European Journal of Orthodontics 26 (2004) 209–215 European Journal of Orthodontics vol. 26 no.

2
 European Orthodontic Society 2004; all rights reserved.

Quantitation of transverse maxillary dimensions using computed

tomography: a methodological and reproducibility study
B. Podesser*, S. Williams**, H.-P. Bantleon* and H. Imhof***
*Universitätsklinik für Zahn, Mund und Kieferheilkunde, Vienna, **Private Practice, Wiener Neustadt,
***Universitätsklinik für Radiodiagnostik, Vienna, Austria

SUMMARY The present investigation was a methodological study of a new method of quantification of
a series of factors in the transverse dimension of the maxilla including the nose, maxillary bones and
dental arches, based on computer tomographic (CT) scanning. The aim was to investigate a series of
parameters thought to be relevant in the differential diagnosis of discrepancies in the morphology of
this area and probably affected by orthodontic appliances.
Based on a standardized CT scanning registration of 10 subjects, a series of points on the scans were
identified and then measured in a special cephalometric computer system (linear and angular values).
The quantitation was repeated by each observer and inter- and intra-observer differences were calculated.
The results demonstrated that virtually all the parameters showed a high degree of reproducibility at
both levels and confirmed the statistical suitability of the method described. The method will be used in
a series of ongoing studies regarding the morphology and treatment of discrepancies of the midface and
therefore supplement the relatively sparse information based on quantitative reports concerning this
important anatomical area.

Introduction defining structures for measurement. In the past, implant

studies have been used to investigate the development
Orthodontic diagnosis is primarily based on a
of the maxillary complex (Björk and Skieller, 1977) or
morphological and quantitative description of structures
the reaction of the structures to orthopaedic forces (Krebs,
in three planes, namely sagittal, vertical and transverse
1958; Skieller, 1964), although for ethical reasons this
(Angle, 1907). In the first two planes, the lateral
type of study cannot now be performed. Timms (1980),
cephalogram provides material that can be quantitated
in a study of rapid maxillary expansion, used computed
as linear or angular variables and, thus, forms the basis of
tomography (CT) for the first time and concluded that
the cephalometric analysis generally used in orthodontic
the method warranted further investigation.
diagnosis. The transverse dimension is naturally of
The aim of the present study was to develop and
equal importance in the formation of the occlusion.
evaluate a simple non-invasive technique for accurate
Anomalies in maxillary transverse dimensions lead to
and reproducible evaluation of maxillary structures
occlusal problems including crossbite and scissors bite
based on a series of frontal CT scans. Using a series
with or without mandibular forced bite. Corrections of
of parameters, the reproducibility of the quantitation
these anomalies are generally considered some of the
method was evaluated at three levels: (1) variation due
most important in orthodontics. The narrow skeletally
to the computer/digitizing process; (2) intra-observer
small maxilla is often related to an increased vertical jaw
variation; (3) inter-observation variation.
relationship and predisposes to open bite and possibly
a Class III skeletal relationship. In the same way that
the diagnosis and treatment of sagittal and vertical
Subjects and method
discrepancies, as well as their differential diagnoses,
include considerations of skeletal structures as well The present study was based on CT registration of
as dento-alveolar structures, these also should be 10 adults (three males, seven females), average age
identified and treated in transverse anomalies. The basic 26 years 3 months (minimum 22 years 1 month; maximum
problem has been to create a radiographic measurement 31 years 6 months). The subjects were chosen irrespective
on which these structures can be measured. of occlusion and facial morphology, although to avoid
The frontal or postero-anterior radiograph, while radiographic/metallic interference, those without dental
occasionally used in the quantitation of cranial morphology restorations were preferred. This study was approved by
(Athanasiou, 1986), does not reveal corresponding infor- the Ethical Committee of the Medical Faculty, University
mation, due to the superimposition of many structures of Vienna, and each subject signed an informed consent
from different planes of space, as well as difficulties in form.
210 B. P O D E S S E R E T A L .

