|Year : 2023 | Volume
| Issue : 1 | Page : 20-27
Accuracy and reproducibility of laser-scanned digital models compared to plaster models – An in vitro study
Mandeep Singh1, Achint Juneja2, Divya Shetty2, Payal Sharma2, Monis Raza2, Shubhangi Jain2
1 Department of Orthodontics, I.T.S Dental College, Muradnagar, Uttar Pradesh, India
2 Department of Orthodontics, IDST Dental College, Ghaziabad, Uttar Pradesh, India
|Date of Submission||27-Sep-2021|
|Date of Decision||25-Feb-2022|
|Date of Acceptance||29-Mar-2022|
|Date of Web Publication||17-Feb-2023|
Department of Orthodontics, I.T.S Dental College, Muradnagar, Ghaziabad - 201 206, Uttar Pradesh
Source of Support: None, Conflict of Interest: None
Purpose: The present study aimed to assess the accuracy and reproducibility of laser-scanned digital models compared to plaster models. Materials and Methods: A total of 50 plaster models were scanned using the 3M ESPE Lava scan ST scanner to construct digital models. Various measurements, encompassing intra-arch and inter-arch from scans and plaster models, were estimated. The plaster models were measured using digital calipers and digital models using MiniMagics 3.0 software. Descriptive statistics were calculated, and the mean difference of the parameters between the two groups was tested by paired t-test. Results: No significant difference was detected between the digital and manual measurements in the majority of the parameters, and the values for the parameters showed significant differences that were not clinically applicable. The digital method showed good reproducibility for all the measurements (r > 0.7). Overbite showed a marked variation between the manual and digital measurements due to overlapping tooth structures in digital models with respect to ideal occlusion. Bolton's analysis did not show a significant difference between the manual and digital methods. Conclusions: Digital models could be compared to the gold standard plaster models.
Keywords: Diagnosis, laser scanning, orthodontics
|How to cite this article:|
Singh M, Juneja A, Shetty D, Sharma P, Raza M, Jain S. Accuracy and reproducibility of laser-scanned digital models compared to plaster models – An in vitro study. Indian J Dent Sci 2023;15:20-7
|How to cite this URL:|
Singh M, Juneja A, Shetty D, Sharma P, Raza M, Jain S. Accuracy and reproducibility of laser-scanned digital models compared to plaster models – An in vitro study. Indian J Dent Sci [serial online] 2023 [cited 2023 Mar 26];15:20-7. Available from: http://www.ijds.in/text.asp?2023/15/1/20/369889
| Introduction|| |
Plaster models have been viewed as the gold standard for three-dimensional (3D) verification of oral structures. Orthodontic models can be used to assess symmetry, monitor treatment, and complement the written record, arch form, the severity of the curves of Spee and Wilson, and axial inclinations. They can also be used to perform analyses, such as Peck and Peck, Bolton, and tooth size-arch length discrepancy. With newer technological advancements, computer-based models are gaining popularity and acceptance in the orthodontic community. 3D scanning of the dental arch was initially used for computer-aided design and computer-aided manufacturing technology in dental restorations about 30 years ago. In 1973, the notion of intraoral scanning for dental purposes was introduced. Digital technology has been a feature of orthodontic operations since 1990.
Companies have recently developed scanning systems that use direct plaster scans and direct alginate impression scans to create digital models. While stone models offer a cost-effective fabrication, the disadvantages relate to the physical nature of the model itself. They are prone to breakage and damage, which is the disadvantage for the storage of orthodontic models. Furthermore, the stone models do not allow for easy sharing of information and assessment by other professionals based on distance.,, Rapidly gaining popularity of digital model scanners can be attributed to their cost-effectiveness, space-saving nature, and ease of maintenance compared to physical models. The scanners also provide convenient access to study the models. However, the only drawback of digital models is that they require sophisticated hardware/software and training.
