Evaluation of the repeatability and matching accuracy between two identical intraoral spectrophotometers: an in vivo and in vitro study

Color matching would be one of the essential factor to achieve good esthetic results of dental restorations. The perceived color can be determined by complex phenomena, such as specular transmission, specular reflection, diffuse light reflection, absorption, and scattering as a result of light interaction with the tooth.1 Visual color determination of a natural tooth can vary individually depending on anatomic and environmental conditions, leading to unreliable color assessment.2, 3, 4 Electronic instrumental color measurement can be more consistent because the shade taking devices are not influenced by subjective variables of color analyses.5 ISO defined the device repeatability as consistency of measurements of color parameters or matching to shade systems in repeated measurements, while the inter-device reliability as the degree of agreement among devices in color measurements or matching scales.6

Numerous studies have analyzed the repeatability7, 8, 9, 10, 11, 12, 13, 14 and accuracy7, 9, 13, 14, 15 of intraoral color-measuring devices. The evaluated intraoral color-measuring devices provided relatively good repeatability of color values and thus, they could be useful for determining and communicating tooth color. On the contrary, the measured inter-device agreement was poor to high depending on the types of devices and the measurement conditions. In addition, the operating systems of the devices could affect the reliability and accuracy of tooth color measurements. A spectrophotometer is an analytical instrument used to quantitatively measure the reflection or transmission properties of a material as a function of wavelength, while a colorimeter detects the amount of light passing through a series of color filters.16 According to the previous study, spot-measurement intraoral spectrophotometer displayed higher reliability than intraoral spectrophotometer of entire tooth measurement or colorimeter.17 In another study, higher color values were noted with the intraoral spectrophotometer compared to the colorimeter.8

Several factors could affect clinical performance of the intraoral spectrophotometer.

The devices measure the curved tooth surface instead of flat surface and small aperture could yield edge-loss effect which can cause a deviation of color interpretation.18 The translucency of the specimens can affect the device repeatability and accuracy.19 Furthermore, it is not always possible to position the aperture on the identical tooth surface.13 VITA Easyshade V (VITA Zahnfabrik, Bad Säckingen, Germany), which is the subject of this study, is a portable spectrophotometer and a new fifth generation of Easyshade. A technology has been developed that allow shade information to be interpreted in the VITA 3D-Master system, VITA classical system, VITABLOCS shades, or tooth bleaching shades. Color communication can be possible with the use of smartphone app VITA mobileAssist via Bluetooth connection.

Although clinical performances of several color measuring devices have been investigated, there is only a limited number of in vivo evaluations of the repeatability and accuracy between identical intraoral spectrophotometers. Therefore, the purpose of this study was to evaluate the repeatability and accuracy of 2 identical intraoral spectrophotometers with different serial numbers. The null hypotheses to be tested were that there were no differences in color values and shade within and between 2 identical intraoral spectrophotometers to measure tooth color and that the color difference between recorded tooth shade given by the device and the matching shade guide would be within the range of 50:50% perceptibility threshold (ΔE₀₀ < 1.3).

This study was approved by the institutional review board of Ajou University Hospital (no. AJIRB-MED-SUR-17-330). Thirty subjects, patients of the Department of Dentistry of Ajou University Medical Center (16 women and 14 men, mean age 32 ± 8 years), were recruited. Every participants received written information and signed an informed consent for the study. The maxillary right central incisor, canine, and the mandibular left central incisor of each of 30 participants were measured using 2 identical intraoral spectrophotometers with different serial numbers (VITA Easyshade V, Table 1). Un-restored natural teeth without caries, heterogeneous staining,17 and irregular surface texture8 were selected. Before the measurement, any specks from the tooth surface were removed and the tooth was wiped off with gauze to prevent any sliding. Subjects sat in the treatment chair and rested their head on a headrest for stabilization with their mouths slightly open.

