configuration, i.e. the analyzer was rotated 90$^{\circ }$ relative to the polarizer, (d) PLM in red-plate configuration, i.e. with a fullwave plate at 45$^{\circ }$ prior to the analyzer. For depolarized light microscopy, the analyzer and fullwave plate were removed and raw polarization images (e) were acquired.
As demonstrated in Fig. 3, the thin sections were examined in three different settings: First, the thin sections were examined using the conventional cross-polarization configuration in polarized light microscopy (Fig. 3(c)), meaning that light rays passed immediately through the analyzer after passing the sample while the polarizer was perpendicularly oriented to the analyzer. Second, a fullwave plate at ${45}^{\circ }$ was placed in front of the analyzer, often referred to as red-plate configuration, which maximizes the color contrast by adding 550 nm retardation (Fig. 3(d)). Third, the analyzer and fullwave plate were removed and thus an examination using transmitted light microscopy was performed (Fig. 3(b)). The photographs during examination were acquired with a digital camera (Sony $\alpha$7, Sony, Tokyo, Japan) using an exposure time of 1/160 s, ISO sensitivity of 125 and white balance of 3200 K.
Finally, advanced PLM (Fig. 3(e)) was performed using a 532 nm band-pass filter (FL 532-10, Thorlabs, Newton, NJ, USA), 0.67x video relay optics (EO58-379 and EO58-377, Edmund Optics, UK), and a USB polarization camera (BFS-U3-51S5P-C, FLIR Integrated Imaging Solutions Inc., Ludwigsburg, Germany). For this purpose, the thin section was fixed on the object stage while the polarizer was rotated manually by slightly more than ${180}^{\circ }$ starting from the ${0}^{\circ }$ position. An image sequence was recorded during the continuous and manual rotation whereby the polarization camera measured the transmitted linear polarization states for each position of the polarizer and thus illuminating linear polarization. The images were acquired at a gain of 1, 16 bit depth and an exposure time of 50 ms or frame rate of 20 Hz, respectively, using a custom LabVIEW program (National Instruments, Austin, TX, USA). An acquisition took about 20 s or 400 images, which corresponds to approximately ${0.5}^{\circ }$ steps assuming a nearly uniform manual rotation. The effective frame rate of about 20 Hz was practically limited by the implemented preview processing and data saving. All processing including the degree of linear polarization (DOLP) from each image and the final degree of polarization (DOP) representation of the thin section was done in MATLAB (MathWorks, Natick, MA, USA).
The entire collection of $\mathrm{\mu}$CT [47] and raw and processed microscopy datasets (polarized [48] and depolarized [49] microscopy) as well as photographs of the occlusal surface [50] of all 17 teeth are available at OpARA (Open Access Repository and Archive, TU Dresden, Dresden, Germany, https://opara.zih.tu-dresden.de).
2.4 Calculation of depolarization images
As visualized in Fig. 3, the polarization camera provides a 2 x 2 pattern of polarization filters attached to adjacent pixels on the sensor (Sony IMX250MZR) with ${0}^{\circ }$, ${45}^{\circ }$, ${90}^{\circ }$ and ${135}^{\circ }$ or horizontal (H), diagonal (D), vertical (V) and anti-diagonal (A) orientation, respectively. As such a pattern is similar to standard color image cameras with an RGGB or Bayer pattern of spectral filters for red (R), green (G) and blue (B), interpolation algorithms developed for restoring an RGB color triplet for each pixel, the debayering or demosaicing, would be applicable also for the reconstruction of the polarization image from the polarizer array. For the sake of simplicity, neighboring pixel of a 2 x 2 cluster were combined here to form a polarization pixel and thereof a polarization image (Fig. 4) with the array size width by height by linear polarizations (H,D,V,A). From the measurement of the linear polarization states, at least three out of four Stokes parameters can be calculated by
Fig. 4. Measurement and calculation of the DOP. (a) From the acquired images with a polarizer pattern on the sensor, linear polarization images for horizontal (H), diagonal (D), anti-diagonal (A) and vertical (V) polarization were extracted. Following the conversion to the linear Stokes parameters $S_0$, $S_1$ and $S_2$, the degree of linear polarization (DOLP) can be calculated for each individual acquisition. (b) By rotating the polarizer in the illumination path of the microscope, multiple DOLP images can be acquired with varying incident states. After median filtering of the sequence, a maximum projection extracts the DOP.
To evaluate the correlation of the intensity of the transmitted light and the depolarization, Fig. 6 compares the values for the DOP (a) and the intensity (b) based on the advanced PLM measurement and shows the benefit of depolarized light microscopy. While the DOP often seems to be related to the polarization-independent intensity ($S_0$), some regions, especially at 0.5 mm of the plotted section show low intensity (Fig. 6(c)) and high DOP (Fig. 6(b)). Therefore, the evaluation of the DOP provides added value compared to pure intensity evaluation, as it similarly analyzed in transmitted light microscopy.