The scans were obtained using a high-resolution bone

algorithm. In the present study, a computed tomogram
(Tomoscan 7000R, Philips, Eindhoven, The Netherlands)
was used. A yoke was mounted in the centre of the
gantry which carried the X-ray tube and detectors. The
X-ray tube was mounted directly opposite the detectors.
The resultant attenuated X-ray beam was transmitted
through the anatomical area and received and measured
by the detectors. A 5.25 inch optical disc was integrated,
enabling the loading of software and recording of the
results, either temporarily or permanently.
The CTs were obtained using a standardized frontal
plane orientation, i.e. with the subject positioned face-
down with the neck hyper-extended and the scanning
plane at 90 degrees to the plane of the hard palate.
To minimize distortion due to poor orientation of the Figure 2 A computer tomographic scanning slice at the canine.
subject, the head was supported with a contoured
pillow. In order to choose appropriate slices, a lateral
overview with the proposed slices was performed with
the first slice slightly anterior to the canine teeth and the
last slice just posterior to the distal surface of the first
permanent molars. The inter-slice distance was set
arbitrarily at 1.5 mm and the slice thickness at 1.5 mm.
The CT scanning process was performed only once for
each subject.
Quantitation was based on two slices. For the molar
slice (Figure 1) the most anterior slice showing the
entire palatal root of the first maxillary molar was
chosen. The canine slice was chosen as the most anterior
slice on which the crown and root could be seen in their
entire length (Figure 2).
Quantitative evaluation of the parameters was
based on the identification and registration of a series
of points (Figures 3 and 4) on standard cephalometric
acetate paper using a light box (i.e. without
magnification). Points 1 and 2 represent the lower limit Figure 3 The parameters for the molar (see text for details).
(left and right, respectively) of the nasal cavity and are
defined from a line as a tangent to the base of the nose,
termed the baseline 1–2: Points 3 and 4 (left and right,
respectively) the most coronal points on the maxillary
alveolar process on the left and right side, respectively:
Points 5 and 6 the lateral limits of the maxillary base
created as the point at which the almost vertical line of
the alveolar process turns to the almost horizontal plane
of the lower border of the zygomatic bone process:
Points 7 and 8 represent the tips of the mesiobuccal cusp
of the maxillary first molars, left and right, respectively:
Points 9 and 10 the apices of the palatal roots of the first
molars, left and right, respectively: Points 11 and 12
(Figure 4) the highest points on the crown on the
maxillary canines, left and right, respectively: Points 13
and 14 (Figure 4) the apices of the maxillary canines, left
and right, respectively: Points 15 and 16 (Figure 4) (for
the canines) the lower limit, left and right, respectively,
Figure 1 A computer tomographic scanning slice at the molar of the nasal cavity and are defined from a line as a
fulfilling the criteria for choice. tangent to the base of the nose (the baseline 15–16):
C T QUA N T I TAT I O N O F M A X I L L A RY D I M E N S I O N S 211

Experimental procedure
The reproducibility of the digitizing process was evalu-
ated by means of double determination of each set of
registration points by the same person (BP) based on
one set of recordings. In order to determine the intra-
observer variation, the entire process, i.e. the selec-
tion of the registration slice, tracing and digitization,
was repeated by a different author (SW) at least 10 days
after the first registration. Inter-observer variation was
determined by comparing the average of the two readings
performed by one person with a new series of tracings
(averages of double determination) by another experi-
enced observer (SW). This process again involved all the
steps of the quantitation, although the radiographic pro-
cedure was not repeated in order to minimize radiation.
In order to investigate variation due to the choice of
slice, the defined parameters were measured in one
patient using all the slices fulfilling the named criteria.

