|Year : 2015 | Volume
| Issue : 3 | Page : 86-93
Single color attribute index for shade conformity judgment of dental resin composites
Director, Institute for Clinical Performance of Biomaterials (ICPB) and ETN Dental Clinic, Seoul, South Korea
|Date of Web Publication||28-Aug-2015|
Director, Institute for Clinical Performance of Biomaterials (ICPB) and ETN Dental Clinic, Seoul
Source of Support: None, Conflict of Interest: None
Introduction: Commercial dental resin composites under the same shade designations show color discrepancies by brand. Moreover, three Commission Internationale de l'Eclairage (CIE) color coordinates show significant variations by measurement method; therefore, direct comparisons of the color coordinates based on different methods are meaningless. This study aimed to assess a hypothesis that a new color attribute index (CAI), which could reduce the color coordinate variations by measurement method, was applicable for the shade conformity judgment of dental resin composites. The Hypothesis: CAI is applicable in the shade conformity judgment of commercial dental resin composites. Using the CIE color coordinates of shade guide tabs and resin composites, combined color indices such as Wa = CIE a* × DE* ab /C ab * and Wb = CIE b* × DE* ab /C ab * were defined, in which DE* ab was the color difference with a standard white tile. Ratio of Wa/Wb to that of an arbitrary reference shade (A2) in the same brand and measurement was defined as the CAI. The CAI values were significantly different by the shade designation and showed a logical trend by the shade designation number. The CAI of commercial resin composites and the keyed shade guide tabs showed overlaps. Evaluation of the Hypothesis: The CAI might be used to judge the shade conformity of resin composites using the values based on different measurement methods. The application of the CAI, instead of conventional three-color coordinates, could efficiently simplify the shade conformity judgment of commercial resin composites. Although the hypothesis of the present study was partially confirmed, further studies for the practical application of this index are highly recommended.
Keywords: Color attribute index (CAI), dental resin composite, shade conformity, shade guide, universal color index
|How to cite this article:|
Lee YK. Single color attribute index for shade conformity judgment of dental resin composites. Dent Hypotheses 2015;6:86-93
| Introduction|| |
Quantitative color measurements in dentistry usually employ the CIE (Commission Internationale de l'Eclairage) color system,  in which the CIE L*, a*, and b* coordinates are used for color specification. The CIE L* is a measure of lightness, a* is a measure of redness or greenness, and b* is a measure of yellowness or blueness. Chroma is calculated as C* ab = (a*2 + b*2 ) 1/2 , and color difference is calculated as DE* ab = (DL* 2 + Da* 2 + Db* 2 ) 1/2 . 
The CIE color coordinates vary significantly by measurement methods including the kind of instrument, such as spectrophotometer (SP) or spectroradiometer (SR); instrumental settings; standard observer; and illumination. ,,, As to the discrepancy by SP and SR measurements, a small-window color measurement by SP applied to turbid materials gives incorrect color coordinates.  Edge-loss effect is avoided by the SR measurement, as there are no apertures between the light source, instrument, and the object. 
A variety of measurement methods have been used in dental color studies and it is well known that direct comparisons of the color coordinates measured with different methods are meaningless. Moreover, it is not easy to compare the color values of objects with the 3-dimensional color coordinate system, simultaneously considering the influence of the measurement method. Therefore, a universal color attribute index (CAI), in which discrepancies in color values by measurement method are reduced or eliminated, would simplify the color conformity judgment of dental materials.
Shade designations, such as A2 or A3, of aaesthetic dental materials are commonly keyed to a commercial shade guide (Vitapan Classical [VPC]; VITA Zahnfabrik, Germany). This shade guide is divided into four hue groups (A, B, C, and D) and each group includes several tabs. , Other shade guides such as Chromascop (CMS; Ivoclar Vivadent, Liechtenstein) and Toothguide 3D-Master (3DM; VITA Zahnfabrik) are also available. However, perceivable color differences have been reported among the aesthetic materials under the same shade designations by manufacturer, brand from the same manufactuer, and even shade group in the same brand, which also indicated substantial color discrepancies of these materials from their keyed shade guide tabs. ,,,,,,,, As a series of the shade conformity studies, firstly the influence of the variables in instrumental color measurements was evaluated.  Secondly, shade compatibility of aesthetic restorative materials was reviewed.  In this study, a hypothesis whether a CAI was applicable for the shade conformity judgment of dental resin composites was assessed.
