J Oral MaxillofacSurg

4&1161-1167,199O

Accuracy of Model Surgery:

Evaluation of an Old Technique and Introduction of a
New One
EDWARD ELLIS III, DDS, MS*

This article evaluates the ability of 4th-year residents to accurately perform

maxillary model surgical repositioning in the conventional manner. Using
a special model surgery-measuring platform, measurements of the max-
illary dental casts were recorded before and after model surgery was per-
formed by residents in 20 bimaxillary cases. The model surgery was per-
formed in the usual manner described in several texts, using reference
lines scored on the casts and measurements made to the incisal pin and
upper member of the articulator. Analysis of the differences between the
planned movements and the postsurgical position of the maxillary model
surgery casts showed statistically significant errors in maxillary reposi-
tioning for all measures. This indicates that the manner in which model
surgery was performed by the residents, as reported in the literature, was
inaccurate. A new technique and instrument for more accurately recording
measurements and repositioning the maxillary cast is introduced.

Mock surgery on dental casts is a routine proce-
dure for purposes of both diagnosis and treatment
planning in patients with dentofacial deformities.
The technical details of model surgery have ap-
peared in the clinical literature for years, with little
difference in their content.lm5 However, all tech-
niques described contain potential errors that
lead to inaccuracy and difficulties. These can be
summarized as 1) improper mounting of the casts;
2) errors in the placement of reference lines on the
casts; and 3) errors in measuring the surgical dis-
placement of segments.
The most difficult aspect in performing model
surgery is in the repositioning of the maxillary cast
* Associate Professor, Oral and Maxillofacial Surgery, The
University of Texas Southwestern Medical School, Dallas.
This research was supported in part by a grant from the
Chalmers J. Lyons Academy-James R. Hayward Research
Fund.
Address correspondence and reprint requests to Dr Ellis: Di-
vision of Oral and Maxillofacial Surgery, The University of
Texas Southwestern Medical School, 5323 Harry Hines Blvd.
Dallas. TX 75235.
0 1990 American Association of Oral and Maxillofacial
geons
0278-2391/90/4811-0005$3.00/0
during bimaxillary surgery. In the usual sequence of
steps performed during bimaxillary model surgery,
the first is repositioning the maxilla. After section-
ing the maxilla, repositioning it in all three planes of
space to the desired location is extremely difficult.
In isolated maxillary model surgery, the mandible
serves to determine the proper anteroposterior and
transverse position of the maxilla once the proper
vertical dimension has been attained. In contrast,
during bimaxillary model surgery the mandible can-
not be used as a guide for maxillary repositioning
because it will also undergo changes in position.
The necessity of holding the maxillary cast on the
articulator against the pull of gravity adds to the
difficulty; frequently, it becomes displaced before it
can be secured.
Because surgical splints are made from the repo-
sitioned models, it is important that model surgery
be performed accurately as errors will be trans-
posed to the patient. To date, no studies have been
performed that determine the accuracy of maxillary
model repositioning during bimaxillary model sur-
gery. The purpose of this investigation was to de-
Sur- termine how accurately 4th-year residents can per-
form maxillary model repositioning during bimaxil-
lary cases.