Fig. 5. Exemplary results of a DOLP measurement series with an animated version available online (see Visualization 1). During manual rotation of the polarizer, the measured DOLP (a) varies as the provided measurement points (b) show. Besides three regions of varying DOLP, the AOLP was extracted from a region without birefringence in the upper left corner (a) and plotted for a total of 415 acquired images (b).
Fig. 6. Quantitative comparison with plot of the values (a) for DOP (b) and intensity (c) of the DLM results: Although the DOP and the amount of transmitted light, i.e. $S_0$ or the intensity, seem to be proportional or related, some regions appear with a low intensity while showing a high DOP as for example at 0.5 mm of the plotted section.
3. Results
The pathohistology of caries is initially displayed utilizing the developed depolarized light microscopy (DLM) technique in Fig. 7 and compared to $\mathrm{\mu}$CT for a quantitative interpretation in Fig. 8. Subsequently, the visualization of occlusal carious lesions with different microscopic techniques is presented for various teeth. The benefits and limitations of DLM are demonstrated by means of image panels, comparing this technique with conventional microscopic methods. In the following, these are structured into initial carious lesions with incipient demineralization (Figs. 9, 10 and 11) and advanced dentin caries (Figs. 12, 13 and 14).
Fig. 7. Pathohistology of caries in depolarized light microscopy. Representative DLM images of three thin sections with a stitched (a) and two single images (b,c) to display the different pathohistological zones of carious lesions. SE: sound enamel, DE: demineralized enamel, $\rightarrow$: Dentin-Enamel-Junction, DD: demineralized zone in dentin, TD: translucent zone in dentin, SD: sound dentin. The different zones can be distinguished by analyzing the DOP within the thin sections. Demineralized areas in enamel can be clearly identified due to the black appearance in DLM with a DOP value less than 0.5. Both healthy and carious dentin do not exhibit a uniform depolarization signal, as sound dentin appears to be slightly depolarizing.
3.1 Pathohistology of dentin caries in DLM
DLM allows the characterization of the different pathohistological zones of caries with a quantitative approach by displaying the degree of polarization (DOP). The following zones can be distinguished: sound enamel (SE), demineralized enamel (DE), demineralized zone in dentin (DD), translucent zone in dentin (TD), sound dentin (SD) (Fig. 1). The appearance of the pathohistological zones with DLM is shown in Fig. 7. Sound and cariously altered enamel and dentin can be characterized by evaluating the DOP with a higher resolution and contrast than in $\mathrm{\mu}$CT, which is often used as a reference procedure in studies examining caries lesions of intact teeth ex vivo. Depolarizing regions with birefringence in enamel can be displayed color-independently. As it can be seen in Fig. 7 for the pathohistology as well as in the methodological comparisons in Figs. 9–14, demineralized areas in the enamel can be precisely quantified by the value of the DOP. Healthy enamel exposes light gray in the DLM image with a DOP value of typically 0.9, but in general greater than 0.6. Cariously altered, demineralized enamel appears darker or black in the DLM image with a DOP value of about 0.5 or below. This enables the observer to clearly identify demineralized areas on first sight. Further histological subdivision of cariously altered enamel cannot be made on the basis of the DOP. The method offers insights into the optical properties of dentin. No clear statement can be made concerning the value of the DOP in the various zones. Demineralized dentin is located within a carious lesion directly beneath the DEJ, as depicted in Fig. 1. Figure 7 provides a more detailed representation of the individual zones, including demineralized dentin (DD). In clinical dentistry, this area is also referred to as peripheral dentin, as it is distant from the pulp and directly adjacent to the DEJ and thus the demineralized enamel. In most DLM images, this zone appears dark gray to black. However, the DOP in this zone is not the same for every thin section; it appears to vary. In some lesions, the dentin adjacent to the lesion also exhibits a strong depolarization signal. The translucent zone in dentin is directly adjacent to the dark gray to black demineralization zone and appears as a light gray, washed-out area. As it can be seen from Fig. 7, the zone can be clearly distinguished from demineralized dentin on the basis of the DOP. The distinction from healthy dentin in the vicinity of the pulp cannot always be clearly defined. For translucent dentin, a comparable DOP is most likely to be found in all lesions. The DOP is similar to healthy enamel. Sound dentin borders on the translucent dentin. As already described for demineralized dentin, no uniform DOP value can be determined for this zone. Presumably healthy dentin irregularly exhibits a slight depolarization, which can be observed in Fig. 7.