Shade conformity or consistency judgment of all the marketed dental resin composites based on the same measurement method is not possible because there are so many brands of materials. Combined color indices instead of the CIE color coordinates might be an alternative. However, there have been no studies that tried to devise a combined CAI. The objectives of this study were to define a universal CAI that might be used to compare the color values measured with different methods primarily based on SP, and to evaluate the applicability of this index.
| The Hypothesis|| |
The hypothesis was that the CAI, devised in the preset study, was applicable in the shade conformity judgment of commercial dental resin composites.
| Experimental methods and results supporting the hypothesis|| |
The influence of the color measurement method on the CIE color coordinates of a standardized object and shade guide tabs was evaluated (part 1 and 2), and several combined color indices that could reduce the discrepancies in color values by the measurement method were defined (3). After then, the applicability of the CAI for the shade conformity judgment was considered (4 and 5).
Color of standardized object
The measurement of the CIE color coordinates of a standardized object (white tiles used for the calibration of SP) was the starting point. To determine the variations by measurement method based on SP, the color coordinates of a standard white tile (white reference standard, X-Rite, USA) were measured by two SPs with three aperture sizes. Color was measured under external light excluded condition relative to the standard illuminant D65.  3-SP and 8-SP values were measured by a SP (CM-3500d, Minolta, Japan) with the aperture diameters of 3 mm and 8 mm under the illuminating and viewing configuration of diffuse/10Ί. Values of 3×8-SP were measured by another SP (ColorEye7000A, X-Rite) with the aperture size of 3×8 mm 2 under the configuration of diffuse/8Ί.
Based on one-way analysis of variance (ANOVA), there was no significant difference in each of three-color coordinates (P > 0.05). Therefore, it was confirmed that the standard white tile could be used as a standardized object with different measurement methods based on SP.
Color consistency of shade guide tabs by measurement method
Correlations between two sets of SP-based and one set of SR-based  color coordinates of VPC, CMS, and 3DM shade guides were assessed. In SP measurements, the color was measured under an external light excluded condition relative to the standard illuminant D65.  3-SP and 3×8-SP values were measured following the measurement methods for the white tile. SR values were obtained from a published paper.  In this reference, the spectral reflectance values were measured with a spectroradiometer at wavelengths from 380-780 nm at 2-nm intervals using 0Ί observer and 45Ί illumination. And the mean CIE L*, a*, and b* values of the VPC guide were 74.9, 0.04, and 19.5, respectively.  Color coordinates of three shade guides based on three measurements were compared.
Correlations in the color coordinates of three shade guides with three methods are presented in [Table 1]. Differences in the CIE L*, a*, and b* values by the measurement method were in the ranges of 7.1-25.3, -0.3-1.4, and 1.8-14.6 CIE units, respectively. However, significant correlations were found between the corresponding color coordinates measured with different methods (range of the coefficient of determination [r2 ]: 0.79-0.99, P < 0.05). Therefore, it was confirmed that relative values, instead of the three-color coordinates themselves, could be used.
|Table 1: Correlations and differences in the CIE color coordinates of shade guides by measurement method |
Click here to view
Combined color indices of VPC shade guide
To define the CAI, several preliminary indices were devised. Wa (relative red-green index) and Wb (relative yellow-blue index) were defined as Wa = DE* ab × CIE a*/C ab * and Wb = DE* ab × CIE b*/C ab *, in which DE* ab was the color difference of each shade tab with a standard white tile under the same measurement and C ab * was the chroma of the tab. Wa and Wb values of the VPC guide based on three methods (3-SP, 3 × 8-SP and SR) were compared.
Based on Wa and Wb values, one set of two color-index system (WaR and WbR) and one color-index system (CAI) were devised. WaR and WbR were defined as the ratios of Wa and Wb values of a shade tab to those of an arbitrary reference shade tab (A2), respectively, under the same measurement. The A2 shade was selected because this shade was the most frequently included shade in the cited papers of the present study. The CAI was defined as the ratio of Wa/Wb values of a shade tab to that of the A2 tab under the same measurement. In summary, WaR = Wa/Wa-A2 tab, WbR = Wb/Wb-A2 tab and CAI = (Wa/Wb)/(Wa-A2 tab/Wb-A2 tab). WaR, WbR and the CAI values of the VPC guide with three methods were compared.