1161
1162 ACCURACYOFMODELSURGERY

Methods

Twenty consecutive cases of bimaxillary surgery

that met the following criteria were included in this
study: 1) a dentition at least first molar to first molar
bilaterally; 2) the entire maxilla was to undergo re-
positioning in any direction without segmentation;
3) sufficient details on the proposed maxillary sur-
gery were available: the planned change in antero-
posterior and vertical position of the incisor, the
planned change in vertical position of the molars,
the planned change in the position of the dental mid-
lines, and the planned mediolateral position of the
posterior dental arch. The dental casts of each case
were mounted on a semiadjustable articulator using
a face-bow transfer and centric relation interoc-
clusal registration. Following mounting and verifi- FIGURE 2. Method of measuring the vertical position of the
maxillary incisor.
cation that the casts were mounted correctly, the
maxillary dental cast was removed from the articu-
lator and mounted on the Erickson Model Surgery tical position of the mesiobuccal cusp tip of the first
Platform (Great Lakes Orthodontic Ltd, Ton- molars bilaterally (Fig 3); 3) the anteroposterior po-
awanda, NY). The model surgery platform is a de- sition of the most prominent central incisor (Fig 4);
vice that allows mounting of the models, after re- 4) the dental midline (Fig 5); 5) the mediolateral
moval from the articulator, on a square model-block position of the first molars (Fig 6). The maxillary
for measurement before, during, and after model cast was then replaced on the articulator and one of
surgery (Fig 1). Measurements of the casts are four senior residents performed maxillary model
made using a digital caliper that is mounted in the surgery in the classic manner. This included scoring
base of the model surgery platform. In this manner, reference lines on the sides of the cast to record
all measurements are made either perpendicular or changes in position of the maxilla. Most residents
parallel to the platform and, therefore, parallel or also measured from the facial surface of the central
perpendicular to Frankfort horizontal. The follow- incisor to the incisal pin for determination of antero-
ing measurements were recorded by the author for posterior movements of the maxilla. Some also
each maxillary cast before model surgery: 1) the measured from the incisal edges of the teeth to the
vertical position of the incisors (Fig 2); 2) the ver- upper member of the articulator or the mounting

FIGURE 3. Method of measuring the vertical position of the

maxillary molar. The vertical position of the first molars are
FIGURE 1. The Erickson Model Surgery Platform (Great recorded bilaterally so that any mediolateral “cant” of the oc-
Lakes Orthodontic Ltd, Tonawanda, NY). clusal plane can be identified.
EDWARD ELLIS III
FIGURE 4. Method of measuring the anteroposterior position
of the maxillary incisor. The model-block is simply turned on its
side (heel) and a measurement is taken to the facial surface of the
incisors.
ring for determination of the vertical position of the
maxilla. Once the resident felt that the maxillary
model was in its proper postsurgical position, it was
returned to the model-block for repeated measure-
ments by the author to determine the actual changes
induced by model surgery.
The planned and actual values for the change in
anteroposterior and vertical position of the incisor,
the vertical position of the molars, the position of
the dental midlines, and the mediolateral position of
the dental arch at the first molars were tabulated.
All measures were converted to positive numbers
for statistical analysis to factor out differences in
1 vertical position of the central incisor had a mean
FIGURE 5. Method of measuring the mediolateral position of
the maxillary dental (incisor) midline. This is performed by tum-
ing the model-block on its side and sliding the caliper to the
interdental space between the incisors.
1163
FIGURE 6. Method of measuring the mediolateral position of
the posterior maxillary dental arch. The caliper is brought in
contact with the facial surface of the first (or second) molar or
orthodontic tube.
direction. Thus, mean error values reported repre-
sent total variance from the planned value, not di-
rection of movement. Comparison between the
planned and actual movements for each variable
were made using the paired t test. Pearson’s corre-
lation coefficients were calculated to determine the
significance of interactions between variables.
Results
The data for each of the 20 cases are presented in
Table I. The anteroposterior position of the central
incisor in the twenty cases had a mean deviation of
1.2 mm (SD, 0.9) from the planned value. This was
a statistically significant difference (P < 301). The
deviation of 1.1 mm (SD, 1.2; P < .OOl) from the
planned value. The vertical position of the right and
left molars had mean deviations of 1.1 mm (SD, 0.9;
P < 301) and 1.2 mm (SD, 1,l; P < .OOl), respec-
tively, from planned values. The mediolateral
change in position of the maxillary midline varied
0.9 mm (SD, 0.5; P < 301) from the plan. The
mediolateral change in position of the first molar
(posterior dental arch) varied 1.9 mm (SD, 0.7; P <
401). Thus, the mean movements of the maxilla for
all variables were significantly different from the
planned movements at greater than the .OOl level of
confidence. This statistical test, in effect, analyzed
the mean difference between the planned and actual
positioning of the maxilla to determine if they were
significantly different from 0, which they were. The
data were also subjected to the paired t test to de-
termine the significance of differences between the
1164 ACCURACY OF MODEL SURGERY

