Open Access
Add to BibSonomy
Abstract
In this work, a backscattering imaging method based on near infrared random fiber laser is shown to provide a high contrast optical image between carious and sound enamel. The obtained contrast is 0.70, which is more than 8 times higher than the contrast obtained from radiographic imaging. Caries and cracks in enamel could clearly be identified against healthy enamel using the optical system. The near infrared wavelength, high spectral density and low coherence of random fiber laser contribute to its deep penetration, high brightness and low speckle contrast, using the method in a backscattering configuration opens potential clinical use.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Corrections
15 April 2020: A typographical correction was made to the funding section.
1. Introduction
Examination of oral cavity diseases rely on visual, tactile and technology-based methods. Besides visual and tactile, the most commonly employed imaging examination method in dental practice for the detection and evaluation of caries lesions is radiography [1], which has evolved from analogic to digital, and can be used for a single tooth [2] or for panoramically view [3]. Three-dimensional radiographic imaging has also been exploited [4]. However, the size or depth of such lesions can be underestimated or even overlooked when the caries is at its early stage or in proximal regions, and can be dependent on dental staff [5].
Recent investigations have been made to find more accurate optically based diagnostic approaches for early dental caries detection, which includes auto-fluorescence [6,7], transillumination [8–10], and reflection [11,12]. Compared with radiography, imaging based on near infrared (NIR) light source has much higher sensitive, uses non-ionizing radiation, handling and operation is simple, which makes it suitable for tooth caries detection. For example, the dental enamel has high transmittance and the dentin has high reflectivity in the NIR region. The mineral loss caused by dental caries leads to an increase in scattering coefficient, thus leading to higher photon scattering (in all directions, including back reflected) than sound enamel and behave differently when imaged (i.e., high imaging contrast). The contrast between sound and demineralized enamel regions was investigated at different wavelengths, and the results showed that wavelengths beyond 1400nm have better performance for caries detection since the photons in sound enamel are absorbed by water [13,14]. Thus, a NIR light source with low cost, high brightness, deep penetration and low spatial coherence is highly desirable for caries detection systems at those wavelengths, and the so-called random laser is very suitable for this purpose, outperforming LEDs, narrow line lasers and radiography. In brief, random lasers (RLs), are optical sources whose coherent emission arise due to multiple scattering feedbacks within a disordered gain medium, instead of having a cavity formed by two mirrors. They were first unambiguously demonstrated in 1994 [15], and their development, characteristics and several applications have been reviewed in [16]. One of their most interesting characteristics of RL is that speckle patterns caused by independent optical modes are superimposed and averaged out to provide a speckle-free radiation which is favorable for imaging [17,18]. Besides, RLs have the property of high brightness and high spectral density, which are useful to improve the performance of speckle-free imaging [19]. Carvalho et al [20] used a RL as the light source in an epi-illumination configuration and demonstrated that it has potential for biological imaging applications due to its image quality and spectral density.
A more compact and flexible light source can be directly generated in an optical fiber, and a fiber-based amplified spontaneous emission (ASE) sources is an example [21]. Random fiber lasers (RFL) have also been demonstrated in 2007 [22], and a review on this subject can be found in [23]. Among the advantages already mentioned for bulk RL, RFL have the additional advantage of directional emission, and can be made very compact. Random fiber laser based on nanoparticles [24] or disordered microstructure polymer optical fiber [25] has been investigated, which promotes the development of coherent random fiber laser regime. It has been recently demonstrated, that RFL has better imaging capability than ASE for deeper penetration due to its high spectral density [26–28].
In this manuscript, we demonstrate a tooth imaging system based on the backscattered radiation from a NIR-RFL, which is based on the RFL developed in Ref. 26. In there, the proof of concept for using the RFL was introduced. Here, we perform bio-imaging application in extracted human teeth. Our hypothesis is that, since the scattering coefficients of the carious tissue compared to healthy tissue in enamel and dentin are different, the number of photons collected by the detector changes, which can be translated into an image and the basic component of teeth can be analyzed. Our experimental results show that using the RFL as a source has much better imaging quality due to its speckle-free property than narrow line lasers (NLL) and has higher imaging contrast of caries and enamel than ASE or LED. Our results are also compared with radiography and optical microscope imaging, and it indicates that the RFL imaging can better detect caries thanks to its wavelength-related merits.