Bootvong et al., Zilberman et al., and Quimby et al. compared plaster and digital models and showed that both methods are effective and can be reproduced when measuring tooth size and dental arch widths. Quimby et al. added that features, such as convenient storage and short duration required for measuring with the digital system, might render this method attractive for orthodontists. However, Garino and Garino found statistically significant differences between measurements obtained using digital and plaster models, thereby disagreeing with the observations that both digital and manual measurements are equally reliable.,
As a result, the goal of this study was to evaluate the accuracy and reproducibility of laser-scanned digital models to plaster models. The goal was to measure tooth sizes, occlusal correlations, and arch dimensions on plaster models; measure tooth sizes, occlusal correlations, and arch dimensions on laser-scanned digital models; and assess the accuracy and reliability of digital models with respect to various parameters on laser-scanned digital models.
| Materials and Methods|| |
A total of 50 pretreatment study models were selected from the archives of the Department of Orthodontics, I. T. S-CDSR, Muradnagar, Ghaziabad, India. The statement is about the selection/inclusion criteria based on which the study models were selected. The exclusion criteria were as follows: supernumerary teeth, carious, heavily restored or teeth with altered anatomy, cleft palate and study models with voids, and fractured teeth or any other damage.
Scanning was carried out using a 3M ESPE Lava scan ST scanner (3M, USA) to construct digital models. The measurements taken on the plaster and digital models were categorized as Groups 1 and 2, respectively.
Digital study models
One laser projector and one adjustable table on which the cast is situated make up the 3M ESPE Lava scan ST scanner. The model is put in the scanner and moved through various orientations while the high-resolution camera records several photographs of the laser-projected points onto the model surface. The STL format is used by the software to save the images of the scanned models. The scanned photos are then transferred into Lava Design software, which is installed on a desktop computer. MiniMagics 3.0 (Materialise, Leuven, Belgium), which contains a magnifying capability that can zoom in on areas of the digital image for easy landmark recognition, was used to import the STL format 3D images. The measurements were taken with a 0.01-mm precision.
Plaster study models
Manual measurements were carried out on the plaster models using digital Vernier calipers. The measurement precision of the Vernier calipers was taken up to 0.01 mm.
The individual tooth measurements of all teeth mesial to the second molar were taken. The crown height of all teeth was measured from the incisal edge to the midpoint of the cervical margin on the facial aspect, and the mesiodistal width of all the teeth was measured as the significant crown height from the midpoint of the cervical margin to the bucco-occlusal tip [Figure 1].
|Figure 1: Measurement of mesiodistal width and crown height on the digital model|
Click here to view
The intra-arch measurements taken were as follows:
- Upper and lower intermolar width (IMW) was measured from the tip of the mesiobuccal cusp of the right first molar to the tip of the mesiobuccal cusp of the left first molar [Figure 2]
- Upper and lower intercanine width (ICW) was measured from the cusp tip of the right canine to the cusp tip of the left canine
- Upper and lower interpremolar width (IPMW) was measured from the buccal cusp tip of the right first premolar to the buccal cusp tip of the left first premolar [Figure 3]
- Little's irregularity index (LII) [Figure 4].
|Figure 2: Measurement of intercanine, interpremolar, and intermolar width on digital model|
Click here to view
|Figure 4: Measurement of Little's irregularity index on the digital study model|
Click here to view
Upper and lower arch perimeters were calculated by dividing each arch into four segments as follows:
- Segment 1 was measured as the distance between the mesial contact point of the central incisor and the distal contact point of the canine on the right side
- Segment 2 was measured as the distance between the mesial contact point of the first premolar to the mesial contact point of the first molar on the right side
- Segment 3 was measured as the distance between the mesial contact point of the central incisor and the distal contact point of the canine on the left side
- Segment 4 was measured as the distance between the mesial contact point of the first premolar and the mesial contact point of the first molar on the left side.
The arch perimeter was obtained by adding the values obtained from the four segments. The inter-arch measurements included the anterior and overall Bolton's ratio, overjet, overbite, and centerline discrepancy measured as the horizontal distance between the upper and the lower centerline [Figure 5].
|Figure 5: Measurements taken using a digital Vernier caliper on plaster models|
Click here to view
All measurements were taken by a single operator to eliminate interoperator variability.