Figure 4 The parameters for the canine (see text for details).
Statistical analysis
The statistical analysis was performed at the following
stages:
Points 17 and 18 (Figure 3) the lateral limits of the nasal
cavity: Point 17 is the lateral point on the lateral wall of (a) Differences between the two readings for both
the left nasal cavity using a line perpendicular to the intra- and inter-observer variation were tested using a
baseline (1–2). Point 18 is constructed on the lateral wall paired “Student’s” t-test. The results for the digitization
of the right nasal cavity from point 17 using a line process and the intra- and inter-observer analyses are
perpendicular to 1–2. Digitization of the points was then shown in Tables 1–3.
performed using a Scriptel© (Scriptel Corp., USA) (b) The correlation between the first and second read-
digitizer in conjunction with a user-defined computer ings was calculated using Spearman’s correlation analysis
cephalometric analysis program (Cephaloplot©, Randers and represented with the r value. The individual differ-
Computers, Randers, Denmark) calculating the linear ences between the first and second determinations are
and angular variables. All digitization was performed by presented in Tables 4 and 5, respectively.
the same person (SW). Subsequent statistical analysis
was undertaken using a standard computer program For the above statistical analysis, the material was
(Microsoft Excel®). tested for the normal requirements for using these two

Table 1 Double determination based on double point digitation by one observer (n = 10).

Parameter Observation 1 Observation 2 t Significance r

Mean Variation Mean Variation

1. Width of maxillary base 39.99 24.89 40.06 24.76 1.172 ns 0.99

2. Maxillary alveolar width 39.46 22.38 39.45 22.10 0.149 ns 0.99
3. Inter-molar width (apex) 21.60 11.22 21.57 10.89 0.818 ns 0.99
4. Inter-molar width (crown) 37.79 19.46 37.74 19.25 0.667 ns 0.99
5. Nasal width 22.85 5.89 22.84 5.85 0.148 ns 0.99
6. Inter-canine width (crown) 25.88 16.72 25.71 16.02 1.306 ns 0.99
7. Inter-canine width (apex) 18.94 6.11 19.08 6.09 1.401 ns 0.99
8. Right molar angulation 122.78 32.89 122.54 32.29 0.459 ns 0.96
9. Left molar angulation 122.98 35.50 123.35 21.28 1.106 ns 0.98
10. Right canine angulation 101.52 32.29 101.01 31.96 1.146 ns 0.97
11. Left canine angulation 102.63 59.84 102.09 55.02 0.914 ns 0.97
212 B. P O D E S S E R E T A L .

Table 2 Results of the intra-observer analysis based on double point identification (tracing on acetate paper) and digitizing
by one observer (n = 10).

Parameter Observation 1 Observation 2 t Significance σ r

Mean Variation Mean Variation

1. Width of maxillary base 39.99 24.87 40.09 24.38 0.684 ns 0.32 0.99
2. Maxillary alveolar width 39.46 22.38 39.56 20.88 0.400 ns 0.53 0.98
3. Inter-molar width (apex) 21.60 11.22 21.75 8.28 0.468 ns 0.68 0.97
4. Inter-molar width (crown) 37.79 14.46 37.71 17.09 0.463 ns 0.37 0.99
5. Nasal width 22.85 5.89 22.92 5.96 0.378 ns 0.40 0.97
6. Inter-canine width (crown) 25.88 16.73 25.81 16.65 0.385 ns 0.39 0.99
7. Inter-canine width (apex) 18.94 6.11 19.16 6.90 1.004 ns 0.49 0.96
8. Right molar angulation 122.78 32.89 123.48 40.55 0.738 ns 2.07 0.88
9. Left molar angulation 122.98 34.50 121.51 18.25 1.088 ns 3.05 0.98
10. Right canine angulation 101.52 32.29 101.34 43.26 0.244 ns 1.56 0.94
11. Left canine angulation 102.63 59.84 102.16 30.82 0.394 ns 2.55 0.89

Table 3 Individual differences with descriptive statistics for the evaluation of intra-observer differences. All differences
expressed as positive values.

