Dental Arch Relationships on Three-Dimensional Digital Study Models and Conventional Plaster Study Models for Patients with Unilateral Cleft Lip and Palate

Abstract

Objective

To determine if three-dimensional (3D) digital study models could replace plaster study models for the evaluation of dental arch relationships for patients with unilateral cleft lip and palate.

Design

Observational study involving plaster study models from a records archive

Setting

U.K. National Health Service.

Patients, Participants

Thirty sets of study models of 5-year-old patients with unilateral cleft lip and palate were identified and scanned to produce 3D digital study models by ESM Digital Solutions Ltd. (Swords, Co. Dublin, Ireland) using an R250 Orthodontic Study Model Scanner (3Shape A/S, Copenhagen, Denmark).

Interventions

None.

Main Outcome Measure(s)

The plaster and 3D digital study models were scored using the 5-year-olds' and modified Huddart Bodenham indices and analyzed using the Friedman test (p < .05) and two-way ANOVA, respectively. Intra-observer and interobserver reproducibility were calculated from the 5-year-olds' index data using the weighted kappa statistic for both the plaster and 3D digital models.

Results

Intra-observer and interobserver reproducibility were good (0.62 to 0.83 and 0.64 to 0.78, respectively). There were no statistically significant differences between the scores for the 3D digital study models when compared to the plaster study models for either the 5-year-olds' index (p = .12) or for the modified Huddart Bodenham index (p = .506).

Conclusions

Three-dimensional digital models are a valid alternative to traditional plaster study models for the evaluation of dental arch relationships in patients with unilateral cleft lip and palate.

Keywords

3D models cleft lip cleft palate study models

For patients with clefts of the lip and/or palate, models are obtained throughout childhood and into adulthood. Dental arch relationships are assessed using the 5-year-olds' index (Atack et al., 1997), the GOSLON index (Mars et al., 1987), and the modified Huddart Bodenham index (Gray and Mossey, 2005) among others. However, plaster models are fragile and need to be archived for clinical reference, audit, and research. Alternatives to plaster study models include photocopies, photographs, holographic films, and stereophotogrammetric images of study models. In orthodontics, photocopies have been shown to be useful for comparing pre- and posttreatment arch forms and for the investigation of tooth rotations but are less useful for the measurement of arch length (Champagne, 1992). Photographs of plaster study models in patients with clefts are reliable when assessed using the GOSLON and modified Huddart Bodenham indices (Nollet et al., 2004; Ali et al., 2006). Holography is clinically useful (Keating et al., 1984), but the images are expensive, difficult to produce, and cannot be manipulated as easily as plaster models. Stereophotogrammetry was found to be accurate, but the accuracy of measurements in the incisor region may be inferior (Bell et al., 2003). Linear laser scanning of study models has also been demonstrated to be accurate (Lu et al., 2000; Hirogaki et al., 2001; Kusnoto and Evans, 2002) but is not used frequently. The major drawback of these techniques relates to the loss of detail in areas of undercut.

Three-dimensional (3D) digital study models became available in orthodontics around 10 years ago. They can be stored, manipulated, and measured using a standard personal computer. Storage space at less than one megabyte per set of models is negligible when compared with conventional plaster models (Peluso et al., 2004). The 3D digital study models can be copied easily and integrated into a patient's electronic file along with other digital records. Retrieval is fast and efficient, and the models can be viewed at multiple locations simultaneously. However, 3D digital models are associated with drawbacks. They cannot be held and viewed in the same way as plaster models, and time is needed to become familiar with their use. Furthermore, although the digital model is 3D, the image is viewed on screen in only two dimensions.

The 5-year-olds' index was developed to assess dental arch relationships in patients with unilateral cleft lip and palate (UCLP) at 5 years of age (Atack et al., 1997; Clark et al., 2007). As a result, this categorical index provides a surrogate assessment of the outcome of early surgical cleft care. The modified Huddart Bodenham system is an alternative method of objectively measuring relative maxillary arch constriction using a numerical scoring system (Gray and Mossey, 2005). It provides an evaluation of the severity and the outcome of cleft care.

Digital orthodontic study models have become a valid alternative to traditional plaster study models in orthodontic treatment planning (Whetten et al., 2006), and they are potentially of use for cleft lip and palate as well as for measurements performed on neonatal models (Oosterkamp et al., 2006). The objective of this investigation was to determine if 3D digital models can be used to evaluate dental arch relationships in patients with UCLP at 5 years of age. Our null hypothesis was that there is no statistically significant difference between measurements made using plaster study models compared with digital study models.