2. Experimental Setup
The experimental setup consists of light source and imaging part, as shown in Fig. 1. The RFL is generated through a half-opened structure composed of a fiber loop mirror (i.e., formed from a 3dB couple) and 25 km-length single mode fiber (SMF). A Raman pump with central wavelength of 1455 nm is injected into the SMF through a 1455/1550 nm wavelength division multiplexer (WDM 1). Another WDM, WDM2, is used at the end of the SMF in order to split out the residual pump light and make sure only the generated RFL injects into the imaging part. An isolator (ISO) is used to eliminate light reflection, making sure that the laser is generated only through randomly distributed Rayleigh scattering. To adjust the output light in appropriate power for imaging, a variable optical attenuator (VOA) is used. For comparison, an ASE light source based on Erbium doped fiber and a 1550 nm narrow line width laser (NLL) are also used separately to replace the RFL for imaging
Fig. 1. Schematic of experimental setup. WDM, wavelength division multiplexer. SMF, single mode fiber. ISO, isolator. VOA, variable optical attenuator. MMF, multimode fiber. BS, beam splitter
The imaging part employs a backscattering configuration. A 30 m extra-large mode area step-index MMF (core and cladding diameter are 105 and 125 µm respectively, NA is 0.22, YOFC) is spliced after the VOA and before the imaging system to reduce the spatial coherence of light source(s). The illumination light from the MMF is collimated by lens1 (focal length of 6.2 mm) and directed by a beam splitter to the sample. The reflected light from the sample is collimated by lens2 (focal length of 120 mm) and directed by the beam splitter to the CCD (Xenics, Bobcat-640-GigE).
Ethical approval was obtained from the ethical committee of the Federal University of Pernambuco, Pernambuco, Brazil. Newly extracted human permanent molar teeth were cleaned and stored in deionized water. Each tooth was sectioned mesiodistally in parallel with the long axis of the crown using a double-sided diamond disco (Buehler Ltda., Lake Bluff, IL, USA) coupled to a precision cutter (Isomet 1000 Speed Saw, Buehler Ltda., Lake Bluff, IL, USA). Samples with a mean thickness of d ≈ 1.00 ± 0.1mm were obtained. Afterwards, both sides of each section were examined under a stereomicroscope at 20x magnification (Stemi 2000; Carl Zeiss, Jena, Germany). Teeth were identified with a sound or carious lesion area in the enamel and/or dentine. Examination of the 5 slices of teeth surfaces revealed a distribution carious and healthy areas of the studied sample teeth. Figure 2 shows a representative tooth slice and the cutting procedure.
3. Results and discussion
We initially show the optical spectra of the three light sources used in this experiment (RFL, ASE, and NLL) are shown in Fig. 3(a). The full width at half maxima (FWHM) of the RFL is ∼1.5 nm, which is obtained at the pump power of 33.4 dBm. The FWHM of ASE is 5 times broader than the RFL, which is about 7nm and the NLL has the narrowest FWHM of less than 0.01 nm. The coherence length, $\Delta c$, is calculated as $\Delta c = {\lambda ^2}/\Delta \lambda$, where $\lambda$ and $\Delta \lambda$ are the center wavelength and spectral bandwidth respectively. RFL, thus, has much shorter coherent length than NLL. Our previous work has demonstrated that when light is injected into the MMF, the spatial coherence decreases with excitation of high-order transverse modes, and speckle contrast reduces significantly, after the decoherence of a 30m extra-large mode area step index MMF, the RFL, and ASE can reach a speckle contrast of ∼0.049 and 0.039, which is near the threshold of human perception and is lower enough for speckle-free imaging [26–28]. The power reaching the teeth specimens from each of the sources was controlled to be at most 4dBm. Care was taken to avoid saturation of the CCD detector. The intensity fluctuation of the RFL that we used is very low for frequency range smaller than GHz, and the CCD response is less than MHz Thus, the average intensity is quite stable, and the speckle contrast value, in our case, is mainly determined by the number of transverse modes. The backscattering imaging result of RFL, ASE, and NLL are shown in Figs. 3(b)–3(d), respectively. The enamel has high transmittance, most of the light penetrates the enamel and only a small percentage is scattered back to the detector. Therefore, the enamel part appears as dark region, while the dentin has high reflection and appears as a bright region. Moreover, the mineral loss of the carious region would cause more than two orders of magnitude increase in scattering coefficient and thus also appear as bright region in backscattering image [28]. This difference in scattering coefficient makes enamel and caries high contrast and easy to be identified. For RFL and ASE, the enamel, dentin and demineralized enamel regions of tooth specimen can be clearly identified, while for NLL, strong speckle patterns occur and blur the image. It is because RFL and ASE have relatively broad spectrum and short coherence length, which lead to low speckle contrast after decoherence of the MMF. For NLL, its long coherence length causes strong modal interference and obvious speckle patterns that makes the image too obscure to distinguish between the demineralized from healthy areas.