Subject Var. 1 Var. 2 Var. 3 Var. 4 Var. 5 Var. 6 Var. 7 Var. 8 Var. 9 Var. 10 Var. 11

1 0.2 0.0 0.1 0.0 0.1 0.1 2.9 1.2 1.8 0.5 0.2
2 0.1 0.4 1.2 0.5 0.0 0.9 1.5 2.6 0.0 2.7 0.1
3 0.2 1.4 0.4 0.3 0.7 1.6 2.1 2.4 4.7 5.4 0.4
4 0.4 0.8 0.5 0.6 0.4 0.2 4.3 3.4 0.5 2.7 0.2
5 1.1 0.2 0.9 0.0 1.0 0.2 1.9 5.0 0.3 2.4 0.8
6 0.1 0.3 1.2 0.9 0.1 0.4 0.2 5.8 3.4 5.4 0.4
7 0.5 1.3 1.6 0.4 0.0 0.6 1.9 8.9 2.5 4.2 0.7
8 0.2 0.7 1.0 0.9 0.9 0.3 1.3 0.8 0.1 4.8 1.0
9 0.1 0.8 0.3 0.1 0.2 0.8 4.6 3.4 2.2 1.2 0.2
10 0.5 0.1 1.3 0.8 0.7 0.3 4.7 3.2 0.9 3.0 0.4
Mean 0.34 0.60 0.85 0.45 0.38 0.54 2.54 3.67 1.64 3.23 0.44
Maximum 1.10 1.40 1.2 0.90 1.00 0.60 4.70 8.90 4.70 5.40 1.00
Minimum 0.50 1.30 1.60 0.90 0.70 1.60 4.60 0.80 0.00 0.50 0.90
SD 0.14 0.25 0.32 0.18 0.18 0.22 0.95 0.75 0.47 1.68 0.29

SD, standard deviation.

Table 4 Inter-observer variation based on the comparison of the average of two determinations by each observer (n = 10).

Parameter Observation 1 Observation 2 t Significance σ r

Mean Variation Mean Variation

1. Width of maxillary base 40.14 24.57 40.36 24.55 2.132 ns 0.67 0.98
2. Maxillary alveolar width 39.51 21.47 40.23 24.31 2.917 * 1.31 0.96
3. Inter-molar width (apex) 21.75 9.56 22.05 11.44 1.154 ns 1.32 0.96
4. Inter-molar width (crown) 37.75 18.11 38.94 27.38 1.848 ns 1.57 0.94
5. Nasal width 22.89 5.85 21.96 2.77 2.229 ns 0.57 0.97
6. Inter-canine width (crown) 25.85 16.61 24.89 13.31 1.432 ns 0.93 0.86
7. Inter-canine width (apex) 18.99 6.20 19.68 8.48 1.208 ns 0.81 0.92
8. Right molar angulation 123.13 34.57 124.15 52.79 1.213 ns 1.40 0.93
9. Left molar angulation 122.25 21.82 121.58 17.64 0.438 ns 2.05 0.43
10. Right canine angulation 101.43 36.42 103.07 86.09 0.997 ns 3.48 0.85
11. Left canine angulation 102.39 41.78 98.40 51.49 2.076 ns 3.87 0.60

ns = not significant; * P < 0.05.

C T QUA N T I TAT I O N O F M A X I L L A RY D I M E N S I O N S 213

Table 5 Individual differences with descriptive statistics for the evaluation of inter-observer analysis. All differences
expressed as positive values.