Table 1. Predicted and Actual Measures of Maxillary Position After Model Surgery
A-P Ul v UI VU6R v U6 L ML M-L U6
Case P A P A P A P A P A P A

1 + 3.0 + 0.8 0.0 J 2.2 t 6.0 t 5.4 t 6.0 t5.1 0.0 0.4R 0.0 2.lR
(error) 3.8 2.2 0.4 2.1
2 -+ 5.0 -+ 4.4 t 6.0 t 5.0 t 6.0 t::: t 6.0 t::; 2.OR 1.6R 0.0 1.4R
(error) 0.6 0.3 0.4 1.4
3 + 7.0 -+ 7.1 0.0 1::: 0.0 7::; 0.0 t 1.3 0.0 0.6L 0.0 2.2L
(error) 0.1 0.7 0.6 2.2
4 0.0 +- 2.4 t 2.0 r::: t 7.0 f 6.8 t 7.0 t::: 1.5L 1.8L 0.0 2.6L
(error) 2.4 1.3 0.4 0.3 2.6
5 + 4.0 + 4.7 4 3.0 4 2.1 J 3.0 r::: 4 3.0 i 2.9 0.0 1.4L 0.0 3.lL
(error) 0.7 1.4 3.1
6 + 2.0 + 2.1 t 2.0 7;:: t 2.0 r::: t 2.0 r;:: 2.OL 1.5L 0.0 0.9L
(error) 0.1 0.2 0.5 0.9
7 + 3.0 -+ 2.1 J 3.0 r::: 4 4.0 r::; J 4.0 J 7.0 2.OR 2.5R 0.0 l.lR
(error) 0.9 0.1 3.0 0.5 1.1
8 + 5.0 + 5.8 0.0 J 0.8 0.0 1;:: 0.0 J 1.3 0.0 1.7L 0.0 2.8L
(error) 0.8 1.0 1.3 1.7 2.8
9 -+ 2.0 + 1.1 13.0 i;:: J 4.0 13.0 0.0 J 0‘7 2.OL 0.9L 0.0 2.4L
(error) 0.9 0.9 1.0 0.7 1.1 2.4
10 + 5.0 + 4.7 J 4.0 J 3.5 5_4.0 12.5 4 4.0 J 3.9 0.0 2.lR 0.0 2.lR
(error) 0.3 0.5 1.5 2.1 2.1
11 0.0 + 0.8 0.0 J 5.5 t 3.0 t 1.6 t 3.0 r::: 0.0 1.4L 0.0 2.5L
(error) 0.8 1.2 1.4 2.5
12 + 4.0 + 3.3 4 2.0 $:Z t 3.0 t::: t 3.0 t 2.9 0.0 1.8R 0.0 1.7R
(error) 0.7 1.8 1.7
13 + 5.0 -+ 2.4 t 4.5 ts:; t 6.0 7;:: T 6.0 t;:t 1.OL 0.2R 0.0 2.lR
(error) 2.6 1.5 2.2 2.9 1.2 2.1
14 + 2.0 + 3.8 t 7.0 t 8.1 t 7.0 T 6.7 t 7.0 16.7 2.OR O.9R 0.0 1.8R
(error) 1.8 1.1 0.3 1.1 1.8
15 0.0 + 0.8 t 3.5 t 4.0 13.5 7 1.9 t 3.5 T;:: 0.0 0.6 0.0 I .8R
(error) 0.8 0.5 1.6 0.6 1.8
16 -+ 2.0 --+ 1.3 J 3.5 J 3.5 -13.5 J 4.0 4 3.5 I::: 0.0 0.8L 0.0 1.5L
(error) 0.7 0.8 1.5
17 -+ 5.0 -+ 5.6 t 1.0 r::; t 1.0 t;:; t 1.0 r;:: 0.0 0.7R 0.0 1.3R
(error) 0.6 1.1 0.5 0.7 1.3
18 0.0 + 1.9 t 2.0 t;:; t 1.0 t4.6 t 1.0 $ 1.6 1.OR 0.4R 0.0 0.8R
(error) 1.9 0.3 0.6 0.8
19 + 10.0 + 12.0 0.0 to.2 0.0 4;:; 0.0 $ 0.0 0.8R 0.0 2.4R
(error) 2.0 1.5 0.8 2.4
20 0.0 --f 1.2 t 1.0 t::‘6 t 2.0 4::; t 2.0 12.1 0.0 0.4R 0.0 0.9R
(error) 1.2 0.4 3.2 4.1 0.4 0.9