Materials and Methods

The principles outlined in the Declaration of Helsinki were followed throughout (http://www.wma.net/en/30publications/10policies/b3/index.html). Thirty sets of plaster study models from UCLP patients aged 5 years of age were selected from the CLEFTSiS (Scottish National Managed Clinical Network for patients with clefts of the lip and/or palate) record archive at Perth Royal Infirmary, Scotland. Records were selected on the basis of having no plaster defects. The study models were scanned by ESM Digital Solutions Ltd. (Swords, Co. Dublin, Ireland) using an R250 Orthodontic Study Model Scanner (3Shape A/S, Copenhagen, Denmark), with the plaster models positioned on the articulating table while a laser projected a series of points in a line on the surface of the model. The model was tilted, rotated, and translated in the scanner in darkness so all that was visible to the two charge coupled device cameras positioned at 30 degrees to each other was a red line on the surface of the model. Many points from one pass of the model through the scanner may have overlapped other points, with the lowest resolution being 0.2 mm. This was the distance between two adjacent points in the point cloud produced from the x, y, and z coordinates.

The point cloud was then converted into a series of triangles constructed from three adjacent points (Fig. 1). Algorithms produced curvature for the triangular surfaces to reproduce the naturally occurring surface. During postprocessing, 30% decimation was used to remove redundant data relating to “flat” surfaces because such data compounded file size and processing time.

Figure 1

Triangulation based on three adjacent points or restoration of curvature by algorithm.

There were three scanning processes involved in generating the 3D digital study models: (1) complete scan of the upper model, (2) complete scan of the lower model, and (3) partial scan of the models when in occlusion—the outer surfaces of the teeth were scanned with the models in occlusion

Three points were then manually selected on the upper model scan, and these were identified on the “in occlusion” partial scan. These and all the other components of the images were then matched to overlap the full digital model of the upper with the partial digital image of the upper model. The same process was performed on the lower model.

The digital models were then imported to the Ortho-Analyzer^™ (3Shape A/S) program for analysis. Three examiners (all specialist orthodontists) scored the plaster and digital models in random order on two separate occasions at least 1 month apart using the 5-year-olds' index and the modified Huddart Bodenham index. The data were imported into a spreadsheet (Microsoft Corporation, Redmond, WA) for analysis.

Statistical Analysis

The statistical significance of any differences in the 5-year-olds' index data was analyzed using the Friedman test, a nonparametric test similar to the parametric repeated measures ANOVA and sometimes called a nonparametric randomized black analysis of variance (p < .05). The modified Huddart Bodenham index data was assessed for statistical significance using a two-way ANOVA including model type (plaster or digital) and examiner as fixed factors. Intra-observer and interobserver reproducibility were calculated from the 5-year-olds' index data using the weighted kappa statistic (Landis and Koch, 1977) for both the plaster and 3D digital models.

Results

The weighted kappa scores (Table 1) demonstrated intraobserver reproducibility to be good or very good (0.62 to 0.83) and interobserver reproducibility to be good (0.64 to 0.78). Tables 2 through 4 demonstrate the variability among the 5-year-olds' index data and the modified Huddart Bodenham data to be similar. There were no statistically significant differences between the scores for the digital study models when compared with the plaster study models for either the 5-year-olds' index (p = .12) (Table 5) or for the modified Huddart Bodenham index (p = 0.506) (Table 6).

Table 1

Reproducibility of 5-Year-Olds' Index Data: Plaster and Digital Models^*

	GM 1 (plaster)	GM1 (digital)	TG1 (plaster)	TG1 (digital)	LD1 (plaster)	LD1 (digital)
GM2 (plaster)	.625		.758		.645
GM2 (digital)		.801		.742		.714
TG2 (plaster)			.827		.700
TG2 (digital)				.813		.769
LD2 (plaster)					.785
LD2 (digital)						.754

GM = Grant McIntyre; TG = Toby Gilgrass; LD = Lorna Dobbyn.

