Int. J. Oral Maxillofac. Surg.

2018; 47: 470–479

http://dx.doi.org/10.1016/j.ijom.2017.09.008, available online at http://www.sciencedirect.com

Clinical Paper
Orthognathic Surgery

SPECT bone scintigraphy for B. H. Chan, Y. Y. Leung

Oral and Maxillofacial Surgery, Faculty of
Dentistry, The University of Hong Kong,
Prince Philip Dental Hospital, Sai Ying Pun,

the assessment of condylar Hong Kong

growth activity in mandibular

asymmetry: is it accurate?
B. H. Chan, Y. Y. Leung: SPECT bone scintigraphy for the assessment of condylar
growth activity in mandibular asymmetry: is it accurate?. Int. J. Oral Maxillofac.
Surg. 2018; 47: 470–479. ã 2017 International Association of Oral and Maxillofacial
Surgeons. Published by Elsevier Ltd. All rights reserved.

Abstract. The comparison of serial radiographs and clinical photographs is

considered the current accepted standard for the diagnosis of active condylar
hyperplasia in patients with facial asymmetry. Single photon emission computed
tomography (SPECT) has recently been proposed as an alternative method. SPECT
can be interpreted using three reported methods absolute difference in uptake,
uptake ratio, and relative uptake. SPECT findings were compared to those from
serial comparisons of radiographs and clinical photographs taken at the time of
SPECT and a year later; the sensitivities and specificities were determined. Two
hundred patient scans were evaluated. Thirty-four patients showed active growth on
serial growth assessment. On comparison with serial growth assessment, the
sensitivity and specificity of the three methods ranged between 32.4% and 67.6%,
and 36.1% and 78.3%, respectively. Analysis using receiver operating characteristic
(ROC) curves revealed area under the curve (AUC) values of <0.58. The average
Key words: condylar hyperplasia; SPECT bone
age (mean  standard deviation) of patients with active growth was 18.6  2.8
scintigraphy; serial radiograph tracing; sensitiv-
years, and average growth in the anteroposterior, vertical, and transverse directions ity; specificity.
was 0.94  0.91 mm, 0.88  0.86 mm, and 1.4  0.66 mm, respectively. With such
low sensitivity and specificity values, it is not justifiable to use SPECT in place of Accepted for publication
serial growth assessment for the determination of condylar growth status. Available online 9 October 2017

Facial asymmetry is an aesthetically and
functionally unsatisfactory condition that
may arise from a number of causes, such
as joint resorption, infection-related
growth disturbances, and neoplastic
changes, among others1–3. Amongst these
causes, condylar hyperplasia is a patho-
logical condition that presents as a pro-
gressive and excessive growth of one or
both mandibular condyles3–5. It has an
unknown aetiology, although possible
causes including trauma, functional influ-
ence, developmental causes, and heredity
or genetics, have been suggested1. Pro-
gressive condylar growth eventually
results in alterations to the dimensions
of the condylar neck, ramus, and body
of the mandible6. A gradual progression

0901-5027/040470 + 010 ã 2017 International Association of Oral and Maxillofacial Surgeons. Published by Elsevier Ltd. All rights reserved.

Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.