Fig. 8. Comparison of demineralization in DLM and µCT. The figure illustrates that the results of DLM and µCT correlate with regard to the extent of demineralization in enamel. This can be visualized and quantified by evaluating the DOP within the thin section and the mineral content derived from the µCT slice within the same plane. The size of the carious lesion seems to be more distinct and extended than in µCT. (a) depolarized light microscopy with marked plotting region, (b) µCT slice with marked plotting region, (c) comparison of DOP (blue) from the DLM image and mineral density as determined from the µCT slice (red).
The results of DLM and $\mathrm{\mu}$CT correlate with respect to the spatial extent and degree of progression of demineralization in enamel. This is presumed when comparing the DOP value in depolarized light microscopy and the mineral density derived from the µCT slice within the same plane. Figure 8 displays the extent of demineralization in carious enamel in direct comparison of depolarized light microscopy and $\mathrm{\mu}$CT volume. The figure indicates that the extent of demineralization can be visualized and quantified by evaluating the DOP within the thin section, and that it correlates with the reduced mineral content of the lesion. The size of the carious lesion seems to be more distinct in DLM than in $\mathrm{\mu}$CT. However, a significant difference is apparent in Fig. 13, where no mineral loss can be observed in the enamel when analyzing the $\mathrm{\mu}$CT data while the DOP image shows a strong depolarization in enamel.
3.2 Visualization of initial carious lesions
For each tooth, four different imaging methods are displayed for the same sectional plane. The histological standard procedure of transmitted light microscopy is compared with polarized light microscopy and depolarized light microscopy. $\mathrm{\mu}$CT can be directly compared based on the corresponding sectional plane and provides a reference for demineralization in the carious lesion. The section of the carious lesion shown with depolarized light microscopy is framed in the other images. In the description of the figures, these are structured as follows: (a) $\mathrm{\mu}$CT slices with marked region of DLM, (b) transmitted light microscopy with marked region of DLM, (c) depolarized light microscopy (DLM), (d) occlusal photograph with marked slicing region, (e) polarized light microscopy, (f) polarized light microscopy with additional fullwave plate.
Fig. 9. Visualization of an occlusal surface with heavily stained fissures (ICDAS 4), under which one might suspect the progression of caries. Examination with PLM and DLM reveals a clinically and histologically healthy tooth with sound enamel and dentin. The imaging techniques are displayed within the same plane section and the region shown with depolarized light microscopy is framed in the other images. (a) $\mathrm{\mu}$CT volume with marked region of DLM, (b) transmitted light microscopy with marked region of DLM, (c) DLM based on the DOP (d) occlusal photograph with marked slicing region, (e) PLM, (f) PLM with additional fullwave plate.
Fig. 10. Visualization of an initial occlusal carious lesion (ICDAS 3) with typically opaque discolored cuspal inclines, representing the initial clinical manifestations of incipient caries. The examination with PLM and DLM reveals an initial enamel lesion with minimal dentin involvement. Refer to Fig. 9 for a detailed enumeration.
Fig. 11. Visualization of a chronic occlusal carious lesion (ICDAS 3) examined with $\mathrm{\mu}$CT, PLM and DLM. The characteristic indications of demineralized enamel on the cuspal inclines can be observed. The brownish discoloration hints at the prolonged nature of the initial lesion. The demineralization has advanced into the dentin, despite the surface remaining intact. Refer to Fig. 9 for a detailed enumeration.
In Fig. 9, a tooth with suspected occlusal caries is depicted. This represents a typical clinical case, showcasing heavily stained fissures, which often lead to overdiagnosis in clinical practice. The ICDAS code 4 assigned to this tooth reflects the potential for misinterpretation due to the presence of significantly discolored fissures, potentially mimicking dentin caries due to dark shadows underlying from dentin. Upon examination with $\mathrm{\mu}$CT, it becomes evident that there is no demineralized enamel present. This observation is further confirmed by the microscopy images, showing only slight abnormalities in transmission microscopy (b) and a mild reduction in the degree of polarization in DLM (c). The representation of initial demineralization is notably clearer in DLM compared to PLM (e). The tooth and its dental hard tissues enamel and dentin are entirely healthy apart from the slight, subclinical demineralization.
In Fig. 10 and Fig. 11, initial carious lesions are depicted, with Fig. 10 showing a localized occurrence of the demineralization in enamel while Fig. 11 illustrates demineralization affecting the entire enamel. Both teeth do have an ICDAS code (3), indicating localized discontinuities in the enamel surface. The difference lies in Fig. 10 representing an incipient initial carious lesion with white discoloration at the cusp tips whereas Fig. 11 depicts a chronic lesion, evident from the brownish discoloration. In Fig. 10, it is evident that the dentin is not yet affected by the demineralization, whereas in Fig. 11, it is. The advantage of DLM images (c) lies in their clear, color-independent representation of demineralization in the enamel, which can be distinctly visualized compared to PLM images (e). Comparing Fig. 10(b) and Fig. 10(c), it can be observed that in this case, the evaluation of DOP does not provide a significant advantage. When comparing Fig. 11(c) and Fig. 11(e), it becomes evident that the alteration of dentin underneath the Dentin-Enamel-Junction due to the carious process can be clearly delineated through the evaluation of DOP compared to PLM. From a clinical point of view, these stages of caries do not require an invasive treatment of caries removal, yet observation.