Table 1
Characteristics of VITA Easyshade V

• Click for larger image
• Click for full table
• Download as Excel file

All measurements were performed by a single experienced operator. For every measurement, each device was conducted a white balance on its calibration block. The shade was measured at the central tooth area of labial surface8, 15, 17 in “base shade determination” mode with the measuring tip lying flush on the tooth surface to achieve an accurate measurement (Fig. 1). The color measurements were carried out 3 times with an interval of 1 hour for each device.8, 17, 20 Thus, a total of 270 readings per each device were obtained. The measurement results could be displayed in the VITA 3D-Master system, VITA classical shade system, VITABLOC shade, or bleach index. In this study, VITA 3D-Master system was used to analyze matching performance of the devices. Shade coordinates were also obtained by clicking 3D-Master shade. The degree of matching to shade guide system was indicated by a green (good), yellow (average), or red bar (adjust). In this study, only “good” matching results were used (Fig. 2).

Fig. 1
The Easyshade V determined tooth color with the probe tip placing perpendicular to the tooth surface.

• Click for larger image
• Download as PowerPoint slide

Fig. 2
The measurement result was displayed in the VITA 3D-Master and VITA classical systems (A) and the color coordinates could be obtained by clicking the shade value (B).

• Click for larger image
• Download as PowerPoint slide

To verify the measurement repeatability within devices, the intraclass correlation coefficient (ICC) and 1-way ANOVA followed by Bonferroni multiple comparison for L^*,a^*, and b^* values were analyzed. The ICC was obtained as an average measurement with absolute agreement and a 2-way mixed effect model. For matching reliability within devices, whether or not the obtained shade matched the correct 3D-Master shade tab, intra-device matching agreement rate was calculated.

To assess the inter-device measurement reliability, ICC and paired t-test were conducted for L^*,a^*, and b^* values by using average values of three measurements for each tooth. The CIEDE2000 color differences (ΔE₀₀) between color values from both devices were also calculated;

where ΔC′ and ΔH′ are the differences in chroma and hue for a pair of color values from the devices, S_L,S_C, and S_H are the weighting functions, and the parametric factors, K_L,K_C, and K_H are the correction terms for variations in experimental conditions. R_T is a rotation function for the interaction between chroma and hue differences in the blue region.21, 22 For matching reliability between devices, the shade indicated by each device were compared and a percentage of identical matches to the shade tab for a total of 270 pairs was calculated as an inter-device matching agreement rate.

To estimate the clinical matching accuracy of both devices, each shade tab from 3 VITA 3D-Master shade guide (26 shades per each shade guide system) was positioned in the dentiform model inside a black box (30.0 × 15.0 × 14.0 cm) to mimic the oral cavity (Fig. 3), and then the central area of each shade tab was measured 3 times with an interval of 1 hour20 for each device (234 measurements per each device). The 9 readings per each 3D-Master shade were obtained and average L^*,a^*, and b^* values were calculated. The ΔE₀₀ color difference value between natural tooth and corresponding 3-D Master shade tab was calculated for each device. The calculated values were then evaluated based on the 50:50% perceptibility threshold (0.80 – 1.30 ΔE₀₀ units) and 50:50% acceptability threshold (1.80 – 2.25 ΔE₀₀ units) from the previous studies.23, 24, 25, 26

Fig. 3
Each shade tab to be measured was positioned in the dentiform model inside a black box.

• Click for larger image
• Download as PowerPoint slide

In vitro matching agreement rates within devices and between devices were evaluated by counting the identical shade measurements per each shade tab irrespective of whether the indicated shade by the device selected the actual shade tab. In vitro matching accuracy was determined by calculating the percentage of correct shade matches to 3D-Master shade tab. All statistical analyses were performed by using a software (IBM SPSS Statistics for Windows, v23.0, IBM Corp., Chicago, IL, USA) (α = .05).