Table 2

Results of Friedman Test for 5-Year-Olds' Index Data

	n	Rank Sum	Mean Rank
GM (plaster)	30	124.0	4.13
TG (plaster)	30	97.0	3.23
LD (plaster)	30	101.5	3.38
GM (3D digital)	30	106.5	3.55
TG (3D digital)	30	102.5	3.42
LD (3D digital)	30	98.5	3.28

Table 3

Results of Modified Huddart Bodenham Index Data—Mean, Standard Error (SE), and Standard Deviation (SD) for Observers

Observer	N	Mean	SE	Pooled SE	SD
GM	60	-6.9	0.58	0.61	4.3
TG	60	-7.0	0.65	0.61	5.0
LD	60	-7.4	0.61	0.61	4.7

Table 4

Results of Modified Huddart Bodenham Index Data—Mean, Standard Error (SE), and Standard Deviation (SD) for Model Types

Model Type	N	Mean	SE	Pooled SE	SD
Plaster	90	-7.3	0.51	0.50	4.8
3D digital	90	-6.9	0.48	0.50	4.6

Table 5

Results of Friedman Test for 5-Year-Olds' Index Data

Parameter	Value
Friedman's statistic	8.75
χ² statistic	8.75
df	5
P	.1197^*

Chi-square approximation, corrected for ties.

Table 6

Results of Modified Huddart Bodenham Index Data—Interaction Between Model Types and Observers: 2-Way ANOVA

Source of Variation	Sum Squares	df	Mean Square	F Statistic	P
Model types	9.8	1	9.8	0.44	.5063
Observers	7.4	2	3.7	0.17	.8453
Residual	3889.0	176	22.1
Total	3906.2	179

Discussion

We found that there were no statistically significant differences between traditional plaster study models and 3D digital models when assessed using the 5-year-olds' and the modified Huddart Bodenham indices.

Our results were in line with the results from the studies that have compared measurements made on plaster and 3D digital models. Mullen et al. (2007), Gracco et al. (2007), Goonewardene et al. (2008), and Nouri et al. (2009) found measurements on 3D dental models to be accurate and reproducible. Other investigations determined that the use of digital models would not result in different diagnostic and treatment planning decisions being made (Rheude et al., 2005; Stevens et al., 2006), whilst orthodontic peer assessment rating scores derived from digital models were found to be valid and reliable (Mayers et al., 2005).

Not all studies have found 3D digital models to be as reliable as traditional plaster models. Okunami et al. (2007) investigated OrthoCAD (version 2.2, www.cadentinc.com) and determined that it was not adequate for scoring all the parameters as required by the American Board of Orthodontists' objective grading scheme because the assessment of buccolingual inclination is not possible with OrthoCAD. They also determined that occlusal relationships are difficult to assess on the 3D digital models. Asquith et al. (2007) found that of 11 parameters measured, random errors were a cause for concern for three, whilst Keating et al. (2008) found a mean difference between measurements made directly on the plaster models and those made on the 3D digital surface models of 0.14 mm, although this was not statistically significant. Sjogren et al. (2010) on the other hand found that angular and linear variables exhibited poor 95% limits of agreement on their sample of 3D digital models. Interestingly, Leifert et al. (2009) found a slight (0.4 mm) but statistically significant difference in the space analysis measurements on the maxillary models in their study comparing plaster and 3D digital models, whilst measurements made on mandibular models were not significantly different.

Nevertheless, Dalstra and Melsen (2009) have shown that virtual measurements performed on digital models display less variability than the corresponding measurements performed with a caliper on the actual models. Therefore it was appropriate to investigate the utility of 3D digital models for patients with clefts given that photographs have been demonstrated by both Ali et al. (2006) and Liao et al. (2009) to be a viable alternative to models for evaluating dental arch relationships in cleft lip and palate. One drawback to using photographs is that overbite can be difficult to assess on photographs of study models (Malik et al., 2009) and therefore 3D digital study models are more desirable.

The models used in this study were consecutively identified from an archive to avoid bias in sample selection. Only those with plaster defects affecting the assessment of dental arch relationships were excluded. The plaster models were assessed by hand articulation and the digital models were produced using a commercially available scanning system. Following a short period of familiarization, each observer was allowed to manipulate both the plaster and digital models as they wished. Although we did not specifically investigate the time required to assess plaster and 3D digital models, the investigators did not find familiarization with 3D digital models to be a problem. Nonetheless, Gracco et al. (2007) and Mullen et al. (2007) noted that 3D digital models were associated with significantly reduced measurement times in comparison with measurements made on plaster models.