of the asymmetry over a long period of
time of several years has been observed to
occur, but it may also occur within a short
time span of weeks to months. The condi-
tion may present in early childhood at 6–7
years of age, or at any time during the
patient’s adolescence growth period, and
even after the cessation of skeletal bone
growth7,8. This makes the timing of treat-
ment commencement crucial and possibly
difficult to predict. The incidence has been
reported to be higher in females, with
more than 64% of those presenting with
the condition being female according to a
systematic review by Raijmakers et al.9;
however, other studies have not shown this
sex difference10.
Condylar hyperplasia was first described
by Robert Adams in 1836 and then,
amongst others, by George Humphry in
1856, who reported dentofacial deformity
with resultant chin deviation as the char-
acteristics of this condition. This is also
correlated to dental malocclusion with a
non-incident dental midline and unilateral
crossbite or open bite3. In 1986, Obwegeser
and Makek classified condylar hyperplasia
into two main types, namely hemimandib-
ular hyperplasia and hemimandibular elon-
gation11. These presentations may occur
individually or in combination, to result
in the eventual dentofacial asymmetry. Nu-
merous other classifications have been sug-
gested1,12, illustrating the difficulties faced
in describing this condition.
While the condition is generally self-
limiting, the extent of the gross asymmetry
is dependent on the extent of condylar
hyperplastic growth activity10,13. The con-
dition may contribute to functional and
aesthetic deficits, which may result in a
negative impact on the patient’s quality of
life and function. A study conducted by
Naini et al. showed that the patient’s desire
to seek surgical treatment for the correc-
tion of this condition was correlated with
the severity of the asymmetry14. It was
also found that a small 5 mm deviation is
noticeable to the layperson.
The use of serial observation of growth
and serial cephalometric and dental model
comparisons is considered as the current
accepted standard for the determination of
condylar growth status4,15,16. This tech-
nique requires a minimum period of time
between observations (usually 6 months to
1 year) in order to determine the status,
which is time-consuming. The findings
from the comparisons can only reveal past
growth history, and they do not allow the
future condylar growth potential to be
determined15,17.
Bone scintigraphy has been used to aid
in the determination of bone growth activ-
Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.
ity since the early 1980s. The initial meth-
od of scintigraphy involved capture in
planar view, with the amount of uptake
by the condyles correlated to the T4 spine.
This required separate imaging of the T4
spine, as it could not be captured within
the condylar view. In recent years, single
photon emission computed tomography
(SPECT) has been proposed as a more
accurate and reliable form of bone scin-
tigraphy, as it allows the three-dimension-
al evaluation of condylar uptake. The
clivus bone, which is situated at the skull
base, is considered to be metabolically sta-
ble after the fusion of the spheno-occipital
synchondrosis. It is used as a reference point
for comparison of uptake values in the
condyles15,18,19. As the clivus is also cap-
tured along with the condyles during the
scan, there is no need for two separate
images to be captured. However, it has been
reported that bone scans are sensitive but
not specific, and that conditions affecting
the joints, such as healing bone (post-sur-
gery or trauma), joint infection, inflamma-
tion (arthritis), and neoplastic changes, can
result in a positive scan value20–23.
The reported sensitivity and specificity
of SPECT scans has varied amongst stud-
ies. Sensitivity values of between 78% and
98%, with a pooled sensitivity of 90%, and
specificity values of between 60% and
95%, have been reported13,19,24. In the
study by Saridin et al., the optimal cut-
off value for comparison between the
affected and contralateral condyles was
determined to be 55%, and the reported
sensitivity and specificity were around
88%19. Another study suggested that the
optimal cut-off is 56%, with sensitivity of
93% and specificity of 96%, based on the
area under the receiver operating charac-
teristics curve4. However, the results from
these previous studies supporting SPECT
scans were obtained by deriving the out-
comes from analyzing patients with con-
dylar hyperplasia at a single clinical time
point only, and there was no additional
external validation or comparison of the
results to an accepted standard. Hence,
there is a knowledge gap regarding the
true sensitivity and specificity of SPECT
as a diagnostic test. As it appears that there
have been no previous studies comparing
SPECT against a reliable standard, and
with the reported wide range of sensitivity
and specificity values, it remains neces-
sary to measure the accuracy and reliabil-
ity of SPECT. Such an evaluation is also
required to justify exposing patients to the
additional radiation (4 mSv per exposure)
of the SPECT scan.
The aims of this study were to deter-
mine the sensitivity and specificity of
Diagnostic accuracy of SPECT 471
SPECT bone scintigraphy when used in
patients with facial asymmetry arising
from condylar hyperplasia, and to identify
general patterns of condylar growth in the
study population according to sex and age
groups.
Materials and methods
Patients with mandibular asymmetry
caused by condylar hyperplasia, who pre-
sented to the Discipline of Oral and Max-
illofacial Surgery, Faculty of Dentistry,
The University of Hong Kong between
January 2011 and July 2013, and who
underwent SPECT bone scintigraphy for
pre-treatment assessment, were recruited.
Patients who fulfilled the following cri-
teria were included: (1) mandibular asym-
metry as a result of condylar hyperplasia;
(2) availability of lateral and postero-an-
terior cephalometric radiographs taken
around the time of SPECT bone scintigra-
phy and 1 year afterwards.
Patients with the following conditions
were excluded: (1) history of temporo-
mandibular joint surgery; (2) presence
of a neoplastic pathology of the temporo-
mandibular joint, e.g. osteochondroma;
(3) presence of systemic diseases that
could potentially affect the temporoman-
dibular joint, e.g. autoimmune conditions;
(4) congenital conditions and syndromes
that are associated with facial asymmetry,
e.g. hemifacial microsomia.
SPECT bone scintigraphy and analysis
SPECT scans were performed 2–4 h after
the intravenous administration of 20–25
mCi technetium 99m-methylene dipho-
sphonate (99mTc-MDP). A region of inter-
est (ROI) of 16 pixels  3 slices was
drawn over both condylar heads. The in-
dividual condylar counts within the ROIs
were calculated and expressed as an aver-
age against the clivus bone count.
The determination of active condylar
growth from the condylar counts was
assessed using three reported methods.
The first was the absolute difference in
uptake (i.e. active inactive)25, in which
the condylar count of the inactive condyle
was subtracted from that of the active
condyle. A difference in scan values of
0.34  0.40 (mean  standard deviation
(SD)) between the active and inactive
condyles was considered to indicate
active growth. The second method was
the uptake ratio, i.e. (active/inactive) 
100%5,16,26, in which the ratio of the
active condylar count against the inactive
condylar count was calculated as a per-
centage. Growth was considered to be
472 Chan and Leung