Abbreviations: A, actual; A-P Ul, anteroposterior position of the maxillary central incisor; L, left; ML, maxillary dental midline (right
c) left); M-L U6, mediolateral position of maxillary first molar (either); P, predicted; R, right; V Ul, vertical position of the maxillary
central incisor(s); V U6 L, vertical position of the maxillary left first molar; V U6 R, vertical position of the maxillary right first molar;
+ , anteriorly; + , posteriorly; t , superiorly; J , inferiorly.

planned and actual maxillary movements greater
than 1 mm. When the data were analyzed in this
manner, the only variable that still showed a signif-
icant difference was the mediolateral position of the
maxillary molars (P < .OOl).
Pearson’s correlation coefftcients showed that
there was a significant correlation between the error
in vertical positioning of the right and left maxillary
molars (r = 31; P < .OOl). Thus, when the maxilla
was improperly positioned vertically on one side,
the other side also was improperly positioned in the
same direction. The only other significant correla-
tion noted was between the error in positioning the
maxillary midlines and the error in mediolateral po-
sitioning of the maxillary molar (r = .47; P < 0.05).
This showed that when there was an error in posi-
tioning of the maxillary midline, there was also an
error in the mediolateral positioning of the posterior
maxilla, and vice versa.
Discussion
The results of this investigation suggest that max-
illary model surgery performed in the classic man-
EDWARD ELLIS III
ner is variable, frequently imprecise, and in some
instances induces extreme errors. This is especially
remarkable given the fact that the maxillae used in
this study were not segmented. Five of the six vari-
ables examined had a mean error in repositioning of
at least 1 mm; the sixth variable, the positioning of
the dental midlines, was almost 1 mm (0.9 mm)
away from the planned location. Because means
and standard deviations may be misleading, an anal-
ysis of each case becomes extremely useful. If we
accept inaccuracies of less than 1 mm in model sur-
gery, a review of individual cases shows that 19 of
the 20 cases had at least 2 (of the 6) measurements
with greater than I mm difference between the
planned and actual measures (Table 1). Thirteen
had at least three errors greater than 1 mm and four
cases had at least four errors greater than 1 mm. In
fact, only one case had no errors greater than 1 mm
(case 6).
The reasons for these inaccuracies are multiple,
but perhaps the most important reason is the differ-
ence between what one measures along the refer-
ence lines and what happens at the level of the
teeth.6,7 The further the reference lines are removed
from the dentition, the greater may be the differ-
ence between the movements at these areas. This is
especially important when the occlusal plane is
changed with surgery, for instance, in the correc-
tion of an anterior open-bite, when posterior intru-
sion of the maxilla is planned. Cases 1 and 4 had
much greater intrusion planned in the posterior
maxilla than in the anterior. In both of these cases,
large errors in the anteroposterior position of the
maxillary incisor occurred during model surgery. In
both, the incisors moved posteriorly more than
planned. This is easily explained if one considers
the geometry involved in using reference lines
scored along the sides of the dental cast to deter-
mine the anteroposterior position of the maxilla.
With posterior intrusion, the anterosuperior aspect
of the maxilla (ie, anterior nasal spine) is moved
forward. When measurements are made at the level
of the bone cut, it appears as if the entire maxilla
has moved anteriorly, when it is only the superior
aspect that has done so (Fig 7). As cases 1 and 4
illustrate, the incisor may not move anteriorly at all,
and, in fact, may move posteriorly. This same phe-
nomenon occurs during segmental surgery when
tilting of the maxillary segments and, therefore, ref-
erence lines can occur. It therefore behooves the
surgeon to measure directly to the dentition instead
of, or in addition to, the area of the bone cuts. Mea-
suring from the facial surface of the incisor to the