A total of 540 measurements for the in vivo evaluation (270 measurements per each device) and 468 measurements for the in vitro evaluation (234 measurements per each device) were made in this2018-06-14 study. The mean L^*,a^*, and b^* values (SD) for both devices in the in vivo models shown in Table 2. Intra-device ICCs for both devices and inter-device ICCs based on L^*,a^*, and b^* measurements of natural teeth are shown in Table 2. Intra-device ICCs for both devices (0.913 – 0.993) and inter-device ICCs (0.897–0.994) were very high and b^* values exhibited the highest repeatability for intra- and inter-device assessment. Although a high level of measurement repeatability among L^*,a^*, and b^* color components was observed between and within devices (P < .001), the CIEDE2000 color difference between the devices was 2.28 (1.61) ΔE₀₀ units for in-vivo readings, which might be an unacceptable color match based on the previous studies.23, 24, 25, 26 ANOVA revealed that there were no significant differences among L^*,a^*, and b^* color parameters within each device, while paired t-test performed on these values showed significant differences between devices for all color coordinates (Table 3). The matching agreement rates of devices in matching teeth to 3D-Master shade guide systems are given in Table 4. The matching agreement rates within devices were higher than the agreement rate between devices.

Table 2
CIE L^*,a^*, and b^* values and intraclass correlation coefficients (ICCs) indicating device repeatability and interdevice reliability in measuring tooth color (n = 270)

• Click for larger image
• Click for full table
• Download as Excel file

Table 3
Results of ANOVA test within devices and paired t-test between devices indicating in vivo measurement differences in CIE L^*,a^*, and b^* color parameters

• Click for larger image
• Click for full table
• Download as Excel file

Table 4
In vivo and in vitro matching agreement rates of shade measurements within devices and between devices

• Click for larger image
• Click for full table
• Download as Excel file

The mean L^*,a^*, and b^* values (SD) for each shade tab in the 3D-Master shade were shown in Table 5. The color difference ΔE₀₀ values by shade tabs between devices were 2.26 (0.43), which would be an unacceptable color match to the human eye based on the previous studies.23, 24, 25, 26

Table 5
Means (SD) of CIE L^*,a^*, and b^* for each shade tab in the 3D-Master shade guide system (n = 9)

• Click for larger image
• Click for full table
• Download as Excel file

As an estimation of in vivo matching accuracy, the color difference ΔE₀₀ values between recorded shade given by the device and matching shade tab were calculated; 3.05 (1.19) for device 1 and 2.86 (1.02) for device 2. Those values were beyond the 50%:50% acceptability threshold based on the previous studies.23, 24, 25, 26

In vitro matching agreement rates of shade measurements within devices and between devices were presented in Table 4. The in vitro matching agreement rates within devices were higher than in vivo agreement rates, while poor agreement rate between devices were noted especially for the in vitro setting. With regard to in vitro matching accuracy, 20.94% for device 1 and 12.82% for device 2 were found.

Based on the results of this study, 2 identical intraoral spectrophotometers (VITA Easyshade V) with different serial numbers exhibited high intra-device repeatability (ICC > 0.9), while the noticeably different tooth color (P < .001) could be measured between the devices. The color differences between recorded tooth shade given by the device and the matching 3D-Master shade guide were beyond the range of 50%:50% perceptibility threshold (ΔE₀₀ > 1.3) for each device. Thus, the null hypothesis was rejected.

Teeth are usually not one color throughout, but a variety of optical characteristics should be understood before the color interpretation.27 Thus, clinical shade identification with specific shade guide system would be difficult. In this study, in vitro matching agreement rates of shade measurements within devices were higher than in vivo matching agreement rates for both devices. A previous study reported that VITA 3D-Master shade guide system included more adequate color range of natural tooth than VITA classical system28 and that the ceramic restoration fabricated using VITA 3D-Master system revealed smaller color difference with the natural tooth than that fabricated using VITA classical system.29 Therefore, the measured shade in 3D-Master shade system was used in this study.