We used the 5-year-olds' and modified Huddart Bodenham indices to investigate dental arch relationships because one is categorical and the other is based on continuous data. As a result, any effect of bias resulting from one or the other type of data would become evident when the data were analyzed. Furthermore, the 30 sets of models were assessed 1 month apart to avoid any bias resulting from remembering the models from the initial scoring session. In order to make the results comparable to those from previous investigations, intra- and interobserver reproducibility were calculated from repeat assessments using the 5-year-olds' index. The observers were all calibrated in the use of the 5-year-olds' index and each frequently used the modified Huddart Bodenham index to avoid any bias arising from a lack of familiarity with both indices.

Nevertheless, there were weaknesses in this study. We limited the investigation to study models of subjects with UCLP and only assessed them at 5 years of age. This is because the 5-year-olds' index was developed for UCLP cases, whilst the modified Huddart Bodenham index can be used with all cleft types. However, it is unlikely that there would be any bias arising from this. Still, it would be useful to confirm that 3D digital study models are useful for assessing severity and outcome in older patients with cleft palate and bilateral cleft palate and lip. Furthermore, the development of intra-oral scanning and subsequent 3D digital modeling may offer the potential to eliminate the need for serial impressions during childhood and adolescence, which would be beneficial because compliance for impression taking in patients with clefts is less than 100%.

Many cleft and craniofacial teams are developing electronic care records incorporating two-, three-, and four-dimensional image records (Ali et al., 2006), and corroborative results from investigations of the reliability of 3D study models in cleft care could lead to the routine use of 3D study models in determining the severity and outcomes of cleft care. Indeed in due course, it may be possible to perform a whole-arch scan, avoiding the need for impressions to be taken. This development would be welcomed by patients and orthodontists alike.

Conclusion

Three-dimensional digital models are a valid alternative to traditional plaster study models for the evaluation of dental arch relationships in patients with UCLP.

Footnotes

Acknowledgments

We would like to thank Mark Barry (ESM Digital Solutions, Swords, Co. Dublin, Ireland) for scanning the models and providing the software used in this study, and Dr. Toby Gillgrass and Dr. Lorna Dobbyn (CLEFTSiS) for taking the time to score the models.

References

Ali

S.A.

Mossey

Gillgrass

A study model based photographic method for assessment of surgical treatment outcome in unilateral cleft lip and palate patients. Eur J Orthod. 2006; 28: 366–372.

Asquith

Gillgrass

Mossey

Three-dimensional imaging of orthodontic models: a pilot study. Eur J Orthod. 2007; 29: 517–522.

Atack

Hathorn

Mars

Sandy

Study models of 5 year old children as predictors of surgical outcome in unilateral cleft lip and palate. Eur J Orthod. 1997; 19: 165–170.

Bell

Ayoub

A.F.

Siebert

Assessment of the accuracy of a three-dimensional imaging system for archiving dental study models. J Orthod. 2003; 30: 219–223.

Champagne

Reliability of measurements from photocopies of study models. J Clin Orthod. 1992; 10: 648–650.

Clark

S.A.

Atack

N.E.

Ewings

Hathorn

I.S.

Mercer

N.S.

Early surgical outcomes in 5-year-old patients with repaired unilateral cleft lip and palate. Cleft Palate Craniofac J. 2007; 44: 235–238.

Dalstra

Melsen

From alginate impressions to digital virtual models: accuracy and reproducibility. J Orthod. 2009; 36: 36–41.

Goonewardene

R.W.

Goonewardene

M.S.

Razza

J.M.

Murray

Accuracy and validity of space analysis and irregularity index measurements using digital models. Aust Orthod J. 2008; 24: 83–90.

Gracco

Buranello

Cozzani

Siciliani

Digital and plaster models: a comparison of measurements and times. Prog Orthod. 2007; 8: 252–259.

10.

Gray

Mossey

P.A.

Evaluation of a modified Huddart/Bodenham scoring system for assessment of maxillary arch constriction in unilateral cleft lip and palate subjects. Eur J Orthod. 2005; 27: 507–511.

11.

Hirogaki

Sohmura

Satoh

Takahashi

Takada

Complete 3-D reconstruction of dental cast shape using perceptual grouping. IEEE Trans Med Imaging. 2001; 20: 1093–1101.

12.

Keating

A.P.

Knox

Bibb

Zhurov

A.I.

A comparison of plaster, digital and reconstructed study model accuracy. J Orthod. 2008; 35: 191–201.

13.