active if the percentage was more than
10%, and vice versa. The third method
was the relative uptake (condylar
activity percentage), i.e. (active/(active
+ inactive))  100%5,10,16,19,27, in which
the active condylar count over the total
condylar count was calculated as a per-
centage. Growth was considered to be
active if the uptake value percentage
was 55% or more.
Serial radiographic tracings and growth
assessment
Lateral and postero-anterior (PA) cepha-
lometric radiographs (magnification
1.23, Orthoralix 9200 X-ray system;
Gendex, Hatfield, Pennsylvania, USA)
were traced using an overlay ortho/
trace film 0.003 matte acetate for the
baseline radiographs taken at the time
of the SPECT scan. The lateral and PA
cephalometric radiographs that were taken
1 year after were then traced and com-
pared to the baseline radiographs of the
same patient. Stable anatomical land-
marks on the radiographs were traced.
For the lateral cephalometric radio-
graphs, the base of the skull, orbital rims,
and the condylar head were marked as the
reference landmarks. To determine the
presence of growth, the posterior and in-
ferior borders of the mandible and the chin
region were traced. Any visible growth in
the vertical and anteroposterior direction
was detected and measured from the men-
ton and pogonion, respectively, with a
clear plastic ruler with 0.5-mm-interval
markings (Fig. 1).
For the PA cephalometric radiographs,
a horizontal reference plane (x-axis) was
drawn between the right and left lateral
orbitale (LO) landmarks. A vertical line
(y-axis) was dropped perpendicular to the
horizontal plane using the crista galli as
the reference point, towards the inferior
border of the anterior mandible. The dis-
tance on the x-axis from menton to the
vertical line was measured and compared
to detect any changes in the transverse
direction (Fig. 2).
Additional comparisons were per-
formed for patients with clinical photo-
graphs available that were taken at the
time of the SPECT scan and a year later.
Intraoral photographs of the same patient
taken at a 1-year interval were also com-
pared. The frontal, profile, and intraoral
photographs were overlapped using Dol-
phin Imaging software (version 11.9.07.24
Premium; Dolphin Imaging and Manage-
ment Solutions, Chatsworth, CA, USA)
for correlation with the findings from
the serial radiographic tracings (Fig. 3).
Growth was considered to be active if
any of the three assessments (lateral ceph-

Fig. 1. Tracing of a lateral cephalometric radiograph.

Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.