The data was entered into MS Excel and analyzed using Statistical Package for the Social Sciences software version 16 (IBM Corp., Armonk, NY, USA) and MedCalc version 17.9.2 (MedCalc Software Ltd, Belgium). The descriptive statistics of all parameters, including mean, median, standard deviation (SD), and 95% confidence interval, were calculated. The normality of data was estimated using the Shapiro–Wilk test. The mean difference of the parameters between the two groups was tested by paired t-test. The mean difference was significant when P < 0.05, highly significant when P < 0.001, and not significant when P > 0.05.
| Results|| |
The intra-examiner reliability for both methods was high: 0.728–0.985 for the plaster and digital models. IPMW for the lower arch presented the highest intra-examiner reliability (0.985), while the overjet presented the lowest value (0.728). The variability in the differences was reflected in the correlation coefficient, which was satisfactory for all measurements (r > 0.7) [Table 1].
|Table 1: Reliability of measurements taken over the plaster models and digital models|
Click here to view
[Table 2] summarizes the descriptive statistics of mean ± SD. The deviation of mesiodistal width of maxillary teeth in Groups 1 (plaster models) and 2 (laser-scanned models) was calculated. Paired t-test was used to compare the means of the two groups. The tooth numbers 13, 15, 21, 22, and 25 showed a statistically significant difference (P = 0.001, 0.005, < 0.001, < 0.001, and <0.001, respectively), whereas the tooth numbers 11, 12, 14, 16, 23, 24, and 26 did not differ significantly.
|Table 2: Descriptive statistics and paired t-test for the comparison of mesiodistal width of Group 1 and Group 2 for the maxillary teeth|
Click here to view
[Table 3] shows the descriptive statistics with mean ± SD. The deviation of mesiodistal width of mandibular teeth in Groups 1 and 2 was calculated. Paired t-test was used to compare the means of the two groups. When comparing the means, the tooth numbers 31, 32, 36, 43, and 45 showed a statistically significant difference (P = 0.026, 0.013, 0.040, < 0.001, and 0.027, respectively).
|Table 3: Descriptive statistics and paired t-test for the comparison of mesiodistal width of Group 1 and Group 2 for the mandibular teeth|
Click here to view
In [Table 4], the descriptive statistics are presented as mean ± SD. The deviation of crown height of maxillary teeth was calculated in Groups 1 and 2. Paired t-test was used to compare the means between the groups. When comparing the means, the tooth numbers 11, 12, 13, 16, and 22 showed a statistically significant difference (P = 0.008, 0.026, 0.001, <0.001, and 0.004, respectively).
|Table 4: Descriptive statistics and paired t-test for the comparison of crown heights of Group 1 and Group 2 for the maxillary teeth|
Click here to view
In [Table 5], the descriptive statistics are presented as mean ± SD. The deviation of crown height of the mandibular teeth was calculated in Groups 1 and 2. Paired t-test was used to compare the means between the groups. When comparing the means, the tooth numbers 31, 33, 35, 43, and 44 showed a statistically significant difference (P = 0.009, 0.040, 0.002, 0.037, and 0.012, respectively), whereas the tooth numbers 32, 34, 36, 41, 42, 45, and 46 did not show a statistically significant difference.
|Table 5: Descriptive statistics and paired t-test for the comparison of crown heights of Group 1 and Group 2 for the mandibular teeth|
Click here to view
In [Table 6], the descriptive statistics are presented as mean ± SD. The deviation of intra-arch perimeters of maxillary and mandibular teeth in Groups 1 and 2 was calculated. Paired t-test was used to compare the means of intra-arch perimeters of maxillary teeth between the two groups. ICW showed a statistically significant difference (P = 0.022), whereas the IPMW, IMW, and arch perimeter did not differ significantly.
|Table 6: Descriptive statistics and paired t-test for the comparison of intra arch parameters of Group 1 and Group 2 for the maxillary teeth|
Click here to view
Paired t-test was used to compare the means of intra-arch perimeters of mandibular teeth between the two groups. Consequently, IPMW showed a statistically significant difference (P = 0.021), whereas the ICW, IMW, and arch perimeter did not show any statistically significant difference.
[Table 7] summarizes the descriptive statistics as mean ± SD. The deviation of inter-arch perimeters was calculated in Groups 1 and 2. Paired t-test was used to compare the means of inter-arch perimeters between the two groups. Consequently, overbite showed a statistically significant difference with P < 0.001, whereas the overjet and centerline discrepancy did not differ significantly.
|Table 7: Descriptive statistics and paired t-test for the comparison of inter-arch parameters of Group 1 and Group 2|
Click here to view
| Discussion|| |
Easy access, minimal physical space requirements, and the ability to exchange information with other professionals over the Internet are all advantages of digital record storage. Orthodontists can now digitally assess intra-arch and inter-arch correlations thanks to new breakthroughs in 3D dental and orthodontic software. Transverse correlations can also be effectively analyzed when 3D models are viewed in occlusion with various viewpoints on the screen. Digital castings can also be used for “virtual treatment” and “virtual setup.”