incisal pin, and from the incisal edge to the upper
member or mounting ring on the articulator, is
1165
FIGURE 7. Illustration showing how measurements made at
the level of the osteotomy provide inaccurate information about
the movement of the dentition when the occlusal plane is
changed. A, Prior to model surgery. E, After model surgery to
close an anterior open-bite by posterior maxillary impaction. If
measurements are made at the level of the osteotomy, they
would indicate that the maxilla had moved anteriorly approxi-
mately 5 qm. However, the arrows in both A and B are at the
same location with respect to the base of the dental cast, showing
that the incisal edge did not move in the anteroposterior plane.
much more accurate in identifying the anteroposte-
rior and vertical position of the maxillary incisors
than reference lines scored on the casts. Another
reason for the inaccuracies in performing model
surgery stems from the difficulties in making pre-
cise measurements along the complex surfaces of
the dental casts and articulator. This is especially
difficult when attempting to hold the maxillary
model against the force of gravity while making the
measurements.
Perhaps the most common yet occult inaccuracy
in performing maxillary model surgery during bi-
maxillary cases is in the mediolateral positioning of
the posterior maxilla. It is extremely difficult to de-
termine the symmetry of the dental arch in the me-
diolateral plane when model surgery is performed
on the articulator. In fact, most cases with preex-
isting asymmetries in the mediolateral position of
the posterior dental arch go unnoticed. For in-
stance, in the 20 cases reported in this article, 16
had greater than 1 mm, 11 had greater than 2 mm,
and 4 had greater than 3 mm of asymmetry before
model surgery. This means that if the case was
properly mounted on the articulator, asymmetries
were present that went unnoticed and were there-
fore untreated. A perusal of Table 1 shows that not
a single case had correction of posterior dental arch
asymmetry planned in the treatment, a reflection of
the fact that maxillary asymmetries are difficult to
determine clinically, radiographically, and on the
articulated models without accurate instrumenta-
tion.
Of even more concern is the finding that maxil-
lary asymmetries are induced during model surgery
in a significant number of patients. The mean error
1166
in mediolateral positioning of the posterior maxil-
lary dental arch was 1.9 mm when measured at the
first molar (P < .OOl). Seventeen of the 20 cases had
greater than 1 mm, 10 had greater than 2 mm, and 1
had greater than 3 mm of change in the mediolateral
position of the posterior maxilla that was not
planned. When this occurred, it usually accompa-
nied movements in the dental midline to the same
side (r = 47; P < .05). Although minor maxillary
arch asymmetries may seem inconsequential, one
must remember that the maxilla, once secured in its
planned position, serves as the target for mandibu-
lar repositioning during bimaxillary surgery. There-
fore, any asymmetry in the position of the maxilla
will be magnified in the mandible, especially at the
gonial angles, as these are a greater distance from
the axis of the asymmetry.
Those surgeons who look critically at their pa-
tients may have noticed an occasional patient who
appears asymmetric after bimaxillary surgery to
correct a symmetrical deformity. Most commonly,
one side of the face appears more prominent than
the other, especially along the lateral surface of the
rami. An examination of the dental midlines may
show alignment with the facial midline. However,
alignment of the dental and facial midlines should
not lull the clinician into a false sense of security
that the surgery has been performed correctly, for
the posterior maxilla may be asymmetrical, ie,
skewed to one side (Fig 8). If the maxilla is skewed
FIGURE 8. Illustration demonstrating how a facial asymmetry
can be present even while the maxillary dental midlines are co-
incident with the facial midline. If the posterior maxilla is rotated
to one side or the other during maxillary surgery, the mandible
will follow because it will be occluded with the maxilla during
mandibular surgery.
ACCURACYOFMODELSURGERY
to one side in the posterior area and the occlusion is
correct, the mandible will also be skewed the same
direction, resulting in facial asymmetry.
An advantage to the use of the Erickson Model
Surgery Platform for performing maxillary model
surgery in bimaxillary cases is that preexisting
asymmetries can be easily identified before model
surgery so that they can be corrected during model
surgery. An asymmetry readily identifiable with
this instrument is the one described above, ie, me-
diolateral asymmetries of the posterior maxilla.
These are difficult to determine any other way and
readily lend themselves to correction during bimax-
illary surgery.
Another common asymmetry one finds is differ-
ence in the vertical position of the maxillary molars
from one side to the other, ie, the “cant” of the
maxilla. These are usually small; when larger than 1
to 2 mm, they can also be seen on an accurate ceph-
alogram and identified clinically with careful mea-
surements between the canines and the medial can-
thi using a caliper. Note in Table 1 that only one
case (case 9) had differential vertical movements of
the posterior maxilla planned. Of interest, however,
is that 13 cases had greater than 1 mm, 5 had greater
than 2 mm, and case 9 had greater than 3 mm
of vertical difference from one side to the other
present before model surgery. This does not mean
that a cant was present in every case, as an asym-
metry in the position of the auditory canals may
also have been present, distorting the face-bow
mounting. However, only case 9 was planned for
correction of this discrepancy. If minor cants were
diagnosed either clinically, radiographically, or on
the model platform, they could have easily been
eliminated with bimaxillary surgery.
Because of the difficulties performing maxillary
model surgery for bimaxillary cases on the articu-
lator, the author and associates have used the
model surgery platform for over 5 years. We have
found it simpler and more rapid than performing
maxillary model surgery on an articulator. Because
the maxillary model surgery is not performed on the
articulator, but on the model-block, ie, “upside-
down,” gravity works to the advantage of the sur-
geon and greatly facilitates repositioning of the
model. Any point on the cast is measured by simply
sliding the caliper to the point on the cast being
measured. These points may be dental, skeletal, or
soft tissue. Measurements along the dentition pro-
vide information on the location of the teeth before
and after model surgery. Measurements can also be
recorded at any level along the proposed osteotomy
side if one desires to determine the osseous changes
at that level.
EDWARD ELLIS III
Using the model surgery platform, we have
learned that minor malpositions, so easily intro-
duced when performing model surgery on an artic-
ulator, are eliminated, resulting in more precise re-
positioning of the maxillary dental cast. The making
of measurements directly to the teeth (and bones as
desired) assures that no surprises will result during
and after surgery.
References
1. Bell WH: Correction of the short-face syndrome-vertical
maxillary deficiency: A preliminary report. J Oral Surg
35:110, 1977
2. Bell WH, Creekmore TD, Alexander RG: Surgical correc-
tion of the long face syndrome. Am J Orthod 7140, 1977
1167
3. Epker BN. Fish LC: Surgical-orthodontic correction of
open-bite deformity. Am J Orthod 71:278, 1977
4. Hohl TA: Use of an adjustable (anatomic) articulator for
case prediction in segmental surgery, in Bell WH, Profftt
WR. White RP (eds): Sureical Correction of Dentofacial
Deformities. Philadelphia: PA, Saunders, 1980, pp 169-
177
5. Hill SC: Cephalometric planning and mode1 surgery. In Bell
WH (ed): Surgical Correction of Dentofacial Deformi-
ties-New Concepts. Philadelphia, PA, Saunders, 1985,
pp 217-226
6. Ellis E, Gal10 WJ: A method of accurately positioning the
maxillary incisor position in two-jaw surgery. J Oral Max-
illofac Surg 42402, 1984
7. Stanchina R, Ellis E, Gal10 WJ, et al: A comparison of two
measures for repositioning the maxilla during orthog-
nathic surgery. Int J Adult Ortho Orthogn Surg 3:149.
1988