Fig. 2. Comparison of transverse growth on postero-anterior cephalometric radiographs. LO,
lateral orbitale; CG, crista galli; Me, menton.
alometric serial tracing, PA cephalometric
serial tracing, and clinical photographic
changes) showed any changes. If none
of these assessments showed any changes
across the 1-year interval, the subject was
considered to be inactive for growth.
Comparison of SPECT bone scintigraphy
and serial growth assessment
The results from SPECT bone scintigra-
phy were compared to the growth status
results determined in the serial growth
assessment. For each patient, the results
for the three SPECT analysis methods
were compared against the findings of
the serial growth assessment. The numbers
of true-positives (where a positive SPECT
finding corresponded with a positive serial
growth status) and true-negatives (where a
negative SPECT finding corresponded
with a negative serial growth status) were
calculated. The sensitivity was calculated
as the percentage of true-positives over the
total positive results for each SPECT anal-
ysis method, and the specificity was cal-
Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.
culated as the percentage of true-negatives
over the total negative results.
Statistical analysis
The statistical analysis was performed
using IBM SPSS Statistics version 23.0
(IBM Corp., Armonk, NY, USA). De-
scriptive statistics, including the
mean  SD, were used to describe the
demographic background of the samples.
The mean movements of menton and
pogonion for patients with active growth
were also obtained.
Significant differences in growth be-
tween sex and age groups were determined
using the parametric t-test for data follow-
ing a normal distribution, and the non-
parametric Mann–Whitney U-test for data
that did not. Significant differences in
proportions of patients with active growth
according to sex and age groups were
calculated with the Pearson x2 test. A
5% level of significance was applied.
The sensitivity and specificity of the
three assessment methods were calculated
Diagnostic accuracy of SPECT 473
against the findings of the serial growth
assessment. Receiver operating character-
istics (ROC) curves were also plotted to
compare the different methods, and the
method that had a curve closest to the
upper left quadrant had the best diagnostic
ability. The area under the curve (AUC)
was also calculated for each method.
Results
A total of 265 patients were initially
recruited into the study, of whom 65 were
excluded based on the exclusion criteria.
Of the 200 patients included, 82 (41%)
were male and 118 (59%) were female,
giving a male to female ratio of approxi-
mately 1:1.43. The mean age of the
patients was 21.0  4.3 years; the mean
age of male patients was 21.7  4.0 years
and of female patients was 20.6  4.4
years. The general demographic data of
the 200 patients are presented in Table 1.
Serial tracings of the scans of 200
patients were analyzed, and 34 patients
(17%) were found to have active growth in
at least one of the three measured dimen-
sions; of these patients, 21 (61.8%) were
female and 13 (38.2%) were male.
The SPECT scans were analyzed and
the findings are presented in Table 2. With
regard to the absolute difference in uptake
measurement, 47 patients were found to
have active growth. When the uptake ratio
was calculated, 129 patients were found to
have active growth, and when relative
uptake was measured, 68 patients had
active growth. There were statistically
significant differences (P < 0.001) be-
tween the proportions of active scan
results between the three methods of
SPECT scan measurement, indicating
poor validity amongst them.
The findings of the SPECT scans were
compared to those from serial growth
assessment, as illustrated in Table 3. Each
SPECT analysis method was compared
against the serial growth assessment,
and the numbers of true active and inactive
findings were determined. The absolute
difference in uptake measurement identi-
fied 11 true active patients, which was the
lowest result among the three methods.
The uptake ratio identified 23 true active
patients and the relative uptake identified
14 true active patients.
When the true inactive findings were
compared, the uptake ratio identified the
lowest number of true inactive findings
with 60 patients; the absolute difference in
uptake identified 130 true inactive patients
and relative uptake identified 112 true
inactive patients.
474 Chan and Leung
Fig. 3. Comparison of clinical photographs. (a) Frontal photographs showing vertical and
transverse growth. (b) Lateral photographs showing vertical growth at menton. (c) Frontal
intraoral photographs of the dentition, using the upper anterior teeth as reference; growth is
noted from the difference in lower anterior teeth positions.
The sensitivity and specificity of the
three methods are presented in Table 4.
The absolute difference in uptake yielded
the lowest sensitivity at 32.4% and the
uptake ratio had the highest sensitivity
at 67.6%; the sensitivity of relative uptake
was 41.2%. With regard to specificity, the
uptake ratio had the lowest specificity at
36.1%, relative uptake had a specificity of
Table 1. Patient demographic characteristics.
Number (%)
Male 82 (41)
Female 118 (59)
Total 200
SD, standard deviation.
Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.
was only slightly more than random
chance. An AUC of 0.50 is equivalent
to random chance.
Based on the findings of the primary
outcome of sensitivity and specificity, and
supported by the AUC calculations, the
three SPECT analysis methods proposed
are not suitable to be used as a diagnostic
tool for condylar hyperplasia.
Further analyses of the findings from
the serial growth assessment were per-
formed; the results are presented in Table
5. Thirty-four of the total 200 patients
(17%) had active growth. The average
age of the patients with true active growth
as determined from the serial growth as-
sessment was 18.6  2.8 years. The aver-
age age of the male patients with active
growth was 19.2  2.9 years, and the av-
erage age of the female patients with
active growth was 18.3  2.7 years. There
was no statistically significant difference
in the proportion of patients with active
growth between the male and female
patients (P = 0.719).
Further stratification of the sample pop-
ulation by age groups was done (Table 6).
Of the 34 patients with active growth, 10
were younger than 17 years old. Eleven
patients were aged 17 to <19 years, seven
were between 19 and <21 years old, and
six were 21 years old. Patients below the
age of 19 years accounted for more than
half (61.7%) of the patients with active
growth. The age group of <17 years had
the highest percentage of patients with
active growth (47.6%) when compared
to the other age groups; correspondingly,
the age group 21 years had the lowest
percentage of patients with active growth
(7.6%). A general trend was observed, in
which approximately half of the patients
below 17 years of age had active growth.
This percentage then further halved in
each subsequent age group, with 7.6%
67.5%, and the absolute difference in up- of patients with active growth in the
take had the highest specificity at 78.3%. 21 years age group (Fig. 5).
The sensitivity and specificity of all three There were significant differences in the
methods are considered to be low. proportions of patients with active growth
The ROC curves for the three methods in those <17 years of age compared to the
are presented in Fig. 4. All three methods other three age groups. A significant dif-
had low AUC values of less than 0.58, ference was also found between the 17 to
indicating that the ability of the three <19 years group and the 21 years group
SPECT analysis methods to detect growth (P = 0.021). There was no other difference
in the proportions of patients with active
growth for the other comparisons.
The amounts of growth in the antero-
posterior, vertical, and transverse direc-
Age (years), Active growth, tions in the 34 patients with active growth
mean  SD number (%) were recorded; the results are presented
21.7  4.0 13 (38.2) according to sex in Table 7. The average
20.6  4.4 21 (61.8) growth was 0.94  0.91 mm in the ante-
21.0  4.3 34 roposterior direction, 0.88  0.86 mm in
the vertical direction, and 1.4  0.66 mm
 Diagnostic accuracy of SPECT 475