Keating

P.J.

Parker

R.A.

Keane

Wright

The holographic storage of study models. Br J Orthod. 1984; 11: 119–125.

14.

Kusnoto

Evans

C.A.

Reliability of a 3D surface laser scanner for orthodontic applications. Am J Orthod Dentofacial Orthop. 2002; 122: 342–348.

15.

Landis

J.R.

Koch

G.G.

The measurement of observer agreement for categorical data. Biometrics. 1977; 33: 159–174.

16.

Leifert

M.F.

Leifert

M.M.

Efstratiadis

S.S.

Cangialosi

T.J.

Comparison of space analysis evaluations with digital models and plaster dental casts. Am J Orthod Dentofacial Orthop. 2009; 136: 16.e1–4.

17.

Liao

Y.F.

Huang

C.S.

Lin

I.F.

Intraoral photographs for rating dental arch relationships in unilateral cleft lip and palate. Cleft Palate Craniofac J. 2009; 46: 415–419.

18.

Wang

Chen

Zhao

The research and development of noncontact 3-D laser dental model measuring and analyzing system. Chin J Dent Res. 2000; 3: 7–14.

19.

Malik

O.H.

Abdi-Oskouei

Mandall

N.A.

An alternative to study model storage. Eur J Orthod. 2009; 31: 156–159.

20.

Mars

Plint

D.A.

Houston

W.J.B.

Bergland

Semb

The Goslon Yardstick: a new system of assessing dental arch relationships in children with unilateral clefts of the lip and palate. Cleft Palate J. 1987; 24: 314–322.

21.

Mayers

Firestone

A.R.

Rashid

Vig

K.W.

Comparison of peer assessment rating (PAR) index scores of plaster and computer-based digital models. Am J Orthod Dentofacial Orthop. 2005; 128: 431–434.

22.

Mullen

S.R.

Martin

C.A.

Ngan

Gladwin

Accuracy of space analysis with emodels and plaster models. Am J Orthod Dentofacial Orthop. 2007; 132: 346–352.

23.

Nollet

PJPM

Katsaros

Van ‘T Hof

M.A.

Bongaarts

C.A.M.

Semb

Shaw

W.C.

Kuijpers-Jagtman

A.M.

Photographs of study casts: an alternative medium for rating dental arch relationships in unilateral cleft lip and palate. Cleft Palate Craniofac J. 2004; 41: 646–650.

24.

Nouri

Massudi

Bagheban

A.A.

Azimi

Fereidooni

The accuracy of a 3-D laser scanner for crown width measurements. Aust Orthod J. 2009; 25: 41–47.

25.

Okunami

T.R.

Kusnoto

BeGole

Evans

C.A.

Sadowsky

Fadavi

Assessing the American Board of Orthodontics objective grading system: digital vs plaster dental casts. Am J Orthod Dentofacial Orthop. 2007; 131: 51–56.

26.

Oosterkamp

B.C.

van der Meer

W.J.

Rutenfrans

Dijkstra

P.U.

Reliability of linear measurements on a virtual bilateral cleft lip and palate model. Cleft Palate Craniofac J. 2006; 43: 519–523.

27.

Peluso

M.J.

Josell

S.D.

Levine

S.W.

Lorei

B.J.

Digital models: an introduction. Semin Orthod. 2004; 10: 226–238.

28.

Rheude

Sadowsky

P.L.

Ferriera

Jacobson

An evaluation of the use of digital study models in orthodontic diagnosis and treatment planning. Angle Orthod. 2005; 75: 292–296.

29.

Sjogren

A.P.

Lindgren

J.E.

Huggare

J.A.

Orthodontic study cast analysis— reproducibility of recordings and agreement between conventional and 3D virtual measurements. J Digit Imaging. 2010; 23: 482–492.

30.

Stevens

D.R.

Flores-Mir

Nebbe

Raboud

D.W.

Heo

Major

P.W.

Validity, reliability, and reproducibility of plaster vs digital study models: comparison of peer assessment rating and Bolton analysis and their constituent measurements. Am J Orthod Dentofacial Orthop. 2006; 129: 794–803.

31.

Whetten

J.L.

Williamson

P.C.

Heo

Varnhagen

Major

P.W.

Variations in orthodontic treatment planning decisions of Class II patients between virtual 3-dimensional models and traditional plaster study models. Am J Orthod Dentofacial Orthop. 2006; 130: 485–491.