3.3 Visualization of advanced carious lesions
Figure 12 – 14 visualize examples of teeth with advanced carious lesions, specifically affecting the dentin to a greater extent. In Fig. 12, demineralization has already affected approximately half of the dentin, while the surface of the enamel is still intact. With an ICDAS code 2, indicating early signs of caries at the enamel surface in visual inspection (Fig. 12(d)), demineralization to such great extent as revealed in the microscopy images would not have been expected. Figure 13 illustrates a hidden caries. This means that the surface does not exhibit any clinical signs indicating a carious process (Fig. 13(d)), resulting in an ICDAS code 0, but the decay extends beneath the seemingly intact enamel into half of the dentin, as it can be seen in the $\mathrm{\mu}$CT image (Fig. 13(a)). Figure 14 represents a case where the surface begins to show slight breakdown, as ist can be seen in visual inspection (Fig. 14(d)). The ICDAS code 4 is confirmed by optical imaging techniques: demineralization into dentin is evident in the $\mathrm{\mu}$CT scans as well as in the thin sections.
Fig. 12. Visualization of an initial occlusal carious lesion examined with $\mathrm{\mu}$CT, PLM and DLM. Visual inspection shows opaque fissures, which is a sign of initial demineralization and thus indicates an incipient carious lesion (ICDAS 2). The extent of dentin involvement is greater than anticipated from visual inspection. The examination of the thin sections reveals demineralization into the dentin without loss of surface integrity. The extent of the lesion is consistent with the $\mathrm{\mu}$CT data. Refer to Fig. 9 for a detailed enumeration.
Fig. 13. Visualization of a hidden occlusal carious lesion examined with $\mathrm{\mu}$CT, PLM and DLM. Visual inspection shows a clinically almost inconspicuous, discolored appearing occlusal surface (ICDAS 0). The examination reveals a hidden carious lesion. Demineralization has already progressed into the dentin, whereas surface integrity is still preserved. Refer to Fig. 9 for a detailed enumeration.
Fig. 14. Visualization of an advanced occlusal carious lesion examined with $\mathrm{\mu}$CT, PLM and DLM. Visual inspection of the occlusal surface is indicative of a carious lesion with progression in dentin (ICDAS 4), which is confirmed by examination of the histological thin sections and $\mathrm{\mu}$CT volume. Refer to Fig. 9 for a detailed enumeration.
In all three caries lesions (Fig. 12 – 14), the zone of demineralized dentin directly adjacent to the DEJ can be recognized by reduced DOP. Compared to demineralized enamel areas, it is evident here that the reduction in DOP in demineralized dentin is not as pronounced as in demineralized enamel. Comparing the transmitted light microscopy and the PLM images, the advantage of DLM becomes apparent regarding the alteration of optical properties in dentin due to caries: the zones are well represented due to their depolarizing behavior. As a clinically practicing dentist, one would continue observation in Fig. 12 and Fig. 13, while in Fig. 14, intervention through caries removal would be necessary.
4. Discussion
We have developed and demonstrated a way of displaying dental caries in thin sections with polarized light microscopy by measuring the degree of polarization (DOP) in a microscopic imaging setup. As we thoroughly examined caries lesions in all stages of development, ranging from incipient to advanced, the study provides important information for the interpretation of changes in the optical properties of the dental hard tissues due to caries. It is shown that the carious changes in dentin have a different effect regarding the depolarization response than in enamel. Sound and demineralized enamel can clearly be distinguished color-independently based on the value of the DOP. It is therefore a major improvement to regular polarized light microscopy where these areas are unclear to delineate. As previously demonstrated within single-point measurements of thin section, the examination of carious lesions by evaluating the DOP is a promising method [24]. However, until now DOP imaging has mainly been performed in conjunction with polarization-sensitive optical coherence tomography (PS-OCT) [32–35,46]. By combining polarized light microscopy with DOP assessment, we offer a methodology for characterizing the depolarizing properties of carious dental hard tissues under microscopic examination, realizing depolarization imaging.