Wa and Wb of the VPC guide are presented in [Figure 1] and [Figure 2]. The range of Wa was -3.4-7.6 and they showed similar trends by the shade designation regardless of the measurement method except for the D group. In the D group, the CIE a* values and the corresponding Wa values were distributed near x-axis, which might be the reason for the difference by the measurement method. However, further confirmation is necessary. The range of Wb was 16.6-49.3 and they showed similar trends by the shade designation regardless of the measurement method. Distribution of the Wa/Wb value of the VPC guide is presented in [Figure 3]. The range of Wa/Wb values was −0.2-0.2. They showed similar trends by the shade designation regardless of the measurement method.
|Figure 1: Distribution of Wa of VPC guide with three measurement methods|
Click here to view
|Figure 2: Distribution of Wb of VPC guide with three measurement methods|
Click here to view
|Figure 3: Distribution of Wa/Wb of VPC guide with three measurement methods|
Click here to view
The CAI, WaR, and WbR values of the VPC guide with three methods are presented in [Figure 4]. The range of the CAI was −2.7-3.4, that of WaR was -2.3-4.6, and that of WbR was 0.9-1.4. The CAI value generally followed WaR value.
|Figure 4: Distributions of CAI, WaR, and WbR of VPC guide with three measurement methods (mean ± SD)|
Click here to view
Correlations between the combined color indices of the VPC guide with three measurement methods are listed in [Table 2]. The coefficients of determination were in the range of 0.71-0.98 for Wa and Wb, 0.90-0.98 for Wa/Wb, 0.90-0.98 for CAI, and 0.72-0.98 for WaR and WbR (P < 0.05). Differences in the combined indices were in the range of −18.7-0.5 for Wa and Wb, and −0.7-0.1 for Wa/Wb, WaR, WbR and CAI.
|Table 2: Multiple regression results for combined indices under varied measurement methods of VPC guide |
Click here to view
CAI of resin composites
To determine the applicability of the CAI on the shade conformity judgment of resin composites, the CIE color coordinates of commercial resin composites in 15 published papers were analyzed. ,,,,,,,,,,,,,, As to the specimens in the included papers, the thickness was 1 mm in 5 papers, 1.6 mm and 3.5-4.0 mm in 1 paper, and 2 mm in 9 papers. The same shade designations regardless of body, dentine, enamel, opaque, pediatric, or flowable types were regarded as the same shade.
Wa, Wb, WaR, and WbR were calculated with the same method for the VPC guide. For the calculation of the CAI of resin composites, the A2 shade of the same brand under the same measurement was used as the reference shade. In the A2 shade, the CAI was calculated as Wa/Wb of each material/mean Wa/Wb of the A2 shade in the same measurement. In other shades, the CAI was calculated as Wa/Wb of each shade/mean Wa/Wb of the A2 shade in the same measurement.
Shades that have at least five validated data were included in the analyses. Number of included materials for each shade was as follows: A1 shade, 5; A2, 23; A3, 10; A3.5, 8; B2, 6; B3, 5; and C2, 6. SP-type instruments were used in the included papers. Measurement geometry was specular component excluded (SCE), specular component included (SCI) or not specified,  and measurement aperture size was 3 mm, 8 mm, 3×8 mm 2 or not specified. Each mean value in the included papers was regarded as one value.
Distributions of the lightness (CIE L*) and chroma, and the CIE a* and b* values of commercial resin composites are presented in [Figure 5] and [Figure 6]. The ranges of the CIE L*, a*, b*, and chroma of the included shades were 61.2-71.2, −1.0-3.7, 8.2-22.1, and 8.3-22.4, respectively. Based on one-way ANOVA by the shade designation, the CIE a*, b*, and chroma were influenced by the shade designation while CIE L* was not (P = 0.05).
|Figure 5: Distributions of lightness (CIE L*) and chroma of commercial resin composites|
Click here to view
Distributions of Wa and Wb of resin composites are presented in [Figure 7]. The range of Wa was −6.2-5.3 and that of Wb was 26.0-35.6. Based on one-way ANOVA by the shade designation, Wa was influenced by the shade designation while Wb was not (P = 0.05). Resin composites under the same shade designations showed high color discrepancies, and there was limited shade designation-dependent distributions in Wa and Wb values of resin composites.