Table 2. Analysis of SPECT scans. A further analysis of the amount of

Method Number (%) growth exhibited by true active growth
patients was performed by stratifying
Growth status from serial growth assessment the patients into the age ranges. The
Active 34 (17)
findings are presented in Table 8. Patients
Inactive 166 (83)
<19 years of age appeared to have higher
Absolute difference in uptake yearly growth amounts as compared to
Active 47 (23.5) patients aged 19 years. However, there
Inactive 153 (76.5) was no significant difference between the
Uptake ratio mean anteroposterior (P = 0.169), vertical
Active 129 (64.5) (P = 0.552), and transverse (P = 0.193)
Inactive 71 (35.5) growth between the two age groups.
Relative uptake
Active 68 (34)
Discussion
Inactive 132 (66)
The key findings of this study were (1) the
Differences in proportion of active counts
sensitivity of the three analysis methods of
Absolute difference versus uptake ratio P < 0.001a,* SPECT bone scintigraphy to determine
Absolute difference versus relative uptake P < 0.001a,* condylar growth ranged between 32.4%
Uptake ratio versus relative uptake P < 0.001a,* and 67.6%, and the specificity between
*
Significant difference (P < 0.05). 36.1% and 78.3%; and (2) the AUC of
a
Pearson x2 test. the ROC curves was low at <0.58. Both
key findings show that SPECT scans were
Table 3. Comparison of SPECT scans and serial growth assessment. poor for determining condylar growth sta-
tus. The proportions of patients with active
Serial growth assessment
condylar growth in the age groups <17
SPECT assessment Active Inactive Total years, 17 to <19 years, 19 to <21 years,
Absolute difference in uptake Active 11 36 47 and 21 years were 47.6%, 21.6%,
Inactive 23 130 153 14.3%, and 7.6%, respectively.
Total 34 166 200 The treatment of patients with condylar
hyperplasia is determined by the progres-
Uptake ratio Active 23 106 129
Inactive 11 60 71
sion of the individual’s condylar growth.
Total 34 166 200 There are advantages and disadvantages to
both commencing treatment earlier and
Relative uptake Active 14 54 68 adopting a watch-and-wait approach in
Inactive 20 112 132 the management of such patients. The
Total 34 166 200
condition tends to present during the
patient’s adolescence period and may con-
tinue even after the cessation of skeletal
in the transverse direction. For the the amount of anteroposterior growth bone growth. This gradual dentofacial
male patients, the average anteroposterior, between male and female patients asymmetry and its attendant dental mal-
vertical, and transverse growth was (P = 0.039), with male patients exhibiting occlusion will no doubt result in some
1.3  1.1 mm, 1.2  1.1 mm, and 1.3  significantly more anteroposterior growth negative impact on the patient’s self-
0.43 mm, respectively. For the female than female patients. Analysis of the image and confidence.
patients, the average anteroposterior vertical and transverse growth between The patients in this study presented with
change was 0.69  0.68 mm, vertical the sexes did not reveal any significant the condition at a young age – at an
change was 0.69  0.62 mm, and trans- difference (P = 0.232 and P = 0.753, re- average of 21.0  4.3 years. This is in
verse change was 1.5  0.78 mm. spectively). line with the findings reported by other
There was a significant difference in
Table 5. Serial growth assessment.
Factor Number (%)
Table 4. Sensitivity and specificity of SPECT
Growth status
methods.
Active 34 (17)
Number (%) Inactive 166 (83)
Absolute difference in uptake Patients with active growth
True-positive (sensitivity) 11 (32.4) Male 13 (38.2)
True-negative (specificity) 130 (78.3) Female 21 (61.8)
Uptake ratio Difference in proportion of active growth between the sexes P = 0.719a
True-positive (sensitivity) 23 (67.6) Mean age of patients with active growth, mean  SD
True-negative (specificity) 60 (36.1) Total: 18.6  2.8 years
Male: 19.2  2.9 years
Relative uptake Female: 18.3  2.7 years
True-positive (sensitivity) 14 (41.2)
SD, standard deviation.
True-negative (specificity) 112 (67.5) a
Pearson x2 test.

Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.
476 Chan and Leung

Table 6. Age group characteristics—serial growth assessment. patients. There were also more females
Patients with active growth by age Number (%) within the active growth group, with 21 of
the 34 patients (61.8%) being female.
<17 years 10 (29.4) As determined in this study, the three
Male 3
SPECT methods had sensitivity between
Female 7
32.4% and 67.6%, and specificity between
17 to <19 years 11 (32.4) 36.1% and 78.3%. Further analysis using
Male 4 ROC curves revealed AUC values of
Female 7 <0.58, indicating that the methods are
19 to <21 years 7 (20.6) no better than random chance. This was
Male 4 corroborated by statistically significant
Female 3 differences between the findings of the
21 years 6 (17.6) three methods, indicating poor validity
Male 2 and accuracy. Such low sensitivity and
Female 4 specificity values would result in a large
percentage of false-positive and false-neg-
Proportion of patients with active growth in the age group n/N (%)
ative diagnoses, which would hinder and
<17 years 10/21 (47.6) potentially negatively affect the treatment
17 to <19 years 11/51 (21.6) outcomes of patients.
19 to <21 years 7/49 (14.3)
Two main treatment methods are cur-
21 years 6/79 (7.6)
rently proposed and practiced, namely
Differences in proportion of active counts high condylectomies and orthognathic
<17 years vs. 17 to <19 years P = 0.027a,* surgery, or a combination of both6,28,29.
<17 years vs. 19 to <21 years P = 0.003a,* These two methods are highly dependent
<17 years vs. 21 years P < 0.001a,* on the clinician’s ability to accurately
17 to <19 years vs. 19 to <21 years P = 0.343a determine the patient’s growth status, as
17 to <19 years vs. 21 years P = 0.021a,* high condylectomies are used to arrest
19 to <21 years vs. 21 years P = 0.223a
further growth of the affected condyle
*
Significant difference (P < 0.05). and orthognathic surgery should only be
a
Pearson x2 test. done after the cessation of growth.
However, with such low sensitivity and
specificity values, the use of SPECT scans
authors7,8. The average age of females was was younger, at 18.6  2.8 years. This is would result in a number of false-positive
generally younger than male patients by 1 approximately 3 years lower than the and false-negative cases, which would
year. This indicates that the condition is overall average. Females with active lead to unnecessary over-treatment (high
already present at an earlier age, as the growth also generally presented 1 year condylectomy or unnecessary growth
majority of the patients were already earlier than the male patients. reviews) for patients who are false-posi-
found to have ceased growing when This study also supports the higher in- tive, or premature commencement of
assessed (166 out of 200). The average cidence of the condition in females, as treatment (orthognathic surgery) for
age of the 34 patients with active growth 59% of the scans were done for female patients who are false-negative. Another

Fig. 4. Receiver operating characteristic (ROC) curves for SPECT. AUC, area under the curve; CI, confidence interval.

Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.
 Diagnostic accuracy of SPECT 477

Fig. 5. Percentage of patients with active growth in each age group, as assessed using serial radiographs.