Fig. 3. Imaging from different sources. (a) spectrum of RFL, ASE, and NLL (b)-(d)imaging result of RFL, ASE and NLL, respectively.
Images of the sample for different powers from the RFL and the ASE are shown in Fig. 4 for comparison. It can be seen that the RFL images are a bit brighter than ASE images at the same radiation power, because the former has higher spectral density and lower absorption. For a quantitative comparison, the intensity integration of each figure of Figs. 4(a)–4(f) is calculated, and is given in Fig. 4(g), as a function of the output power. It is observed that, the RFL image is brighter than the ASE image for all the output power, and has comparable speckle contrast to guarantee image quality. Although the bandwidth of RFL is narrower than ASE, its coherence after the MMF is low enough for speckle-free imaging, and the high spectral density and low absorption make the RFL image brighter. Furthermore, it has been shown that even though RL or RFL have narrow bandwidth, they are highly multimode and speckle free [18].
Fig. 4. Imaging at different pump power. (a)-(f) image of different power for RFL and ASE, (g) the integral intensity of RFL and ASE.
As mentioned before, the cavities and fissures could open the way to mineral loss, which lead to two orders of magnitude increase in scattering coefficient at NIR region. Caries and cracks thus appear as bright regions on the image because more light is scattered back to the CCD. Figures 5(a)–5(d) compares the results from RFL and ASE illumination at power of 4 dBm. Figures 5(a) and 5(b) shows the images formed, while Fig. 5(c) gives intensity fluctuation among enamel and crack (along the red dashed curve), and Fig. 5(d) depicts the intensity fluctuation along the red solid curve. The peak in Fig. 5(c) represents the crack region, and other part of the curve represents the enamel region. It can be seen that the RFL image has higher contrast between the crack and health region of the enamel, than the ASE image, this is in agreement with the conclusion of Fig. 4. We also calculate the contrast ratio between crack and sound enamel for RFL and ASE imaging. The contrast I defined as in Eq. (1)
where ${I_{crack}}$ and ${I_{enamel}}$ are the average intensities of the crack and the sound enamel. The contrast is 0.32 and 0.25 for RFL and ASE respectively, indicating that the RFL image can better identify the crack region. Figure 4(d) compares the intensity between enamel and dentin. Note that the two light sources, RFL and ASE, have enough low speckle contrast to fulfill speck-free imaging, thus the fluctuation in intensity arises from different concentrations of dentin. The standard deviation of intensity is calculated to be 0.18 and 0.14 for RFL and ASE, respectively, which means RFL has a better performance for dentin concentration detection.Fig. 5. Imaging comparison between RFL and ASE. (a) RFL image, (b) ASE image, (c) Intensity along the red dotted curve, (d) Intensity along red solid curve.
Furthermore, we compare the RFL imaging with other two different imaging methods, i.e., radiography and microscope. In Fig. 6(b), the radiography could only identify enamel and dentin, while not capable to detect all details in enamel and dentin. For example, the crack observed in Fig. 6(a) is not detected by radiography. In Fig. 6(c), the optical microscope can show crack in enamel but the dentin region appears uniform. By contrast, in Fig. 6(a), the RFL based backscattering image method could further distinguish the mineral concentration showing different brightness in the dentin region thanks to the penetration ability of its NIR wavelength, as well as its high spectral density.
Fig. 6. Imaging comparison between RFL and ASE. (a) RFL image, (b) ASE image, (c) Intensity along the red dotted curve, (d) Intensity along red solid curve.
For quantitative comparison, Fig. 7 depicts the intensity profile of the backscattered photons along the red curve marked in each images of Fig. 6. Note that the RFL (blue curve) and microscope (red curve) detect the backscattered light while radiography (yellow curve) detect the transmitted light, and that leads to different patterns for intensities as seen in Fig. 7.
From Fig. 7, we first calculated the contrast between enamel and dentin from Eq. (2), where ${I_{dentin}}$ and ${I_{enamel}}$ are average intensities of the dentin and enamel, respectively. The values of ${C_{de}}$ are given in Table 1.
Table 1. The contrast of RFL, radiography, and microscope
The RFL has the highest value of ${C_{de}}$, which verifies that the RFL is the best light source to distinguish enamel and dentin.
Secondly, we also consider the contrast between the caries and the sound enamel. The caries regions appear as the brighter region in the reflective image [Figs. 5(a) and 5(c)] and as a darker region in the transmitted image. The quantitative value of contrast is defined as in Eq. (3), which is 0.70, 0.08 and 0.18 for the RFL image, radiography and microscope respectively, as given in Table 1. It is obvious that the RFL imaging has the best performance in distinguishing enamel from both dentin and caries.
4. Conclusion
In conclusion, backscattering tooth imaging method based on NIR-RFL as the optical source has been investigated. The experimental results indicate that, owing to its low coherence, high spectral density and speckle-free nature, the RFL (and random lasers) have the best comprehensive performance as an optical source in finding dental details like demineralized and crack areas, separating enamel from dentin as well as identifying mineral concentration, compared to other NIR light sources such as NLL and ASE. It also outperforms other imaging measuring methods like radiography and optical microscopy, with a higher imaging contrast, as shown in Table 1. Therefore, the RFL is an ideal light source for dental caries lesions diagnosis. We anticipate that technologically it is feasible to design a compact RFL or a semiconductor random laser based backscattering imaging technique, which can potentially be used in clinical environment.
Funding
National Natural Science Foundation of China (11974071, 61635005, 61811530062); Ministério da Ciência, Tecnologia e Inovação (400011/2016-6); Conselho Nacional de Desenvolvimento Científico e Tecnológico; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior; Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco; Instituto Nacional de Ciência e Tecnologia de Fotônica.
Disclosures
The authors declare no conflicts of interest.
References
1. J. Gomez, “Detection and diagnosis of the early caries lesion,” BMC Oral Health 15(S1), S3 (2015). [CrossRef]
2. N. Abogazalah, J. G. Eckert, and M. Ando, “In vitro performance of near infrared light transillumination at 780-nm and digital radiography for detection of non cavitated approximal caries,” J. Dent. 63, 44–50 (2017). [CrossRef]
3. S. Perschbacher, “Interpretation of panoramic radiographs,” Aust. Dent. J. 57(Suppl 1), 40–45 (2012). [CrossRef]
4. J. K. Mitchell, A. R. Furness, R. J. Sword, S. W. Looney, W. W. Brackett, and M. G. Brackett, “Diagnosis of Pit-and-fissure Caries Using Three-dimensional Scanned Images,” Oper. Dent. 43(3), E152–E157 (2018). [CrossRef]
5. M. A. Geibel, S. Carstens, U. Braisch, A. Rahman, M. Herz, and A. Jablonski-Momeni, “Radiographic diagnosis of proximal caries-influence of experience and gender of the dental staff,” Clin. Oral. Invest. 21(9), 2761–2770 (2017). [CrossRef]
6. S. A. Son, K. H. Jung, C. C. Ko, and Y. H. Kwon, “Spectral characteristics of caries-related autofluorescence spectra and their use for diagnosis of caries stage,” J. Biomed. Opt. 21(1), 015001 (2016). [CrossRef]
7. D. M. Zezell, A. d. C. Ribeiro, L. Bachmann, A. S. L. Gomes, C. Rousseau, and J. Girkin, “Characterization of natural carious lesions by fluorescence spectroscopy at 405 nm excitation wavelength,” J. Biomed. Opt. 12(6), 064013 (2007). [CrossRef]
8. S. Chung, D. Fried, M. Staninec, and C. L. Darling, “Multispectral near-IR reflectance and transillumination imaging of teeth,” Biomed. Opt. Express 2(10), 2804–2814 (2011). [CrossRef]
9. L. Karlsson, A. M. A. Maia, B. B. C. Kyotoku, S. Tranæus, A. S. L. Gomes, and W. Margulis, “Near-infrared transillumination of teeth: measurement of a system performance,” J. Biomed. Opt. 15(3), 036001 (2010). [CrossRef]
10. A. M. A. Maia, L. Karlsson, W. Margulis, and A. S. L. Gomes, “Evaluation of two imaging techniques: near-infrared transillumination and dental radiographs for the detection of early proximal enamel caries,” Dentomaxillofac Rad. 40(7), 429–433 (2011). [CrossRef]
11. W. A. Fried, D. Fried, K. H. Chan, and C. L. Darling, “High contrast reflectance imaging of simulated lesions on tooth occlusal surfaces at near-IR wavelengths,” Lasers Surg. Med. 45(8), 533–541 (2013). [CrossRef]
12. J. C. Simon, S. A. Lucas, M. Staninec, H. Tom, K. H. Chan, C. L. Darling, M. J. Cozin, R. C. Lee, and D. Fried, “Near-IR transillumination and reflectance imaging at 1300-nm and 1500-1700-nm for in vivo caries detection,” Lasers Surg. Med. 48(9), 828–836 (2016). [CrossRef]
13. K. H. Chan and D. Fried, “Multispectral cross-polarization reflectance measurements suggest high contrast of demineralization on tooth surfaces at wavelengths beyond 1300 nm due to reduced light scattering in sound enamel,” J. Biomed. Opt. 23(6), 060501 (2018). [CrossRef]
14. V. B. Yang, D. A. Curtis, and D. Fried, “Cross-polarization reflectance imaging of root caries and dental calculus on extracted teeth at wavelengths from 400 to 2350 nm,” J. Biomed. Opt. 11(11), e201800113 (2018). [CrossRef]
15. N. M. Lawandy, R. M. Balachandran, A. S. L. Gomes, and E. Sauvain, “Laser action in strongly scattering media,” Nature 369(6478), 340 (1994). [CrossRef]
16. F. Luan, B. Gu, A. S. L. Gomes, K. T. Yong, S. Wen, and P. Prasad, “Lasing in Nanocomposite Random Media, Review Paper,” Nano Today 10(2), 168–192 (2015). [CrossRef]
17. B. Redding, M. A. Choma, and H. Cao, “Speckle-free laser imaging using random laser illumination,” Nat. Photonics 6(6), 355–359 (2012). [CrossRef]
18. B. H. Hokr, M. S. Schmidt, J. N. Bixler, P. N. Dyer, G. D. Noojin, B. Redding, R. J. Thomas, B. A. Rockwell, H. Cao, V. V. Yakovlev, and M. O. Scully, “A narrow-band speckle-free light source via random Raman lasing,” J. Mod. Opt. 63(1), 46–49 (2016). [CrossRef]
19. B. H. Hokr, D. T. Nodurft, J. V. Thompson, J. N. Bixler, G. D. Noojin, B. Redding, R. J. Thomas, H. Cao, B. A. Rockwell, M. F. Scully, and V. V. Yakovlev, “Lighting up microscopy with random Raman lasing,” in Real-time Measurements, Rogue Events, and Emerging Applications, (2016), pp.9732.
20. M. T. Carvalho, A. S. Lotay, F. M. Kenny, J. M. Girkin, and A. S. L. Gomes, “Random laser illumination: an ideal source for biomedical polarization imaging?” in Multimodal Biomedical Imaging XI, (2016), pp. 9701.
21. B. Redding, P. Ahmadi, V. Mokan, M. Seifert, M. A. Choma, and H. Cao, “Low-spatial-coherence high-radiance broadband fiber source for speckle free imaging,” Opt. Lett. 40(20), 4607–4610 (2015). [CrossRef]
22. C. J. S. de Matos, L. d. S. Menezes, A. M. Brito-Silva, M. A. M. Gámez, A. S. L. Gomes, and C. B. de Araújo, “Random Fiber Laser,” Phys. Rev. Lett. 99(15), 153903 (2007). [CrossRef]
23. D. V. Churkin, S. Sugavanam, I. D. Vatnik, Z. Wang, E. V. Podivilov, S. A. Babin, Y. Rao, and S. K. Turitsyn, “Recent advances in fundamentals and of random fiber lasers,” Adv. Opt. Photonics 7(3), 516–569 (2015). [CrossRef]
24. Z. J. Hu, Q. Zhang, B. Miao, Q. Fu, G. Zou, Y. Chen, Y. Luo, D. Zhang, P. Wang, H. Ming, and Q. Zhang, “Coherent Random Fiber Laser Based on Nanoparticles Scatteringin the Extremely Weakly Scattering Regime,” Phys. Rev. Lett. 109(25), 253901 (2012). [CrossRef]
25. Z. J. Hu, B. Miao, T. X. Wang, Q. Fu, D. G. Zhang, H. Ming, and Q. J. Zhang, “Disordered microstructure polymer optical fiber for stabilized coherent random fiber laser,” Opt. Lett. 38(22), 4644–4647 (2013). [CrossRef]
26. R. Ma, Y. J. Rao, W. L. Zhang, and B. Hu, “Multimode Random Fiber Laser for Speckle-Free Imaging,” IEEE J. Sel. Top. Quantum Electron. 25(1), 0900106 (2019). [CrossRef]
27. R. Ma, W. L. Zhang, J. Y. Guo, and Y. J. Rao, “Decoherence of fiber supercontinuum light source for speckle-free imaging,” Opt. Express 26(20), 26758–26765 (2018). [CrossRef]
28. R. Ma, J. Q. Li, J. Y. Guo, H. Wu, H. H. Zhang, B. Hu, Y. J. Rao, and W. L. Zhang, “High-power low spatial coherence random fiber laser,” Opt. Express 27(6), 8738–8744 (2019). [CrossRef]
