17 AN EXPLORATORY STUDY OF HUMAN TEETH ENAMEL BY USING FT-RAMAN SPECTROSCOPY

Transcription

1 17 AN EXPLORATORY STUDY OF HUMAN TEETH ENAMEL BY USING FT-RAMAN SPECTROSCOPY Afishah Alias 1, Siti Rahayu Mohd Hashim 1, Judith Mihaly 2, Julynnie Wajir 1 & Fauziah Abdul Aziz 1 1 School of Science and Technology, Universiti Malaysia Sabah, Kota Kinabalu, 88502, Sabah, MALAYSIA 2 Chemical Research Center, Hungarian Academy of Sciences, H-1525, Budapest, P.O. Box 17, Hungary ABSTRACT Unaffected, affected and heavily affected teeth enamel were studied by using FT-Raman spectroscopy. The 14 permanent teeth s enamel surface were measured randomly, resulting in total n=43 FT-Raman spectra. The results obtained from FT-Raman spectra of heavily affected, affected and unaffected tooth s enamel surfaces did not show any significant difference. In this study, Kruskal-Wallis and Wilcoxon rank sum tests were used to compare the intensity between the categories of enamel as well as the surfaces of teeth samples. ABSTRAK Tidak terjejas, terjejas dan terjejas dengan teruk enamel gigi telah dipelajari dengan menggunakan spektroskopi FT-Raman. 14 enamel gigi kekal permukaan telah diukur secara rawak, mengakibatkan dalam jumlah n=43 spektrum FT-Raman. Keputusan-keputusan itu diperolehi dari spektrum FT-Raman terjejas dengan teruk, gigi terjejas dan tidak terjejas permukaan-permukaan enamel melakukan tak tunjuk mana-mana perbezaan penting. Dalam kajian ini, Kruskal-Wallis dan Wilcoxon ujian-ujian jumlah lebat adalah sudah biasa bandingkan keamatan antara kategori-kategori enamel serta permukaan-permukaan contohcontoh gigi. Keywords: human teeth enamel, FT- Raman spectroscopy, Kruskal-Wallis, Wilcoxon INTRODUCTION Dental caries is the common oral disease in the developed country. Carious lesions are regularly found at area between adjacent teeth. The detection of the caries lesions in these areas is difficult. These characteristics of the caries problem magnify the limitations of caries detection through conventional diagnostic methods that involve subjective clinical criteria and the use of diagnostic tools such as the dental explorer and dental radiographs (Alex et al., 2006). These conventional methods are adequate for detection of larger, possibly cavitated lesions, but due to poor specify and sensitivity, they are not suitable for detection of early stage, non-cavitated lesions (Alex et al., 2006). Therefore, better diagnostic tools are needed to detect early noncavitated lesions and to monitor their activity (Alex et al., 2006). Polarized Raman spectroscopy and optical coherence tomography have been used to detect these changes and potentially offer a means to detect and monitor early caries development (Michael et al., 2006). 141

2 In this study, unaffected, affected and heavily affected teeth enamel were studied by using FT- Raman spectroscopy and the first attempt was to analyze the data by using General Linear Modeling (GLM) method, two-way ANOVA. Unfortunately, the tests failed to identify significant differences of intensity amongst the three conditions. The results could be due to the critical violation of the normality assumption. This study has carried out non-parametric test as an alternative, as the data failed to follow the normality distribution after several transformations. EXPERIMENTAL METHOD The samples of human adult teeth were obtained from Dr. Rashid Dental Clinic. From these teeth samples obtained, its have been categorized to 3 types of teeth: - heavily affected, affected and unaffected enamel. The enamel that has more than 3 affected surfaces was classified as heavily affected enamel, 3 or less affected surfaces as affected enamel and has non affected surface as unaffected enamel. They were then sterilized with sodium-hypochlorite of 5% water solution for 4 hours. These were then treated in a 6.85 M KOH water solution for 50 hour while changing the solution for every 2-3 hour (Fattibene et al., 2005). The enamel samples were then investigated by FT-Raman spectroscopic method at Chemical Research Center, Hungarian Academy of Sciences, Budapest, Hungary. The measurements were performed by BioRad (Digilab) dedicated FT-Raman spectrometer equipped with liquid nitrogen cooled germanium detector and Nd:YAG excitation laser with radiation line at 1064 nm. No sample preparation was required and the measurements were completely nondestructive. All spectra were collected using 500mW laser power, 4 cm -1 resolution and coaddition of 512 individual spectra. The spectra were than baseline corrected and normalized to the strongest 958 cm -1 peak, using Grams32 spectral evaluation program. The Kruskal-Wallis test is applied to compare two or more populations with the data are either ordinal or interval but non-normal (Keller and Warrack, 2003). In this study, the populations referred to are the classification of enamels and the position of enamel during the spectroscopy test. The test statistic of Kruskal-Wallis test is; H = n 12 k T + j= 1 n ( n 1) 2 j j 3 ( n + 1) (1) Where, n = the number of observations k = the number of populations compared j = 3 ( classification of 2 (enamel positions) T = the total of ranks However, the Kruskal-Wallis can determine only whether a difference exists. To determine whether one population is larger than another, this study applied the Wilcoxon rank sum test. The standardized test statistic is; Where, ( ) T E T Z = σ T (2) 142

3 E (T) = n ( n + n 1) (3) and, T = n ( n + n 1) 1 n σ (4) 12 Both n 1 and n 2 is the number of observations belonged to two groups, lower ranks group and higher ranks group respectively. RESULT AND DISCUSSION The results obtained from FT-Raman spectra of heavily affected, affected and unaffected tooth enamel s surfaces did not show any significant difference. Figure 1 showed FT-Raman spectra of heavily affected (upper spectrum), affected (middle spectrum) and unaffected (lower spectrum) tooth s enamel surfaces. For discriminating the 3 types of tooth enamel with FT- Raman spectral features, Kruskal-Wallis was applied. The result in Table 1 showed that there is significant difference between the types of enamel and the positions of enamel with p-value < for both tests. Figure 2 showed that the mean rank of intensity for side enamel surface, , is significantly higher than the occlusal enamel which is The mean rank of intensity of heavily affected enamel, , is found much higher than the affected and unaffected enamels with the mean rank are and respectively. But to identify which type of enamel is significantly difference with the other type of enamel, the Wilcoxon rank sum tests are carried out and the results can be found in Figure 3. From table 2, the p-values (both < 0.001) showed that the mean rank of intensity for heavily affected enamel is significantly difference with unaffected and affected enamel. Whereas, the mean rank of intensity of affected and unaffected enamel are not differ significantly with p- value = (> 0.05). It showed that the difference of the mean rank of intensity is depends on the categories of the teeth ν 1 PO A CO ν 3 PO ν 4 PO ν 2 PO Figure 1: Spektra FT-Raman of heavily affected (upper spectrum), affected (middle spectrum) and unaffected (lower spectrum) enamel teeth. 143

4 Table 1: The Kruskal-Wallis results on determining a significant difference of intensity between the types of enamel and the surface. Intensity by enamel types Intensity by surface Chi-square statistic Degree of freedom 2 1 Significant value, p < < Mean ranks between side enamel surface and occlusal enamel Intensity Surface side enamel surface occlusal surface Figure 2: The mean ranks between side enamel surface and occlusal enamel. 144

5 Mean ranks of three types of enamel from the Kruskal Wallis test Intensity category heavily affected affected unaffected Figure 3: The mean ranks of three types of enamel from the Kruskal_Wallis test. Table 2: The result of the Wilcoxon rank sum tests. category N Mean Rank Sum of Ranks Z p-value intensity heavily affected unaffected < Total 6928 intensity heavily affected affected < Total 6928 intensity affected unaffected Total

6 CONCLUSIONS The results obtained from FT-Raman spectra of heavily affected, affected and unaffected tooth s enamel surfaces did not show any significant difference. However, by using Kruskal- Wallis and Wilcoxon rank sum tests, the intensity between the categories of enamel as well as the surfaces of teeth samples can be discriminated. The results showed that Wilcoxon rank sum tests can be an indicator to characterize the conditions of the teeth. ACKNOWLEGMENT The authors acknowledge the Hungarian Academy of Sciences and Universiti Malaysia Sabah for full support and cooperation. REFERENCES Alex C.-T. Ko, Lin-P'ing Choo-Smith, Mark Hewko, Michael G. Sowa, Cecilia C. S. Dong, and Blaine Cleghorn., (2006), Detection of early dental caries using polarized Raman spectroscopy Optics Express, Vol. 14, 1: Michael, G. S., Dan P. Popescu, Jeffrey Werner, Mark Hewko, Alex C.-T. Ko, Jeri Payette, Cecilia C. S. Dong, Blaine Cleghorn and Lin-P ing Choo-Smith., (2006), Precision of Raman depolarization and optical attenuation measurements of sound tooth enamel. Analytical and Bioanalytical Chemistry. ISSN Vol. 387, 5: Keller, G. and Warrack, B., (2003), Statistics for Management and Economics. ISBN , 6 th edition, Thomson Learning. 146