Subject Var. 1 Var. 2 Var. 3 Var. 4 Var. 5 Var. 6 Var. 7 Var. 8 Var. 9 Var. 10 Var. 11

1 0.50 0.60 0.15 0.30 0.10 1.95 0.95 2.55 0.10 0.00 3.65
2 0.75 0.00 0.40 0.75 0.45 2.30 1.55 1.25 3.50 7.70 3.65
3 0.50 2.50 5.00 4.20 1.65 0.15 0.00 0.95 2.90 4.35 4.30
4 1.00 0.70 0.45 0.40 0.40 0.90 1.50 0.75 1.30 0.75 4.65
5 0.25 0.80 0.15 0.00 0.50 1.20 0.70 2.85 1.80 10.2 11.0
6 1.55 1.65 1.40 0.10 1.15 1.15 1.30 1.00 3.80 5.30 2.80
7 0.65 1.35 0.20 2.95 0.50 0.40 1.10 0.75 3.95 2.75 0.60
8 1.80 4.25 2.70 4.50 0.45 0.95 1.55 3.35 4.40 1.55 1.30
9 0.85 1.90 0.45 1.40 1.15 0.40 1.00 2.40 1.10 3.90 9.70
10 0.55 0.55 0.05 0.60 0.40 1.85 0.95 1.75 2.70 2.35 2.90
Mean 0.84 1.43 1.09 1.49 0.67 1.125 1.106 1.96 2.565 3.885 4.455
Maximum 1.80 4.25 5.00 4.70 1.65 2.30 1.55 5.35 4.50 10.20 11.0
Minimum 0.25 0.00 0.05 0.00 0.10 0.15 0.00 0.55 0.45 0.00 0.60
SD 0.154 0.391 0.504 0.568 0.151 0.227 0.149 1.424 1.437 1.003 1.062

SD, standard deviation.

parametric tests, namely a normal distribution and correlation analysis), and a summary of the individual
homogeneity of variance. differences between the two readings in Table 5. From
Table 4 it can be seen that, with the exception of
(c) Determination of the method error was calculated parameter 2, the maxillary alveolar width, there were no
using Dahlberg’s formula (Dahlberg, 1940): σ = √(Σd2/2n), significant statistical differences between the readings
where σ is the error of the measurement, d is the differ- of the two observers, although the t-values were higher
ence between the first and second readings and n is the than those for the intra-observer differences (Table 2).
number of subjects. Variable 2 showed a difference corresponding to t = 2.917,
P < 0.05. Both linear and angular variables demonstrated
(average) values of σ that were greater than those for
Results
the intra-observer variance (Table 2). However, the
The results of the repetition of the digitizing process can greatest average value for a linear parameter was still
be seen in Table 1, expressed by the mean values and only 1.57 mm and for the angular variable 3.87 degrees.
variation. For all parameters there was a slight difference Considering the individual values, single differences
between the first and second readings, although this was in the linear variables of up to 5.0 mm were observed
not statistically significant. The correlation between the (Table 5). The angular variables also showed a greater
first and second readings was very high, with values difference than those for the intra-observer analysis.
ranging from 0.48 to 0.98. While there was no statistical difference between the
Correlation of the quantitation of intra-observer readings from the two observers, single differences of up
error is given in Table 2 (paired t-test and Spearman’s to 7.7 degrees could be seen. In all cases, the differences
correlation analysis). It can be seen that double between the first and second readings showed a reason-
registration of the radiographs by the same observer did able distribution of positive and negative values (although
not result in significant differences. The (average) error for statistical purposes represented by positive results in
of the method for the linear variables (1–7) as described the respective tables).
by σ was generally less than 0.5 mm, although linear The effect of choosing alternative slices which still
values involving the apices of the molar and canine fulfilled the criteria for quantitation, although not being
teeth demonstrated relatively high values. The values of the most anterior slice, is shown in Figure 5a, b. For both
t were generally similar to those in Table 1, although the the angular and linear variables very little variation
correlation coefficients were a little lower, once again existed, and the size of the difference between the slices
with the angular parameters demonstrating the lowest did not exceed the differences that could be expected
correlation coefficients. For the individual variation, very when remeasuring a chosen slice.
few of the linear values demonstrated a difference between
the first and second readings of more than 1.0 mm, with
Discussion
a maximum value of 1.4 mm (Table 3). For the angular
parameters, the maximum average value was 3.05 degrees. As stressed by Vanarsdall and White (1992), the lateral
The results of quantitation for inter-observer error dimension can be regarded as the ‘forgotten dimension’
are shown in Table 4 (paired t-test and Spearman’s of dentofacial analysis, due partly to the lack of simple
214 B. P O D E S S E R E T A L .