Wa and Wb values of resin composites and those of the keyed VPC shade tabs are presented in [Figure 8] and [Figure 9]. In both of Wa and Wb, the values of resin composites and shade tabs were in the similar range in most of the shades and showed similar trends by the shade designation.
Application of CAI in shade conformity judgment
The CAI values of resin composites and those of the keyed VPC shade tabs are presented in [Figure 10]. They showed similar trends to those of Wa values [Figure 8]. Based on one-way ANOVA by the shade designation, the CAI was influenced by the shade designation (P < 0.05). In case of the B3 shade, the CAI values of resin composites and shade tab showed some deviations possibly because small number of brands (n = 5) was included in the analysis.
| Evaluation of the Hypothesis|| |
In this study, a single combined CAI based on the CIE color coordinates of dental shade guides and commercial resin composites was assessed for the possible application to the shade conformity judgment of these materials. Comparing the CAI values of resin composites and those of the keyed VPC shade tabs [Figure 1]0], they showed similar distributions following the shade designation. Although they showed broad range of standard deviations, the CAI was influenced by the shade designation (P < 0.05). Since the fact confirmed in the present study was that the CAI of resin composites and keyed VPC shade tabs showed similar range of values, it is hard to argue that the CAI can be efficiently applied to judge the shade conformity of these materials. However, discrepancies in the color coordinates by measurement method have at least partially reduced in the CAI value since the color coordinates of standard color (white tile) and the reference shade (the A2 shade) under the same measurement method were included in the calculation process.
Defining new colorimetric indices, which are simple functions of the well-known CIE coordinates, in order to try to eliminate the measurement variables might seem absurd. Therefore, several factors were considered in the present study. Firstly, the CIE L* coordinate was indirectly included in the CAI because the color difference (DE* ab ) with the standard white tile included the lightness difference by measurement method. Moreover, in the commercial resin composites included in the present study, the CIE L* was not influenced by the shade designation (P > 0.05). Concerning the white tiles, since they are routinely used for the calibration of SP, they can be generally used as a standard. Also CIE a*/C ab * and b*/C ab * indicated the relative CIE a* and b* coordinate portions within the chroma value. Secondly, as compared with the color difference (DE* ab ) with respect to the standard white tile or a shade reference, the differences in color coordinates by the measurement variables have at least partially reflected in the CAI value.
As to the problems of the present study, firstly, the physical meanings of the color indices defined in this study were not confirmed. Therefore, threshold values for visual or perceptual interpretation for shade conformity judgment with new index could not be proposed. Further study should be performed. Secondly, although DE* ab color difference was calculated according to the CIE specification,  application of new color difference formula such as CIEDE2000 (DE 00 ) might have provided more practical information. The CIEDE2000 color difference formula, which includes not only lightness, but chroma and hue weighting functions as well,  was introduced and officially adopted as a new CIE color-difference equation. 
From now, discussions followed the order of previous section for experimental data supporting the hypothesis. As the first preliminary step to define the CAI, variations in the color coordinates of the standard white tile and shade guide tabs by the measurement method were assessed. As to the CIE color coordinates of the standard white tile with different measurement methods, there were no significant differences in the color coordinates (P > 0.05). Therefore, the standard white tiles could be regarded as a standardized object. However, reconfirmation should be done in other independent studies.
The color of shade guide tabs was assessed to determine the influence of measurement method on the CIE color coordinates. Shade guides were investigated because a series of shades encompassing varied hue, lightness and chroma of natural teeth are included in the shade guides. Although the VPC guide is the commonly keyed shade guide, CMS and 3DM guides were also investigated to enlarge the shade ranges. As indicated in [Table 1], they showed significant correlations (P < 0.05). Therefore, it was reconfirmed that the combined color indices, which could reduce or eliminate the influence of measurement method, could be used when comparing the color values based on different methods. SR-based values were included to enlarge the application field. Since the color coordinates of shade guides showed strong correlations regardless of the measurement methods (r2 = 0.79-0.99, [Table 1]), combined color indices were devised. Wa and Wb were defined regarding two aspects. Firstly, the color difference (DE* ab ) with the standard color indicates the distance in the color of an object from the standard color. Secondly, the ratios of the CIE a* or b* value with chroma indicate the hue angle of an object including the distance from zero (gray color) point in the CIE a* (x-axis) and b* (y-axis) plane. Therefore, Wa and Wb values could be regarded to encompass all the three CIE color coordinates. In the present study, Wa and Wb values of the VPC guide [Figure 1] and [Figure 2] showed similar trends with different measurement methods.