The purpose of this study was to evaluate the accuracy and reproducibility of laser-scanned digital models to plaster models. On plaster models and laser-scanned models, readings were made to quantify tooth sizes and analyze occlusal relations and arch dimensions.
The intra-examiner reliability of the manual and digital methods was evaluated to find the error in the methods by repeating the measurements on ten randomly selected models. The variability in the differences was reflected in the correlation coefficients. Both methods demonstrated that the average intra-examiner reliability was 0.728–0.985 for the plaster and digital models. The IPMW for the lower arch presented the highest intra-examiner reliability (0.985), while the overjet presented the lowest value of 0.728.
In the current study, the comparison of the mesiodistal widths revealed that values of five teeth were significantly different in the two groups in the maxillary arch, whereas in the mandibular arch, values of four teeth differed significantly. Although these readings are statistically different, ranging from −0.0634 ± 0.1957 to −0.4190 ± 0.4230, but they are not of sufficient magnitude for them to be considered clinically relevant.
Nonetheless, our findings are consistent with those of Abizadeh et al., who found significant differences between the two approaches, but not clinically important, in their investigation. For 8/16 characteristics, the plaster measurements were repeatable. In a comparable study, Quimby et al. discovered a difference of 0.5 mm, which was not clinically significant.
Abizadeh et al. and Quimby et al. argued that inaccuracies with the digital technology could also be partly attributed to difficulty locating the highest mesiodistal width of the tooth to assess the arch length. Furthermore, the contact points using digitized systems were less well defined, which was exacerbated by crowding. Despite the rotation functions and magnification in the software, pointing accurate location is difficult. The operator's expertise with using a digital model, according to Quimby et al., could make a difference when locating landmarks. Moreover, in digitally scanned models, a 3D image is viewed on the screen, which may lead to inaccuracies in locating landmarks.
Similar findings were observed by Redlich et al. using cross-section planes for measuring tooth width. The study stated that linear-digital measurements of tooth width were significantly smaller in the digital model group than the caliper measurements. However, because this small discrepancy (0.18–0.28 mm) has no clinical significance, computerized models could be used in clinical settings.
Smith et al., in contrast to various studies, compared the accuracy and precision of plaster and digital measurements to a common gold standard, a dental typodont. They found a statistically significant deviation of the mean from zero (P < 0.001). The estimated difference in accuracy between the two methods was 0.017 mm, suggesting that digital measurements were more precise and accurate than those carried out physically on a stone model.
The present study assessed the crown height compared between the two groups. A statistically significant difference was found in five teeth in both the maxillary and mandibular teeth when comparing the two groups. The range of mean difference was from −0.0665 ± 0.1798 to −0.2936 ± 0.6905, which was not clinically significant.
In comparison to traditional dental casts, Kusnoto and Evans discovered that digital models yield precise height and width measurements. These findings differed from those of Camardella et al., who found that the differences between the plaster and digital models, as well as the digital measuring tool used, were of low clinical significance. This could be attributed to structural differences between the plaster and digital models and the digital measuring tool used (Ortho Analyzer software). This may affect measuring precision, particularly on teeth having a buccal inclination. The study concluded that the crown height of the teeth could be reliably measured using digital calipers.
The present study also compared the intra-arch parameters, such as ICW, IPMW, IMW, and arch perimeter, between the two groups in the maxillary and mandibular teeth. The comparison of the intra-arch parameters revealed a significant difference in ICW of the maxillary teeth (0.2494 ± 0.7468) and IPMW of the mandibular teeth (−0.1544 ± 0.4563). This finding was in concordance with that of the study by Quimby et al. and Watanabe-Kanno et al. However, no significant difference was detected in IMW and arch perimeter in either of the groups in maxillary and mandibular teeth. These results were in agreement with the findings of the study by Sousa et al., who found that linear measurements of the arch width on digital models with the 3D shape scanner were accurate and reproducible. The digital measurements tended to produce smaller IMW values than manual measurements.