Figure 5 Variation in dimensions for several slices, each fulfilling the criteria for selection. (a) Linear variables (mm).
Maxillary alveolar width: 3–4; maxillary base: 5–6; inter-molar width (apex): 9–10; inter-molar width (crown): 7–8;
nasal width: 17–18; inter-canine width (crown): 11–12; inter-canine width (apex): 13–14. (b) Angular variables
(degrees). Right molar angulation: 7–9/1–2; left molar angulation 8–10/1–2; right canine angulation 1–13/15–16; left
canine angulation 12–14/15–16.

registration corresponding to the lateral cephalogram would arise from the placement of the patient in the
in sagittal and vertical analysis. As such it can be antici- scanner. For ethical reasons, double determination at
pated that the transverse dimension, and not least its that level was not possible.
quantitation, will be the focus of future studies. The process of digitization revealed that although
Tomographic scanning has already been used in there were no significant differences between the first
orthodontic diagnosis (Fuhrmann et al., 1994), in con- and second readings, some variation in values could be
nection with cleft palate subjects, the temporomandibular expected. The weakness of the digitizer in quantitation
joint (Seren et al., 1994; Moaddab et al., 1985) and in the has been described by Eriksen and Solow (1991) and,
case of ectopic canine teeth (Ericson and Kurol, 1988; consequently, the reproducibility of the digitizer and the
Schmuth et al., 1992). The possibilities for CT scanning digitizing process has to be rigorously controlled.
of transverse discrepancies were recommended by Timms Intra-observer variation revealed no significant vari-
(1982) after a pilot study, although his investigation was ation between the first and second readings, only very
based on axial scanning. In orthodontic diagnosis and small individual variations, both positive and negative,
treatment planning it is customary to differentiate were seen. These values correspond well (subjectively)
between skeletal structures, the morphology and size of with those for standard cephalometric reference points
which are thought to be largely genetic, and dento- (Baumrind et al., 1971b).
alveolar structures, i.e. the structures immediately The results of the inter-observer analysis showed, as
surrounding the teeth and dependent on the presence of would be expected, greater variation, although with the
the teeth and the surrounding environment (Solow, exception of one parameter, representing the maxillary
1980). The efficiency of orthopaedic and orthodontic alveolar width, statistically significant differences were
appliances is usually based on observed changes in these not seen (no reason for the slight significance in variable
areas. 2 can be given, but the phenomenon of mass significance,
All forms of quantitation raise questions as to the occurring when statistical tests are repeated continually,
validity and reproducibility of the structures measured. could be the answer). The variation between observers
For standard cephalometric analyses used in orthodontic was small compared with the biological variation from
diagnosis, these have been well reported by Baumrind subject to subject.
et al. (1971a,b). The present study aimed to investigate In general the errors for the angular parameters
reproducibility at three levels, namely the process of were larger than for the linear variables, which seems
digitization as well as intra- and inter-observer variation. reasonable, as each parameter is constructed on three
It must also be realized that another source of error points each with its own variation. The small distance
C T QUA N T I TAT I O N O F M A X I L L A RY D I M E N S I O N S 215