It is well known that the optical properties of dental hard tissues, i.e. scattering and attenuation, change due to the carious process [27] and that light polarization can be used to probe these changes [31]. However, optical alterations due to caries are much more pronounced in enamel than in dentin because the amount of birefringence in enamel is mainly attributed to anisotropic hydroxyapatite (HAP) crystallites and therefore depending on the orientation between the enamel prisms and the incident probe light [9]. Healthy enamel appears light gray in the DLM image with a DOP of about 0.9, which means that in these areas during the ${180}^{\circ }$ rotation of the polarizer, the optical axis of the HAP crystallites is hit at least once in the healthy enamel. Thus, the measured DOLP reaches a maximum and resembles the DOP. Cariously altered, demineralized enamel appears black in the DLM image with a DOP below 0.6, which is similarly indicated by previous research showing a reduced DOP in enamel lesions [24].
Sound enamel exhibits birefringence, which originates from the mineral as intrinsic birefringence with negative sign and from the non-mineral content as form birefringence with positive sign [17]. Accordingly, the observed birefringence is the sum of both contributions. Based on microscopic examination of carious thin sections and the application of different immersion media [17], it is know that demineralization causes not only a sign change in the observed birefringence from negative (healthy) to positive (carious), but also an increase in pore volume. While also this sign change can be used for assessing the lesion extend [52], the increased pore volume alters also the scattering behavior. Both surface roughness and scattering coefficient significantly increase [26,29] and by that cause an increased depolarization as shown within similar examinations of transmitted light [23], which is mediated by the collected multiply scattered light. Further analysis of depolarization caused by different scatterer sizes and scattering coefficients indicates that the linear and circular DOP can be used to access such sample properties in a back-scattering geometry [22]. Accordingly, an in-depth analysis of the DOP depending on circular or linear illumination in the presented configuration of transmission measurements needs to be performed in the future.
Similarly, this raises the question of whether there is a quantitative relationship between the pore volume of carious enamel and the DOP. While we coarsely investigated this assumption as shown in Fig. 8, differences in contrast and resolution limited the direct relation of DLM and $\mathrm{\mu}$CT. However, both this quantitative comparison and the qualitative evaluations for initial and advanced carious lesions suggest a strong relation between demineralization and thus reduction in pore volume, and increased scattering and thus higher depolarization.
For enamel, the following applies: demineralization causes depolarization. Otherwise, this conclusion cannot be unambiguously drawn for dentin. Dentin is a complex and heterogeneous dental tissue consisting of a network of tubules, collagen fibers, and hydroxyapatite crystals, all of which contribute to its intricate optical and structural properties [5]. Birefringence in dentin is dependent on inorganic (anisotropic HAP crystals) and organic (collagen fiber network) components [13]. It has been shown that the refractive index varies in dentin due to the orientation of the dentinal tubules [28]. Healthy dentin already exhibits a high scattering behavior [24], which makes it difficult to clearly distinguish between carious and healthy dentin [27]. This raises the question of what is responsible for depolarization in sound dentin; possibly collagen-fibrils, the density and arrangement of the dentinal tubules or the varying thickness of the thin sections. Demineralized dentin appears in the DLM image as a dark gray to black area in the peripheral dentin with a DOP value less than 0.5; consequently similar to demineralized enamel, which may indicate significant demineralization in the dentin. This observation would be in line with Arnold et al., who examined the mineral content in dentin with polarized light microscopy and scanning electron microscopy and described that carious lesions are surrounded by demineralized dentin [1]. The structure of HAP crystallites is massively altered in outer zones of carious dentin [40,41]. This could be a cause for the loss of birefringence in this area and an altered DOP signal. The ultrastructure of the dentin and thus the course of the dentinal tubules can still be seen in the depolarization image, which is consistent with the findings of Zavgorodniy et al. and Frank that the remaining crystals align along the collagen fibrils [2,40]. The DOP values of the different pathohistological zones in dentin correspond with previous studies conducted by Darling et al. [24,25]: It seems as if there is a constant slight background noise of backscattering present in dentin, whereas clearly demineralized areas in advanced lesions can be distinguished. The translucent zone in dentin can be clearly distinguished from demineralized dentin on the basis of the DOP as it appears as a light gray area. The DOP value is greater than 0.6 and therefore similar to healthy enamel. The translucent zone has a higher mineral content due to the sclerosis of the tubules, a physiological process where the dentinal tubules are blocked with mineral as a defense against the carious process [7]. It has been shown that in this zone the intertubular dentin is also demineralized compared to healthy dentin, while the dentinal tubules are blocked with mineral [40]. However, due to the intricate nature of dentin’s composition, an understanding how these factors interact and contribute to depolarization is challenging. Despite this pronounced change within the optical properties of enamel, ICDAS II histological validation appears to be feasible for clinically advanced caries in particular. This is in accordance with the conclusions of Braun et al. where ICDAS II classification was of no benefit in predicting enamel lesions, but certainly represents a standardized method in the assessment of advanced dentin lesions [53].