Smith et al., in contrast to our findings, reported a substantial difference in arch length values when comparing the two approaches. On the actual model, 0.01” steel ligature wire was used. The curve was standardized to be the best-fit curve linking all of the teeth, including the first molar's contact sites. The application of a wire to the edges of the teeth on a plaster cast and marking it for accurate measuring of the wire makes estimating the catenary arch lengths on plaster models difficult. These physical limitations are no longer a worry with digital measures, thanks to several key achievements. Trapezoidal measurements were more trustworthy than catenary approaches, regardless of measuring method.
The main goal of comprehensive orthodontic treatment is to obtain optimal final occlusion, overbite, and overjet. The variations in tooth size in the maxillary and mandibular arch are crucial for achieving this aim; yet, in the presence of Bolton's discrepancy, optimal occlusion is unachievable. As a result, inter-arch factors such as overjet, overbite, and Bolton's analysis are critical for orthodontic diagnosis evaluation.
The current study showed a significant difference in the overbite measurements (0.4318 ± 0.4384) when comparing the two groups, whereas overjet, centerline discrepancy, Bolton's ratio, and LII did not differ significantly. These findings were consistent with those of Santoro et al., who found a statistically significant difference in overbite measurements (0.49 mm) between the two groups, with the digital measurements being smaller than the manual measurements. This discrepancy can be explained by the fact that digital tooth sizes were consistently smaller than those determined by plaster measurements. If the teeth are smaller, a small overbite (in millimeters [mm]) must be expected. Measurements in mm, on the other hand, have no effect on overbite expressed as a percentage. The difference between the two groups in this study could be due to the inaccurate landmark identification in the software, especially in the areas with overlapping maxillary teeth, which is considered a major disadvantage.
The present study did not find a significant difference in Bolton's anterior and the overall ratio with a mean difference ranging from −0.3073 ± 1.2663 to 0.1371 ± 0.7243, respectively. This was in accordance with the result of Kim and Lagarvare, wherein the accuracy of Bolton's analysis of digital models scanned with the Ortho Insight 3D laser system was compared to that of cone-beam computed tomography images and traditional plaster models (gold standard). The study found a mean difference of 0.59 ± 0.520 and 0.41 ± 0.305 for Bolton's anterior and overall ratio, respectively, which suggested that the discrepancy in the mesiodistal tooth width measurement between physical and scanned digital models did not exhibit any clinically relevant differences in the anterior or overall Bolton's ratios.
The mean difference of 0.47 mm between Groups 1 and 2 did not vary significantly for centerline discrepancy. Similar results are presented by Reuschl et al.
Furthermore, during the comparison of the two groups, no significant differences in LII were found. This result was similar to that of Massoud et al. and Kumar et al. According to the findings, the designed 3D system and software were effective for virtual dental cast reconstruction. For measurements, the program was sufficiently precise and repeatable. Kim and Lagarvare on the other hand, found that results for LII were significantly different for measures that differed by more than 2 mm. The LII aberrations were found to be within clinically acceptable limits, according to the study.
In conclusion, we discovered significant disparities in a number of characteristics. These changes were not clinically significant, and the disparities in the results compared to previous research were attributable to differences in methodology, scanner type, and digitizing software. Furthermore, interoperator variability is important in detecting landmarks on both plaster and digital models. Despite the fact that plaster models are the gold standard, digital models can be used for orthodontic diagnosis and planning. This was demonstrated by the current investigation, which indicated no significant differences in model analysis between the two methodologies. Digital models are useful in congested dental arches because they allow for 3D magnification and rotation.
| Conclusions|| |
The following conclusion can be drawn from this study:
- No significant difference is detected in the majority of the parameters between the digital and manual measurements, while those that showed statistically significant differences had small values (<2 mm) and were not clinically appreciable
- The digital method showed good reproducibility for all the measurements (r > 0.7)
- Overbite showed a large variation between the manual and digital measurements due to overlapping tooth structures in digital models in ideal occlusion
- Bolton's analysis did not show a significant difference between the manual and digital methods. Therefore, digital models can be compared to the gold standard plaster models
- Further studies are essential to evaluate the efficacy of direct intraoral scanners used in the oral cavity. These scanners might have the advantage of scanning proper bite registration and eliminating the overlapping of tooth structures.