between the points involved will also tend to increase References

the error. The variation in the angular values corresponds Athanasiou A 1986 The frontal cephalometric analysis. Thesis,
well with similar values based on reproducibility in Royal Dental School, Århus, Denmark
standard cephalometrics (Baumrind et al., 1971b), where Angle E 1907 Malocclusion. S. S. White, Philadelphia
parameters involving the apices of teeth seem least Baumrind S, Frantz R C 1971a The reliability of head film
reproducible, as found in the present study. measurements 1. Landmark identification. American Journal of
The measurements performed in this experiment could Orthodontics 60: 111–127
have been carried out on dry skulls. It was, however, Baumrind S, Frantz R C 1971b The reliability of head film
measurements 2. Conventional angular and linear measures.
decided to perform the study on living patients in order American Journal of Orthodontics 60: 505–517
that the experimental set-up, including the process of Björk A, Skieller V 1977 Growth of the maxilla in three dimensions
placing the patient in the CT scanner, could be included as revealed radiographically by the implant method. British
in the source of variation. Journal of Orthodontics 4: 53–64
The interpretation and use of the results of this study Dahlberg G 1940 Statistical methods for medical and biological
are important. Methodological variation must be seen students. George Allen and Unwin, London, pp. 122–132
in relation to the use to which the parameters are to be Ericson S, Kurol J 1988 CT diagnosis of ectopically erupting maxillary
canines—a case report. European Journal of Orthodontics 10:
put. In diagnosis where the individual patient is compared 283–295
with ‘standard values’ obtained from a particular control Eriksen E, Solow B 1991 Linearity of cephalometric digitizers.
population, the methodological variation (error) must European Journal of Orthodontics 13: 337–342
be smaller than the biological variation between the Fuhrmann R A W, Frohberg U, Diedrich P 1994 Treatment
individuals that comprise the control population. When prediction with three dimensional computer tomographic skull
evaluating the effect of therapy on the dimensions models. American Journal of Orthodontics and Dentofacial
Orthopedics 106: 156–160
measured, it is necessary to relate the methodological
Krebs A 1958 Expansion of the midline suture studied by means of
error to the expected effect of the therapy. In both cases, metallic implants. Transactions of the European Orthodontic
an understanding of the size of the methodological error Society, pp. 131–142
will be important in interpreting results, not least negative Moaddab M B, Dumas A, Chavoor A, Neff P, Homayoun N 1985
results. Temporomandibular joint: computer tomography three dimensional
reconstruction. American Journal of Orthodontics 88: 342–353
Schmuth G P F, Freisfeld M, Köster O, Schüller H 1992 The
Conclusions application of computerized tomography in cases of impacted
canines. European Journal of Orthodontics 14: 296–301
The results of the present study indicate that the system Seren E, Akan H, Toller M O, Akyar S 1994 An evaluation of
described represents a reasonable method by which the the condylar position of the temporomandibular joint by
transverse morphology of the maxillary structures can computerised tomography in Class III malocclusion: a preliminary
study. American Journal of Orthodontics and Dentofacial
be described. The difference between intra- and inter-
Orthopedics 105: 483–488
observer variation indicates that when possible, all
Skieller V 1964 Expansion of the midpalatal suture by removable
quantitation should be performed by the same person. plates analysed by the implant method. Transactions of the
Based on the findings regarding values for single European Orthodontic Society, pp. 143–158
patients of a few, although relatively large, differences Solow B 1980 The dento-alveolar compensatory mechanism: back-
in inter-observer variation, the method would be better ground and clinical implications. British Journal of Orthodontics
7: 145–161
for comparisons of groups using average values than for
Timms D J 1980 A study of basal movement with rapid maxillary
individuals.
expansion. American Journal of Orthodontics 80: 500–507
The method constitutes a useful supplement to
Timms D J 1982 A computer tomographic assessment of maxillary
orthodontic diagnosis, not least considering the extreme movement induced by rapid expansion–a pilot study. European
weakness of all other methods available and can, there- Journal of Orthodontics 4: 123–127
fore, be recommended. Vanarsdall R L, White Jr R P 1992 The 1990s. The era of the
transverse dimensions. International Journal of Adult Orthodontics
and Orthognathic Surgery 7: 3 (Editorial)
Address for correspondence
Stephen Williams
Majsvænget 20
Malling 8340
Denmark