The limitations of this study include the small sample size and the qualitative nature of the study. The major disadvantage of depolarized light microscopy is that it can only be used for validation of thin sections on extracted teeth. With regular polarized light microscopy, different zones of carious enamel can be displayed. To this point, this is not possible with depolarized light microscopy. Regarding dentin, the effect of the collagen-fibrils nor the orientation of the dentinal tubules on the degree of polarization within a thin section cannot yet be described.
In summary, depolarized light microscopy represents a complementary tool for basic dental research by leveraging the depolarizing characteristics of dental hard tissues and implementing them within a novel imaging approach. Due to the fact of being an ex vivo procedure, DLM cannot be employed as a diagnostic method in routine day by day dental practice. Nonetheless, examining thin sections based on the DOP with spatial allocation to $\mathrm{\mu}$CT and other volumetric data, helps bridging the gap to non-invasive 3D assessment methods like PS-OCT. The results from fundamental research on the interpretation and validation of optical characteristics in dental caries will therefore serve the future development of diagnostic tools. In a follow-up study we will compare PS-OCT imaging and depolarized light microscopy (DLM) in the same section to evaluate benefits and limitations of PS-OCT for the detection of dentin caries.
5. Conclusion
Depolarized light microscopy (DLM) is an advanced technique of polarized light microscopy, which is considered to be the gold standard when examining carious lesions with different optical imaging techniques. The presented method provides an important contribution to basic research in the histopathological examination of carious lesions and offers additional insights concerning the optical properties of dentin. It bridges the gap to non-invasive assessment methods like polarization-sensitive optical coherence tomography (PS-OCT) as it can support the interpretation of future in vivo data.
Acknowledgments
We want to thank Diana Jünger for preparing the thin sections and the Bone Lab Dresden for $\mathrm{\mu}$CT imaging.
Disclosures
The authors declare no conflicts of interest.
Data availability
All relevant data, materials, and software code used in this research are available upon request from the corresponding author. Furthermore, the entire collection of processed $\mathrm{\mu}$CT [47] volumes and photographs of the occlusal surface [50] as well as raw and processed microscopy datasets (polarized [48] and depolarized [49] light microscopy) of all 17 teeth is provided at OpARA (Open Access Repository and Archive, TU Dresden, Dresden, Germany).
References
1. W. Arnold, S. Konopka, M. Kriwalsky, et al., “Morphological analysis and chemical content of natural dentin carious lesion zones,” Ann. Anat. - Anat. Anzeiger 185(5), 419–424 (2003). [CrossRef]
2. R. M. Frank, “Structural events in the caries process in enamel, cementum, and dentin,” J. Dent. Res. 69(2_suppl), 559–566 (1990). [CrossRef]
3. M. Shahmoradi, L. E. Bertassoni, and H. M. Elfallah, “Fundamental structure and properties of enamel, dentin and cementum,” in Advances in Calcium Phosphate Biomaterials, B. Ben-Nissan, ed. (Springer Verlag, 2014), chap. Fundamental Structure and Properties of Enamel, Dentin and Cementum, pp. 511–547, 1st ed.
4. E. Beniash, C. A. Stifler, and C.-Y. Sun, “The hidden structure of human enamel,” Nat. Commun. 10(1), 4383 (2019). [CrossRef]
5. J. Baranova, D. Büchner, and W. Götz, “Tooth formation: Are the hardest tissues of human body hard to regenerate?” Int. J. Mol. Sci. 21(11), 4031 (2020). [CrossRef]
6. A. T. Weerakoon, N. Condon, and T. R. Cox, “Dynamic dentin: A quantitative microscopic assessment of age and spatial changes to matrix architecture, peritubular dentin, and collagens types I and III,” J. Struct. Biol. 214(4), 107899 (2022). [CrossRef]
7. K. Galler, M. Weber, and Y. Korkmaz, “Inflammatory response mechanisms of the dentine-pulp complex and the periapical tissues,” Int. J. Mol. Sci. 22(3), 1480 (2021). [CrossRef]
8. R. Oldenburg, “Polarized light microscopy: principles and practice,” Cold Spring Harbor protocols 2013, pdb.top078600 (2013).
9. B. W. Colston, U. S. Sathyam, and L. B. DaSilva, “Dental OCT,” Opt. Express 3(6), 230–238 (1998). [CrossRef]
10. F. Massoumian, R. Juškaitis, M. A. A. Neil, et al., “Quantitative polarized light microscopy,” J. Microsc. 209(1), 13–22 (2003). [CrossRef]
11. S. Inoué, “Polarization microscopy,” Curr. Protoc. Cell Biol. 13, 1–27 (2002).
12. J. S. Henriksen, E. Lauridsen, and H. Gjørup, “A pilot study comparing optical coherence tomography, radiography, clinical photography, and polarisation microscopy for studies of hypomineralisation disturbances in enamel,” Heliyon 9(2), e13688 (2023). [CrossRef]
13. M. Wolman, “Polarized light microscopy as a tool of diagnostic pathology,” J. Histochem. & Cytochem. 23(1), 21–50 (1975). [CrossRef]
14. E. Applebaum, “Incipient dental caries,” J. Dent. Res. 12(4), 619–627 (1932). [CrossRef]
15. A. Darling, “Studies of the early lesion of enamel caries with transmitted light, polarized light and radiography,” Br. Dent. J. 101, 289–297 (1956).
16. A. Darling, “Studies of the early lesion of enamel caries: Its nature, mode of spread, and points of entry,” Br. Dent. J. 105, 119–135 (1958).
17. R. de Medeiros, J. Soares, and F. de Sousa, “Natural enamel caries in polarized light microscopy: differences in histopathological features derived from a qualitative versus a quantitative approach to interpret enamel birefringence,” J. Microsc. 246(2), 177–189 (2012). [CrossRef]
18. N. N. Soni and F. Brudevold, “Microradiographic and polarized light studies of initial carious lesions,” J. Dent. Res. 38(6), 1187–1194 (1959). [CrossRef]
19. C. de Mattos Brito, K. Meira, and F. de Sousa, “Natural enamel caries in quinoline: Volumetric data and the pattern of infiltration,” Microsc. Res. Tech. 81(2), 181–190 (2018). [CrossRef]
20. F. E. R. Baldasso, L. R. Cardoso, and V. D. da Silva, “Evaluation of the effect of four final irrigation protocols on root canal dentin components by polarized light microscopy and scanning electron microscopy,” Microsc. Res. Tech. 80(12), 1337–1343 (2017). [CrossRef]
21. S. Alali and I. A. Vitkin, “Polarized light imaging in biomedicine: emerging Mueller matrix methodologies for bulk tissue assessment,” J. Biomed. Opt. 20(6), 061104 (2015). [CrossRef]
22. M. D. Singh and I. A. Vitkin, “Discriminating turbid media by scatterer size and scattering coefficient using backscattered linearly and circularly polarized light,” Biomed. Opt. Express 12(11), 6831–6843 (2021). [CrossRef]
23. N. Ghosh, P. Gupta, and H. Patel, “Depolarization of light in tissue phantoms – effect of collection geometry,” Opt. Commun. 222(1-6), 93–100 (2003). [CrossRef]
24. C. L. Darling and D. Fried, “Polarized light propagation through sound and carious enamel at 1310-nm,” in Lasers in Dentistry XII, vol. 6137 (SPIE, 2006), pp. 151–158.
25. C. L. Darling, J. J. Jiao, C. Lee, et al., “Near-IR polarization imaging of sound and carious dental enamel,” Proc. SPIE 7549, 75490L (2010). [CrossRef]
26. M. D. Singh, L. Lilge, and A. Vitkin, “Spatial analysis of polarimetric images to enhance near-surface sampling sensitivity: feasibility in demineralized teeth and other tissue-like media,” J. Biomed. Opt. 28(10), 102906 (2023). [CrossRef]
27. L. Hoffmann, M. Feraric, and E. Hoster, “Investigations of the optical properties of enamel and dentin for early caries detection,” Clin. Oral Invest. 25(3), 1281–1289 (2021). [CrossRef]
28. I. Hariri, A. Sadr, and Y. Shimada, “Effects of structural orientation of enamel and dentine on light attenuation and local refractive index: An optical coherence tomography study,” J. Dent. 40(5), 387–396 (2012). [CrossRef]
29. M.-T. Tsai, Y.-L. Wang, and T.-W. Yeh, “Early detection of enamel demineralization by optical coherence tomography,” Sci. Rep. 9(1), 17154 (2019). [CrossRef]
30. B. Baumann, “Polarization sensitive optical coherence tomography: A review of technology and applications,” Appl. Sci. 7(5), 474 (2017). [CrossRef]
31. A. Baumgartner, S. Dichtl, and C. K. Hitzenberger, “Polarization–sensitive optical coherence tomography of dental structures,” Caries Res. 34(1), 59–69 (2000). [CrossRef]
32. J. Golde, F. Tetschke, and J. Walther, “Detection of carious lesions utilizing depolarization imaging by polarization sensitive optical coherence tomography,” J. Biomed. Opt. 23(07), 1 (2018). [CrossRef]
33. J. Golde, F. Tetschke, and J. Walther, “Cross-sectional and en-face depolarization imaging for the assessment of dental lesions,” Curr. Dir. Biomed. Eng. 4(1), 301–304 (2018). [CrossRef]
34. J. Golde, F. Tetschke, R. Vosahlo, et al., “Towards quantitative demineralization imaging for the assessment of carious lesions based on PS-OCT,” EPJ Web Conf. 238(2020), 04009 (2020). [CrossRef]
35. J. Golde, F. Tetschke, R. Vosahlo, et al., “Assessment of occlusal enamel alterations utilizing depolarization imaging based on PS-OCT,” in Optical Coherence Imaging Techniques and Imaging in Scattering Media III, vol. 11078M. Wojtkowski, S. A. Boppart, W.-Y. Oh, eds., International Society for Optics and Photonics (SPIE, 2019).
36. B. Dikmen, “ICDAS II criteria (international caries detection and assessment system),” J. Istanbul Univ. Fac. Dent. 49(3), 63–72 (2015). [CrossRef]
37. F. Litzenburger, A. Lederer, and M. Kollmuß, “Near-infrared transillumination with high dynamic range imaging for occlusal caries detection in vitro,” Lasers Med. Sci. 35(9), 2049–2058 (2020). [CrossRef]
38. N. B. Pitts, D. T. Zero, P. D. Marsh, et al., “Dental caries,” Nature Reviews 3, 17030 (2017).
39. H. Meyer-Lueckel and S. Paris, “When and how to intervene in the caries process,” Oper. Dent. 41(S7), S35–S47 (2016). [CrossRef]
40. A. V. Zavgorodniy, R. Rohanizadeh, and M. V. Swain, “Ultrastructure of dentine carious lesions,” Arch. Oral Biol. 53(2), 124–132 (2008). [CrossRef]
41. C. Shimizu, Y. Yamashita, T. Ichijo, et al., “Clinical science: Carious change of dentin observed on longspan ultrathin sections,” J. Dent. Res. 60(11), 1826–1831 (1981). [CrossRef]
42. G. Conrads and I. About, “Pathophysiology of dental caries,” in Caries Excavation: Evolution of Treating Cavitated Carious Lesions, vol. 27F. Schwendicke, J. Frencken, and N. Innes, eds. (Karger, 2018), chap. Pathophysiology of Dental Caries, pp. 1–10, 1st ed.
43. G. der Bundesärztekammer, “Die (Weiter-)Verwendung von menschlichen Körpermaterialien für Zwecke medizinischer Forschung,” Deutsches Ärzteblatt 100, A1632–A1635 (2003).
44. L. Findeisen, J. Bolte, and C. Vater, “Cell spheroids are as effective as single cells suspensions in the treatment of critical-sized bone defects,” BMC Musculoskeletal Disord. 22(1), 401 (2021). [CrossRef]
45. K. Donath, “Die Trenn-Dünnschliff-Technik zur Herstellung histologischer Präparate von nicht schneidbaren Geweben und Materialien,” Der Praeparator 34, 197–206 (1988).
46. J. Golde, F. Tetschke, R. Vosahlo, et al., “Depolarization imaging for caries detection validated by co-registered PS-OCT and µCT volumes (Conference Presentation),” in Lasers in Dentistry XXVI, vol. 11217P. RechmannD. Fried, eds., International Society for Optics and Photonics (SPIE, 2020), p. 1121707.
47. J. Grundmann and J. Golde, “X-ray micro-computed tomography (µCT) volumes of carious teeth in vitro,” OpARA, 2023, https://dx.doi.org/10.25532/OPARA-225.
48. J. Grundmann and J. Golde, “Polarized light microscopy (PLM) images of occlusal caries lesions,” OpARA, 2023, https://dx.doi.org/10.25532/OPARA-224.
49. J. Grundmann and J. Golde, “Depolarized light microscopy (DLM) images of occlusal caries lesions,” OpARA, 2023, https://dx.doi.org/10.25532/OPARA-226.
50. J. Grundmann and J. Golde, “Photographs of the occlusal surface and bitewing X-ray images of carious teeth in vitro,” OpARA, 2023, https://dx.doi.org/10.25532/OPARA-223.
51. D. H. Goldstein, Polarized Light (CRC Press, 2017).
52. P. Tang, N. Le, and J. Lu, “Local axis orientation mapped by polarization sensitive optical coherence tomography provides a unique contrast to identify caries lesions in enamel,” Biomed. Opt. Express 13(8), 4247–4260 (2022). [CrossRef]
53. A. Braun, L. M. J. C. Guiraud, and R. Frankenberger, “Histological validation of ICDAS II and radiological assessment of occlusal carious lesions in permanent teeth,” Odontology 105(1), 46–53 (2017). [CrossRef]