Table 7. Growth in the anteroposterior, vertical, and transverse direction according to sex; mean major drawback of using SPECT is the
 SD values. need to expose patients to the highly ra-
Anteroposterior Vertical Transverse dioactive 99mTc-MDP intravenous con-
trast during scans. The effective dose of
Growth in males 1.3  1.1 mm 1.2  1.1 mm 1.3  0.43 mm
a single SPECT scan is 4 mSv
Growth in females 0.69  0.68 mm 0.69  0.62 mm 1.5  0.78 mm
Average growth 0.94  0.91 mm 0.88  0.86 mm 1.4  0.66 mm (4000 mSv), which is the equivalent of
200 chest X-rays (0.02 mSv)30, or 1.8
Differences according to sex years of background radiation. In contrast,
Anteroposterior growth between males and females P = 0.039a,* a panoramic radiograph has an effective
Vertical growth between males and females P = 0.232b dose of only 36 mSv, and a cone beam
Transverse growth between males and females P = 0.753b computed tomography (CBCT) scan only
SD, standard deviation. 432 mSv31,32.
*
Significant difference (P < 0.05). Wu et al. performed a study to predict
a the cancer risks from dental computed
Parametric t-test.
b
Non-parametric Mann–Whitney U-test performed for data that did not follow a normal tomography and reported that the risk of
distribution. cancer induction was higher if the exposed
patient was younger, and also that females
were at a higher risk than males32. This
could be extrapolated to the use of 99mTc-
Table 8. Growth in the anteroposterior, vertical, and transverse direction according to age; MDP for the purpose of determining con-
mean  SD values. dylar growth activity. SPECT scans for the
Anteroposterior Vertical Transverse purpose of condylar growth determination
are typically done on younger patients,
<17 years 1.0  1.0 mm 0.9  0.77 mm 1.7  0.70 mm
and with the higher prevalence of condylar
17 to <19 years 1.2  1.0 mm 0.86  0.67 mm 1.0  0.38 mm
19 to <21 years 0.42  0.53 mm 1.4  1.3 mm 1.5  0.91 mm hyperplasia in females, the majority of
21 years 0.83  0.6 mm 0.33  0.4 mm 1.7  0.4 mm patients being exposed to radiation would
be young females.
Difference according to age group A more worrying trend that was noted
Anteroposterior growth between <19 years and 19 P = 0.169a during this study was the taking of the raw
years condylar scan results at face value, which
Vertical growth between <19 years and 19 years P = 0.552a led to the diagnosis of a condyle with a
Transverse growth between <19 years and 19 years P = 0.193a higher count as active. Such diagnoses of
SD, standard deviation. active condylar growth were found to be
a
Non-parametric Mann–Whitney U-test performed for data that did not follow a normal prevalent during the reporting of the
distribution. SPECT scans results. A small number of

Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.
478 Chan and Leung
patients were also noted to have been
subjected to repeated SPECT scans during
their follow-up period, and it is the
authors’ view that they were unnecessary.
The main limitation of this study was
the retrospective nature of the research.
This limitation was overcome by the use
of a well-developed treatment protocol
and a detailed records system for patients
who required skeletal growth monitoring
in the unit. The findings of this study,
however, suggest that it would be unethi-
cal to design a prospective study with
similar aims and objectives, as the SPECT
scans were proved to have low sensitivity
and specificity. Also, with the high dosage
of radiation exposure for relatively young
individuals, it might be more valuable to
design prospective studies to investigate
three-dimensional mandibular growth pat-
terns using advanced imaging techniques
with a lower radiation exposure, such as
CBCT. Other limitations of this study
were that it was conducted at a single
centre with a sample population compris-
ing mainly patients of Asian descent. The
study also only included patients with
facial asymmetry caused by condylar hy-
perplasia. Therefore, the extrapolation of
the findings from this study may not be
wholly accurate when applied to a differ-
ent geographical or racial group.
Several considerations regarding the
determination of condylar growth activity
can be summarized from the findings of
this study: (1) females were found to
present with condylar hyperplasia at an
earlier age, and correspondingly ceased
their growth activity at an earlier age.
(2) Male patients in general were found
to present with condylar hyperplasia later
than female patients and ceased their
growth activity later. It would therefore
be prudent not to commence definitive
orthognathic surgery too early in male
patients if condylar hyperplasia is noted.
(3) Approximately half of the patients
under the age of 17 years were found to
have active condylar growth. This propor-
tion halved in each subsequent age group,
until only approximately 7% of patients
21 years of age were found to have
active condylar growth. (4) The average
change at the chin point over 1 year in the
anteroposterior, vertical, and transverse
dimensions was 0.94 mm, 0.88 mm, and
1.4 mm, respectively, with male patients
growing almost twice as much as female
patients in the anteroposterior and vertical
dimensions. The period of highest growth
was found for the age group <19 years.
In conclusion, when compared to serial
growth assessment using cephalometric
radiographs and clinical photographs,
Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.
the sensitivity of SPECT bone scintigra-
phy to determine condylar growth was
32.4–67.6%, and the specificity to detect
growth cessation was 36.1–78.3%. As a
diagnostic tool, SPECT bone scintigraphy
did not achieve an acceptable sensitivity
or specificity, and it is therefore not justi-
fiable to use this tool for the routine as-
sessment of growth in patients with
condylar hyperplasia.
Funding
No sources of funding were obtained for
this research.
Competing interests
No competing interests identified.
Ethical approval
Ethical approval was obtained from the
Institutional Review Board of the Univer-
sity of Hong Kong/Hospital Authority
Hong Kong West Cluster. This study
has been registered with the HKU Clinical
Trials Registry (Study Identifier
HKUCTR-2127).
Patient consent
Patient consent was obtained for the use of
the clinical photographs in Fig. 3.
Acknowledgements. The authors would
like to acknowledge the kind support
and advice from Ms Samantha Li over
the duration of the study.
References
1. Thiesen G, Gribel BF, Freitas MP. Facial
asymmetry: a current review. Dental Press
J Orthod 2015;20:110–25.
2. Hovinga J, Kraal ER, Roorda LA. Difficul-
ties in and indications for the treatment of
facial asymmetry. Int J Oral Surg
1974;3:234–8.
3. Olate S, Netto HD, Rodriguez-Chessa J,
Alister P, de Albergaria-Barbosa J, de Mor-
aes M. Mandible condylar hyperplasia: a
review of diagnosis and treatment protocol.
Int J Clin Exp Med 2013;6:727–37.
4. Saridin CP, Raijmakers PG, Tuinzing DB,
Becking AG. Comparison of planar bone
scintigraphy and single photon emission
computed tomography in patients suspected
of having unilateral condylar hyperactivity.
Oral Surg Oral Med Oral Pathol Oral
Radiol Endod 2008;106:426–32.
5. Pogrel MA. Quantitative assessment of iso-
tope activity in the temporomandibular joint
regions as a means of assessing unilateral
condylar hypertrophy. Oral Surg Oral Med
Oral Pathol 1985;60:15–7.
6. Rodrigues DB, Castro V. Condylar hyperpla-
sia of the temporomandibular joint: types,
treatment, and surgical implications. Oral
Maxillofac Surg Clin North Am 2015;27:
155–67.
7. Obwegeser HL. Mandibular growth anoma-
lies: terminology, aetiology, diagnosis, treat-
ment. Heidelberg, Germany: Springer; 2001.
8. Nitzan DW, Katsnelson A, Bermanis I, Brin
I, Casap N. The clinical characteristics of
condylar hyperplasia: experience with 61
patients. J Oral Maxillofac Surg 2008;66:
312–8.
9. Raijmakers PG, Karssemakers LH, Tuinzing
DB. Female predominance and effect of
gender on unilateral condylar hyperplasia:
a review and meta-analysis. J Oral Maxillo-
fac Surg 2012;70:e72–6.
10. Hodder SC, Rees JI, Oliver TB, Facey PE,
Sugar AW. SPECT bone scintigraphy in the
diagnosis and management of mandibular
condylar hyperplasia. Br J Oral Maxillofac
Surg 2000;38:87–93.
11. Obwegeser HL, Makek MS. Hemimandibu-
lar hyperplasia—hemimandibular elonga-
tion. J Maxillofac Surg 1986;14:183–208.
12. Wolford LM, Movahed R, Perez DE. A
classification system for conditions causing
condylar hyperplasia. J Oral Maxillofac
Surg 2014;72:567–95.
13. Wen B, Shen Y, Wang CY. Clinical value of
99Tcm-MDP SPECT bone scintigraphy in
the diagnosis of unilateral condylar hyper-
plasia. Sci World J 2014;2014:256256.
14. Naini FB, Donaldson AN, McDonald F,
Cobourne MT. Assessing the influence of
asymmetry affecting the mandible and chin
point on perceived attractiveness in the
orthognathic patient, clinician, and layper-
son. J Oral Maxillofac Surg 2012;70:192–
206.
15. Fahey FH, Abramson ZR, