Time-resolved fluorescence spectroscopy of white-spot caries in human enamel Fernanda Ferretti de Oliveira, Amando Siuiti Ito, and Luciano Bachmann* Departamento de Física e Matemática, Faculdade de Filosofia Ciências e Letras de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto, Brazil *Corresponding author:
[email protected] Received 20 November 2009; accepted 4 March 2010; posted 23 March 2010 (Doc. ID 120264); published 13 April 2010
The objective is to differentiate noncavitated caries enamel through time-resolved fluorescence and to find excitation and emission parameters that can be applied in future clinical practice for detection of caries lesions that are not clearly visible to the professional. Sixteen human teeth with noncavitiated white-spot caries were selected for this work. Fluorescence intensity decay was measured by using an apparatus based on the time-correlated single-photon counting method. An optical fiber bundle was employed for sample excitation (440 nm), and the fluorescence collected by the same bundle (500 nm) was registered. The average lifetime for sound enamel was 7:93 0:09, 2:46 0:04, and 0:51 0:02 ns, whereas for the carious enamel the lifetimes were 4:84 0:06, 1:35 0:02, and 0:16 0:01 ns. It was concluded that it is possible to differentiate between carious and sound regions by time-resolved fluorescence and that, although the origin of enamel fluorescence is still uncertain, the lifetime values seem to be typical of fluorophores like collagen I. © 2010 Optical Society of America OCIS codes: 140.0140, 300.2530.
1. Introduction
For years, researchers have been improving the current conventional methods of caries detection, such as clinical inspection or radiographic methods that are commonly used in clinical practice [1,2]. The main purpose in caries inspection is to identify the caries in the earliest stages with good accuracy. The development of new diagnostic techniques is a current theme, and different procedures have been developed and shown to be efficient, although they are not widely applied in the clinical context. Examples of these procedures are fiber optic transillumination, laser profilometry, quantitative lightinduced fluorescence, microradiography, optical coherence tomography, x-ray microtomography, electrical caries detection, dye techniques, and fluorescence spectroscopy. A more extended review of new technologies can be found in the literature [1–5].
0003-6935/10/122244-06$15.00/0 © 2010 Optical Society of America 2244
APPLIED OPTICS / Vol. 49, No. 12 / 20 April 2010
The development of new techniques, such as fluorescence spectroscopy, has opened the possibility of detecting subsuperficial lesions [6]. This technique, known as static fluorescence spectroscopy, detects the emitted light over a wavelength range (fluorescence spectrum) when the tissue is excited at a specific wavelength that promotes a molecular excitation. The fluorescence spectroscopy technique used in this work measures the time dependence of the emission intensity instead of the fluorescence spectrum. Lifetimes are determined from the intensity decay, and they represent the average times during which the fluorophore remains in the first excited electronic state prior to its return to the ground state. An examination of the lifetimes permits identification of different chemical compounds and allows for possible sample differentiation such as caries lesions or soft tissue lesions, since these tissues have different chemical compositions because of lesion installation and, consequently, will also furnish different lifetime values. The carious enamel displays different fluorescence characteristics from those of the sound enamel.
Specifically, carious regions exhibit higher fluorescence when excited between 400 and 420 nm and measured at 520 nm, mainly for white-spot lesions [7]. Other different excitation wavelengths produce distinct fluorescence emission peaks. A recent review about the fluorescence of biological tissues can be found in [8]. Fluorescence has been widely employed in dental research, mainly for our understanding of caries [9]. Compared with sound enamel, carious teeth have been reported to exhibit shorter fluorescence lifetimes when excited at 407 nm [10]. The origin of fluorescence in the sound and carious enamel has not been clearly defined in the literature. In the case of the carious enamel, some evidence has been assigned to porphyrins [10], while sound enamel fluorescence is probably dominated by the collagen and tryptophan amino acid present in the organic matrix. Metal-free fluorescent porphyrin monomers have a lifetime of about 10–20 ns, while most of the endogenous fluorophores like the free or bound amino acid tryptophan exhibit lifetimes of less than 6 ns [11]. In turn, collagen I has an exponential lifetime of the order of 4 ns [12], but other lifetimes have been observed for collagen molecules [13]. The time-dependent fluorescence decay resulting from fluorescence excitation by laser P pulses obeys the following expression: IðtÞ ¼ i αi exp½−t=τi , where the fluorescence intensity I of one fluorophore decreases exponentially with typical lifetime values of picoseconds or nanoseconds. As long as the fluorophore remains in the excited state, its fluorescence lifetime is longer. This lifetime does not depend on the fluorescence intensity, which seems to be advantageous for differentiation between carious and sound enamel. This is because the caries volume, enamel mineralization, and scattering of light that
interfere in the fluorescence intensity do not interfere in its lifetime. 2.
Methodology
A.
Sample Selection
Sixteen human teeth with noncavitated white-spot caries were selected for this work. The teeth were obtained after clinical indication for extraction and had been cleaned and stored in 0:9 wt: % sodium chloride solution until the fluorescence experiments were accomplished. The tooth areas under investigation did not receive any kind of prophylactic procedure in order to avoid modification of the original material constitution. The measurements were performed on dry teeth so that it would be possible to have better visualization of the lesion. The present work was approved by a local research ethics committee. B.
Fluorescence Spectrometer
Fluorescence intensity decay was measured by using an apparatus based on the time-correlated singlephoton counting method; the experimental diagram can be seen in Fig. 1. The excitation source was a Tsunami 3950 Spectra Physics titanium–sapphire laser that emitted pulses of 6 ps and allowed changes of wavelengths tunable between 840 and 1000 nm, pumped by the solid-state laser Millenia X Spectra Physics with emission of 530 nm. The repetition rate of the pulses was set to 4:0 MHz by using a 3980 Spectra Physics pulse picker that allowed operation in variable frequencies between 80 MHz and 8 kHz. The laser was tuned so that a second- and thirdharmonic generator (GWN-23PL Spectra Physics) give an emergent beam tuned in the ranges 420–500 nm and 280–333 nm. Samples were excited at 440 nm. The output pulses were coupled to the
Fig. 1. Experimental setup employed for the fluorescence lifetime experiment. 20 April 2010 / Vol. 49, No. 12 / APPLIED OPTICS
2245
fiber of a fluorescence probe (R400-7-UV-Vis, Ocean Optics); the probe tip was flat and could touch the selected sample for excitation. At the same fiber, a bundle of six fibers collected the fluorescence and directed the signal to an Edinburgh FL900 spectrometer. The fluorescence intensity decay was registered at 500 nm. A photodiode detected the excitation pulse and sent the start signal to a time–amplitude converter (TAC). Emitted photons were detected by a refrigerated Hamamatsu R3809U microchannel plate photomultiplier, and the stop signal was sent to the TAC. An electrical signal corresponding to the time between start and stop events was sent to an analog– digital converter and then transferred to a multichannel analyzer. The repetition of this procedure for a large number of excitation pulses on the sample resulted in a histogram in the multichannel analyzer, which represents the number of counts in each channel, which in turn corresponds to the sample fluorescence decay profile. The FWHM of the instrument response function was typically 100 ps, and the time resolution was 12 ps=channel. C. Sample Measurement
For fluorescence acquisition, the aimed region was touched with the tip of the probe. The fluorescence was measured until a predetermined signal count between 6000 and 10,000 was reached. The measurement was conducted in a dark room without adverse light sources. Each selected tooth with white spots was measured twice. The carious region was measured first, followed by measurement of the opposite side of the tooth free of caries or other anomalies chosen to register the fluorescence decay corresponding to the sound enamel. D. Data Evaluation
The fluorescence was evaluated with a software that extracts the lifetime values; a t-Student statistic was employed for comparison between sound and carious data for determination of the average value and dif-
ferences between the two groups. Software provided by Edinburgh Instruments was employed for analysis of the individual decays, which were fitted to multiexponential curves for the fluorescence intenP sity: IðtÞ ¼ i αi exp½−t=τi , where αi are the preexponential factors and τi are the lifetimes. 3.
Results
The results registered in this work refer to the use of a laboratory procedure, but they could be extrapolated to clinical measurement and used with portable equipment. The use of the above described procedure produces a typical fluorescence intensity decay as shown in Fig. 2. On the left, the fluorescence counts are presented on a linear scale. The time scale starts at 20 ns because a measurement delay was produced by the experimental system. Before this time, there was no fluorescence signal. There is a slight visual difference between the sound and the carious enamel. The same results are shown on a logarithmic scale on the right-hand side of Fig. 2. It is possible to observe that the fluorescence is not linear for a logarithmic scale, thereby indicating a multiexponential decay. After the experimental acquisition, the three-term exponential function was adjusted as shown in Fig. 3. With this adjustment, the lifetime values and the preexponential value were extracted for each sample. The preexponential values together with the exponential term were integrated to extract the relative percentage contribution of each term to the fluorescence. The lifetimes and their relative percentage value areas are listed in Table 1 for the sound and carious regions of all 16 samples. Table 1 also lists the three lifetimes with standard deviations provided by the software that makes the theoretical adjustment of both groups. The last row shows the average value for each lifetime, too. The average lifetimes can be observed in Fig. 4, which reveals a statistical difference (p < 0:05) between the sound and the carious regions for all three lifetimes.
Fig. 2. Typical experimental data for the fluorescence decay of one sound region and one carious region. Left, the fluorescence counts are shown on a linear scale. Right, the same results on a logarithmic scale. 2246
APPLIED OPTICS / Vol. 49, No. 12 / 20 April 2010
The methodology employed here and other methodologies that measure the lifetime fluorescence are not affected by changes in fluorescence intensity; only a change in the signal-to-noise ratio will occur if the fluorescence intensity changes, but this does not significantly affect the lifetime value. In contrast to this technique, the steady-state fluorescence procedure is affected by changes in the fluorescence intensity. The dynamic process of demineralization and remineralization determines mineral loss and changes in the pore structure, which in turn alters the fluorescence intensities of the fluorophores located inside the enamel structure. The optical properties of tissues are not the same at different sites of the tooth, and the fluorescence intensity accompanies all these optical and structural alterations [14]. Also, different carious lesions lead to different band intensities, [15] and this may be one of the reasons for the differences in the values of fluorescence band intensities of a steady-state measurement when compared with literature data. So it is obvious that the static fluorescence procedure is not as specific for detection of carious lesions as the timeresolved fluorescence presented in this study. In the literature, a steady-state fluorescence experiment with an excitation wavelength of 405 nm has demonstrated that the maximum emission peak is near 500 nm for both carious and sound enamel [16]. The present work uses an excitation wavelength of 440 nm and also selects an emission at the maximum peak (500 nm). The origin of the fluorescence signal in the enamel tissue is not clearly defined in the literature [8]. However, when an appropriate excitation wavelength is employed, the fluorescence lifetime should be determined mainly by endogenous porphyrins, collagen, and tryptophan. Metal-free fluorescent porphyrin monomers have a lifetime of about 10–20 ns,
Fig. 3. Typical adjustment to experimental data performed by using a three-term exponential function. Specifically for this sample, the following results were obtained: Counts ¼ 10þ 489 exp½−ðx − 20Þ=0:6 þ 1573 exp½−ðx − 20Þ=2:4 þ 708 exp½−ðx − 20Þ= 7:1. The three lifetimes achieved for all the samples can be observed in Table 1. The preexponential values together with each exponential term were integrated to extract their relative percentage contribution to fluorescence.
All the lifetimes of the sound enamel are higher than the respective values found for the carious enamel. 4. Discussion
The results show that it is possible to employ the lifetime fluorescence parameter to distinguish between carious and sound enamel. According to Fig. 4 and Table 1, the present methodology can differentiate noncavitated natural carious lesions in the human dental enamel at a significant level, p < 0:05. This differentiation can be provided by using any of the three lifetimes evaluated in this work. Table 1.
Lifetimes and Their Relative Percentage Values for 16 Sound and Carious Samplesa
Sound Region Sample
τ1 (ns)
%
τ2 (ns)
%
Carious Region τ3 (ns)
1 9:25 0:08 32.71 2:98 0:04 52.71 0:73 0:02 2 10:74 0:09 39.47 3:41 0:05 46.48 0:85 0:03 3 10:31 0:10 35.71 3:42 0:06 48.43 0:92 0:03 4 9:38 0:08 31.53 3:04 0:05 50.64 0:76 0:02 5 10:11 0:09 31.67 3:45 0:05 51.01 0:97 0:03 6 9:75 0:08 33.39 2:99 0:04 49.71 0:66 0:02 7 9:16 0:09 49.16 2:92 0:05 37.16 0:67 0:04 8 7:45 0:18 38.51 2:47 0:04 52.57 0:51 0:01 9 6:59 0:09 49.94 2:06 0:03 44.56 0:31 0:01 10 6:69 0:06 54.41 1:51 0:02 34.47 0:14 0:01 11 6:56 0:17 42.71 1:64 0:04 27.55 0:13 0:01 12 6:13 0:08 51.12 1:89 0:03 41.81 0:35 0:01 13 7:01 0:16 44.11 2:44 0:06 46.31 0:57 0:01 14 6:22 0:07 53.63 1:79 0:02 39.84 0:21 0:01 15 5:92 0:06 51.59 1:72 0:02 38.51 0:15 0:01 16 5:68 0:06 53.88 1:73 0:02 40.58 0:24 0:01 Average 7:93 0:09 2:46 0:04 0:51 0:02
% 14.58 14.05 15.86 17.84 17.32 16.92 12.89 8.93 5.51 11.13 29.74 7.07 9.59 6.49 9.91 5.54
τ1 (ns)
%
τ2 (ns)
%
τ3 (ns)
4:03 0:05 4:65 0:05 4:12 0:05 3:96 0:06 4:67 0:05 4:57 0:05 4:68 0:05 4:24 0:05 5:18 0:05 4:99 0:05 5:67 0:05 5:49 0:05 4:92 0:15 5:57 0:06 5:58 0:07 5:21 0:07 4:84 0:06
59.06 60.22 56.89 54.31 57.37 57.09 57.37 53.24 52.28 53.19 55.47 55.38 44.02 54.35 47.72 47.05
1:04 0:03 1:24 0:03 1:08 0:03 0:98 0:02 1:25 0:02 1:18 0:02 1:35 0:02 1:07 0:02 1:56 0:02 1:47 0:02 1:68 0:02 1:53 0:02 1:36 0:04 1:63 0:02 1:58 0:03 1:61 0:02 1:35 0:02
31.06 32.28 31.71 33.81 34.26 33.91 34.26 37.26 40.19 38.09 38.15 37.34 29.32 39.23 39.02 40.45
0:17 0:01 0:15 0:01 0:18 0:01 0:13 0:02 0:14 0:03 0:14 0:01 0:14 0:01 0:16 0:01 0:18 0:01 0:17 0:01 0:21 0:01 0:19 0:01 0:12 0:01 0:21 0:01 0:14 0:01 0:13 0:01 0:16 0:01
% 9.91 7.51 11.41 11.89 8.37 9.01 8.37 9.51 7.52 8.72 6.28 7.28 26.65 6.43 13.26 12.51
a This percentage was obtained by integration of the three exponential terms and preexponential terms as shown in the equation of Fig. 3.
20 April 2010 / Vol. 49, No. 12 / APPLIED OPTICS
2247
Fig. 4. Average values of three different lifetimes for carious and sound samples. There was a statistical difference (p < 0:05) between sound and carious enamel for all three lifetimes. The mean lifetime value of sound samples is higher than the lifetime of carious enamel.
while most of the endogenous fluorophores such as the free or bound amino acid tryptophan have a lifetime of 6 ns [11]. It has been shown that collagen I has an exponential lifetime of the order of 4 ns [12]. According to our results, displayed in Table 1, the average lifetime of the sound region was near 8 ns, and that of the carious enamel fell near 5 ns, which is probably due to the effect of collagen I, since this fluorophore is the compound present in larger amounts in the inorganic tooth matrix and is eliminated in a higher proportion in the presence of caries metabolism [17]. In the latter process, changes affecting collagen I are derived from partial hydrolyses of rich amino acid areas and poor polar chain areas. Reactions from most parts of the molecule polar chains, mainly from amino acid chains, with carbohydrate products produce dark complexes, which reduce the fluorescence lifetime of collagen I compared with sound tissues containing the same kind of collagen [10,13]. In the literature the sound enamel has its lifetime increased from 10 to 17 ns in the presence of caries [10]. This lifetime increase apparently disagrees with our results, which reveal a reduction in lifetime in the presence of caries: from 7.9 to 4:8 ns. It is important to notice that there is a difference in the carious process and the selection of emission wavelength in both works. Our samples were noncavitated caries, whereas the carious samples used in the literature are cavitated [10]; moreover, the emission fluorescence in our work was measured at 500 nm, far from the red emission at 600–700 nm assigned to porphyrins [8]. 5. Conclusion
The experimental system employed here consists in a tabletop nonportable time-resolved fluorescence spectroscope. Although the observed results point 2248
APPLIED OPTICS / Vol. 49, No. 12 / 20 April 2010
out its use to differentiate between carious and sound enamel, its application to the clinical setting is not possible because of its financial cost and size. However the same results can be achieved with a similar portable system using a pulsed LED with picosecond width, filters to select the aimed emission, and a time-resolved photomultiplier to detect the fluorescence decay. Once these experimental problems are overcome, the present results can be applied to the detection of noncavitated caries. It is concluded that it is possible to differentiate between carious and sound regions by time-resolved fluorescence. Our results showed that the average lifetime values of carious regions are lower than those of the sound region. It was also concluded that, although the origin of enamel fluorescence is still uncertain, the lifetime values seemed to be typical of fluorophores like collagen I. This study was supported by financial assistance of the Brazilian agencies FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico). A. S. Ito is a member of INCT-FCx. References 1. I. A. Pretty, “Caries detection and diagnosis: novel technologies,” J. Dent. Res. 34, 727–739 (2006). 2. S. M Higham, N. Pender, E. J Jong, and P. W. Smith, “Application of biophysical technologies in dental research,” J. Appl. Phys. 105, 102048 (2009). 3. V. Baelum, J. Heidmann, and B. Nyvad, “Dental caries paradigms in diagnosis and diagnostic research,” Eur. J. Oral Sci. 114, 263–277 (2006). 4. B. T. Amaechi, “Emerging technologies for diagnosis of dental caries: the road so far,” J. Appl. Phys. 105, 102047 (2009). 5. S. Tranaeus, X-Q. Shi, and B. Angmar-Manssson, “Caries risk assessment: methods available to clinicians for caries detection,” Community Dent. Oral Epidemiol. 33, 265–273 (2005). 6. A. Lussi, R. Hibst, and R. Paulus, “DIAGNOdent: an optical method for caries detection,” J. Dent. Res. 83, C80–C83 (2004). 7. W. Buchalla, “Comparative fluorescence spectroscopy shows differences in noncavitated enamel lesions,” Caries Res. 39, 150–156 (2005). 8. L. Bachmann, D. M. Zezell, A. C. Ribeiro, L. Gomes, and A. S. Ito, “Fluorescence of biological tissue: a review,” Appl. Spectrosc. Rev. 41, 575–590 (2006). 9. R. R. Alfano and S. S. Yao, “Human teeth with and without dental caries studied by visible luminescence spectroscopy,” J. Dent. Res. 60, 120–122 (1981). 10. K. König, H. Schneckenburguer, and R. Hibst, “Time-gated in vivo autofluorescence imaging of dental caries,” Cell. Mol. Biol. (Paris) 45, 233–239 (1999). 11. H. Schneckenburguer and K. Konig, “Fluorescence decay kinetics and imaging of NAD(P)H and flavins as metabolic indicators,” Opt. Eng. 31, 1447–1451 (1992). 12. Q Fang, T. Papaioannou, J. A. Jo, R. Vaitha, K. Shastry, and L. Marcu, “Time-domain laser-induced fluorescence spectroscopy apparatus for clinical diagnostics,” Rev. Sci. Instrum. 75, 151–162 (2004) 13. L. Marcu, D. Cohen, J. I. Maarek, and W. S. Grundfest, “Characterization of type I, II, III, IV, and V collagens by
time-resolved laser-induced fluorescence spectroscopy,” Proc. SPIE 3917, 93–101 (2000). 14. C. Robinson, R. C. Shore, S. J. Brookes, S. Strafford, S. R. Wood, and J. Kirkham, “The chemistry of enamel caries,” Crit. Rev. Oral Biol. Med. 11, 481–495 (2000). 15. K. Konig, G. Flemming, and R. Hibst, “Laser-induced autofluorescence spectroscopy of dental caries,” Cell. Mol. Biol. (Paris) 44, 1293–1300 (1998).
16. D. M. Zezell, A. 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, 064013 (2007). 17. A. C. R. Figueiredo, C. Kurachi, and V. S. Bagnato, “Comparison of fluorescence detection of carious dentin for different excitation wavelenghts,” Caries Res. 39, 393–396 (2005).
20 April 2010 / Vol. 49, No. 12 / APPLIED OPTICS
2249