In this study, 2 identical devices showed high intra-device repeatability based on ICCs (0.91 – 0.99) and ANOVA testing (P > .307), even higher ICCs than those of the previous studies (ICCs: 0.68 – 0.93 in Lagouvardos et al.'s study8 and 0.80 – 0.99 in Lehmann et al.'s study10) in measuring tooth color. In terms of inter-device reliability, some previous in vivo studies reported that inter-device measurement reliability was lower than intra-device measurement repeatability,8, 10, 11 while Lagouvardos et al.'s8in vivo study showed that inter-device matching reliability (ICCs: 0.64 – 0.87) was higher than inter-device measurement reliability (ICCs: 0.40 – 0.49). Although higher ICCs (0.90 – 0.99) were calculated in this study than those of the previous studies,8, 10, 11 paired t-test revealed significantly different color values (P < .001) between devices were noted, suggesting that color data cannot be reproduced in a predictable way even with the identical device.

In the present study, the inter-device color difference values in measuring tooth color were 2.28 (1.61) ΔE₀₀ units, which might be an unacceptable color match to the human eye, while those in Lehmann et al.'s study were 5.04 – 14.63 ΔE^*_ab units. The study used three devices of different specifications (operation mode, light source, measurement region, and spectral resolution), and thus large color differences were identified although color difference was quantified using CIELab formula in Lehmann et al.'s study. Recent in vitro study12 investigated the inter-device reliability of 8 identical electronic color measuring devices. They used VITA Easyshade Advance, the fourth generation of VITA Easyshade, and colors were measured on the ceramic samples. Intra-device repeatability of all color parameters (ICCs > 0.99) were slightly higher than those of the present study and the color differences among devices by 3D-Master shade tabs among 8 devices were 0.62 – 1.67 ΔE^*_ab units.

One previous study7 investigated in vitro matching accuracy to shade guide systems using an intraoral shade-matching device, ShadeScan, which is a combination of a colorimeter and digital imaging. The matching accuracy to 3D-Master shade system was 54.2%, which was a higher percentage value than that measured in this study with VITA Easyshade devices (12.82 – 20.94%). The operating system and any manufacturing variations of the device might influence the accuracy performance of the electronic shade selection devices. In this study, to identify the device accuracy in measuring tooth color, the CIEDE2000 color difference between recorded shade given by the device and matching shade tab were calculated. Since the obtained values (2.86 – 3.05 ΔE₀₀ units) could be considered to be unacceptable, 23, 24, 25, 26 unreliable color matching of the device tested in this study would be expected. Several studied investigated the accuracy of color-matching devices. One study measured color deviations of the intraoral devices from the spectrophotometric reference system.9 Zenthöfer et al. used spectroradiometric data as reference values to calculate the color difference15 and another study compared the devices with reference values of the shade tab from SpectroShade's database.13

Several factors, such as an operating mode of the device,9, 13 training and experience of the examiner,30 tooth position,30 matching shade guide,31 and illumination11 can affect the reliability and accuracy of the intraoral color measuring devices. Different training methods affected the color data and more consistent measurements were noted for canines compared to central incisors in Hassel et al.'s study.30 The most consistent shade match between spectrophotometric shade selection and visual shade selection were noted with VITA classical shade guide system.31 The illumination effect on the color measurement would be more pronounced with Spectroshade than with Easyshade in Sarafianou et al.'s study.11 In addition, the measurement area of a spectrophotometer could influence the deviation of color readings.13 In this study, minor variations in device positioning and the convexity of the external tooth surface and the shade tab could affect light interactions with materials causing discrepancies in measurement consistency even between identical devices. In addition, poor matching accuracy for in vitro examination with 3D-Master shade guide system may be attributed to possible variations in color parameters among different 3D-Master shade guides.

The limitation of this study was that only 2 devices of a single type were tested. In addition, the number of measurements in clinical setting (3 per tooth) might affect the device repeatability and accuracy.

Although two Easyshade V devices with different serial numbers show high repeatability of CIE L^*,a^*, and b^* measurements, they could provide different color values and shades for the same tooth due to spectrophotometric technical sensitivity. Thus, unreliable color determination might be expected between identical devices.