Ethical Committee Approval Number was ITSCDSR/L/2018/149.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Abizadeh N, Moles DR, O'Neill J, Noar JH. Digital versus plaster study models: How accurate and reproducible are they? J Orthod 2012;39:151-9.
Correia GD, Habib FA, Vogel CJ. Tooth-size discrepancy: A comparison between manual and digital methods. Dental Press J Orthod 2014;19:107-13.
Asquith J, Gillgrass T, Mossey P. Three-dimensional imaging of orthodontic models: A pilot study. Eur J Orthod 2007;29:517-22.
Moreira DD, Gribel BF, Torres GD, Vasconcelos KD, Freitas DQ, Ambrosano GM. Reliability of measurements on virtual models obtained from scanning of impressions and conventional plaster models. Braz J Oral Sci 2014;13:297-302.
Martin CB, Chalmers EV, McIntyre GT, Cochrane H, Mossey PA. Orthodontic scanners: What's available? J Orthod 2015;42:136-43.
Bootvong K, Liu Z, McGrath C, Hägg U, Wong RW, Bendeus M, et al.
Virtual model analysis as an alternative approach to plaster model analysis: Reliability and validity. Eur J Orthod 2010;32:589-95.
Zilberman O, Huggare JA, Parikakis KA. Evaluation of the validity of tooth size and arch width measurements using conventional and three-dimensional virtual orthodontic models. Angle Orthod 2003;73:301-6.
Quimby ML, Vig KW, Rashid RG, Firestone AR. The accuracy and reliability of measurements made on computer-based digital models. Angle Orthod 2004;74:298-303.
Garino F, Garino GB. Comparison of dental arch measurements between stone and digital casts. World J Orthod 2002;3:1-5.
Sousa MV, Vasconcelos EC, Janson G, Garib D, Pinzan A. Accuracy and reproducibility of 3-dimensional digital model measurements. Am J Orthod Dentofacial Orthop 2012;142:269-73.
Redlich M, Weinstock T, Abed Y, Schneor R, Holdstein Y, Fischer A. A new system for scanning, measuring and analyzing dental casts based on a 3D holographic sensor. Orthod Craniofac Res 2008;11:90-5.
Smith DK, Beaudoin B, Messersmith M, Blume JD. Comparison of digital measurement versus physical model measurement: An analysis of accuracy and precision. Dent Health Curr Res 2017;3:1-4.
Kusnoto B, Evans CA. Reliability of a 3D surface laser scanner for orthodontic applications. Am J Orthod Dentofacial Orthop 2002;122:342-8.
Kim J, Lagravére MO. Accuracy of Bolton analysis measured in laser scanned digital models compared with plaster models (gold standard) and cone-beam computer tomography images. Korean J Orthod 2016;46:13-9.
Watanabe-Kanno GA, Abrão J, Miasiro Junior H, Sánchez-Ayala A, Lagravère MO. Reproducibility, reliability and validity of measurements obtained from Cecile3 digital models. Braz Oral Res 2009;23:288-95.
Horton HM, Miller JR, Gaillard PR, Larson BE. Technique comparison for efficient orthodontic tooth measurements using digital models. Angle Orthod 2010;80:254-61.
Lopatiene K, Dumbravaite A. Relationship between tooth size discrepancies and malocclusion. Stomatologija 2009;11:119-24.
Santoro M, Galkin S, Teredesai M, Nicolay OF, Cangialosi TJ. Comparison of measurements made on digital and plaster models. Am J Orthod Dentofacial Orthop 2003;124:101-5.
Czarnota J, Hey J, Fuhrmann R. Measurements using orthodontic analysis software on digital models obtained by 3D scans of plaster casts: Intrarater reliability and validity. J Orofac Orthop 2016;77:22-30.
Reuschl RP, Heuer W, Stiesch M, Wenzel D, Dittmer MP. Reliability and validity of measurements on digital study models and plaster models. Eur J Orthod 2016;38:22-6.
Kumar AA, Phillip A, Kumar S, Rawat A, Priya S, Kumaran V. Digital model as an alternative to plaster model in assessment of space analysis. J Pharm Bioallied Sci 2015;7:S465-9.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7]