The kind of aesthetic dental materials, such as ceramic or resin composite, influenced the color coordinates differently by measurement method.  Therefore, Wa and Wb indices were further converted to "ratio form"by comparison with the reference shade (A2) of the same kind of material under the same measurement. Based on the distributions of the CAI, WaR, and WbR indices of the VPC guide [Figure 4], the CAI and WaR shared similar trends by the shade designation regardless of the measurement method [Table 2]. Because the CAI generally followed WaR, further analyses by WaR and WbR with resin composites were not tried. Since the CAI value included both of WaR and WbR indices, the CAI could be regarded to represent the all the color attributes of resin composites.
SP-type instruments were used in the 15 included papers of the present study to determine the color of commercial resin composites; therefore, difference in instrument settings in each study might have influenced the color coordinates differentially. Based on the distributions of the CIE L*, a*, b*, and chroma of commercial resin composites [Figure 5] and [Figure 6], discrepancies within the same shade designated materials were high and there was limited logical order in shade distribution by the shade designation. However, it should be mentioned that some other factors such as thickness and surface condition of investigated specimens might also have increased the variations. Specimen thicknesses were reported as 1-4 mm (generally 2 mm). The number of investigated materials for each shade was in the range of 5-23, which might have influenced the validity of the color coordinates because those values based on limited numbers of data might have not represented true values.
The ranges of Wa and Wb indices of commercial resin composites were large within the same designated shades, and limited logical order was also observed in the shade distribution by the shade designation [Figure 7]. Compared with the mean values of the keyed VPC shade tabs measured with three measurement methods [Figure 8] and [Figure 9], although the standard deviations were high in both of the shade guide tabs and resin composites, the values for each shade of resin composite and keyed tab were in the similar range, which also indicated that the shade of the resin composites reflected the shade of the keyed shade tab. The comparison with the mean values of shade tabs were made because the color measurement methods in the included papers were highly variable. Therefore, instead of the values measured under one protocol, it was supposed that mean values based on three methods might have represented relatively subjective values. As to the CAI of resin composites and the keyed VPC tabs [Figure 1]0], they showed similar trends to those of Wa values [Figure 8], and the CAI index was significantly influenced by the shade designation (P < 0.05).
Based on the present results and further studies, the ranges for the CAI for each shade of material that is allowable for successful color reproduction of resin composites should be further decided. The CAI of the same shade designated resin composites should be within predefined range, and the range should be determined based on the determination of the perceptible or acceptable color differences. Since the numbers of the data were limited in the present study, the allowable range could not be decided.
| Conclusion|| |
The CAI might be used to judge the shade conformity of resin composites based on the color coordinates with different measurement protocols because 1) this value was significantly different by the shade designation; 2) this value showed a generalized trend by the shade designation number in each shade group; and 3) the CAI of resin composites and the keyed shade tabs generally showed overlaps. Although the CAI was introduced in the present study, it is obvious that these transformations could not have totally excluded the influence of the measurement method on the color value.
Financial support and sponsorship
The authors do not have any financial support.
Conflicts of interest
There are no conflicts of interest.
| References|| |
CIE (Commission Internationale de l'Eclairage). Colorimetry - Technical Report. 3 rd
ed. Vienna, Austria: Bureau Central de la CIE, CIE Pub. No.15; 2004.
Park SK, Lee YK. Color distribution of commercial resin composites and color difference with shade guide tabs. Am J Dent 2007;20:335-9.
Kim BJ, Lee YK. Influence of the shade designation on the color difference between the same shade-designated resin composites by the brand. Dent Mater 2009;25:1148-54.
Lee YK, Lim BS, Kim CW. Effect of surface conditions on the color of dental resin composites. J Biomed Mater Res 2002;63:657-63.
Lee YK, Lim BS, Kim CW. Influence of illuminating and viewing aperture size on the color of dental resin composites. Dent Mater 2004;20:116-23.
van der Burgt TP, ten Bosch JJ, Borsboom PC, Kortsmit WJ. A comparison of new and conventional methods for quantification of tooth color. J Prosthet Dent 1990;63:155-62.
Bayindir F, Kuo S, Johnston WM, Wee AG. Coverage error of three conceptually different shade guide systems to vital unrestored dentition. J Prosthet Dent 2007;98:175-85.
Miller LL. A scientific approach to shade matching. In: Preston JD, editor. Perspectives in Dental Ceramics: Proceedings of the Fourth International Symposium on Ceramics. 15 th
ed. Chicago: Quintessence Pub. Co.; 1988. p. 193-208.
Schwabacher WB, Goodkind RJ. Three-dimensional color coordinates of natural teeth compared with three shade guides. J Prosthet Dent 1990;64:425-31.
Yap AU, Tan KB, Bhole S. Comparison of aesthetic properties of tooth-colored restorative materials. Oper Dent 1997;22:167-72.
King KA, deRijk WG. Variations of L*a*b* values among Vitapan Classical Shade Guides. J Prosthodont 2007;16:352-6.
Browning WD, Contreras-Bulnes R, Brackett MG, Brackett WW. Color differences: Polymerized composite and corresponding Vitapan Classical shade tab. J Dent 2009;37 (Suppl 1):e34-9.
Paravina RD, Kimura M, Powers JM. Color compatability of resin composites of identical shade designation. Quintessence Int 2006;37:713-9.
Analoui M, Papkosta E, Cochran M, Matis B. Designing visually optimal shade guides. J Prosthet Dent 2004;92:371-6.
Kim HS, Um CM. Color differences between resin composites and shade guides. Quintessence Int 1996;27:559-67.
Lee YK, Yu B, Lee SH, Cho MS, Lee CY, Lim HN. Variation in instrument-based color coordinates of esthetic restorative materials by measurement method - A review. Dent Mater 2010;26:1098-105.
Lee YK, Yu B, Lee SH, Cho MS, Lee CY, Lim HN. Shade compatibility of esthetic restorative materials - A review. Dent Mater 2010;26:1119-26.
Lee YK, Lim BS, Kim CW, Powers JM. Comparison of color of resin composites of white and translucent shades with two shade guides. J Esthet Restor Dent 2001;13:179-86.
Lee YK, Lim BS, Rhee SH, Yang HC, Powers JM. Color and translucency of A2 shade resin composites after curing, polishing and thermocycling. Oper Dent 2005;30:436-42.
Kim IJ, Lee YK. Changes in color and color parameters of dental resin composites after polymerization. J Biomed Mater Res B Appl Biomater 2007;80:541-6.
Hosoya Y, Shiraishi T, Oshiro M, Ando S, Miyazaki M, Powers JM. Effects of specular component on color differences of different filler type resin composites after aging. J Dent 2009;37:585-90.
Woo ST, Yu B, Ahn JS, Lee YK. Comparison of translucency between indirect and direct resin composites. J Dent 2008;36:637-42.
Jeong TS, Kang HS, Kim SK, Kim S, Kim HI, Kwon YH. The effect of resin shades on microhardness, polymerization shrinkage, and color change of dental composite resins. Dent Mater J 2009;28:438-45.
Yu B, Lee YK. Differences in color, translucency and fluorescence between flowable and universal resin composites. J Dent 2008;36:840-6.
Murai H, Katayama T. Color of the Composite Resin for ESTELITE PRO. Bangkok, Thailand: International Dental Materials Congress, PS 1; 2007. p. 178.
Yu B, Lee YK. Comparison of the color stability of flowable and universal resin composites. Am J Dent 2009;22:160-4.
Lee YK. Comparison of CIELAB DE *
and CIEDE2000 color-differences after polymerization and thermocycling of resin composites. Dent Mater 2005;21:678-82.
Technical CIE. Report: Improvement to industrial color difference equation. Vienna: Central Bureau of the CIE, CIE Pub. No. 142-2001; 2001.
Lagouvardos PE, Diamanti H, Polyzois G. Effect of individual shades on reliability and validity of observers in color matching. Eur J Prosthodont Restor Dent 2004;12:51-6.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9], [Figure 10]
[Table 1], [Table 2]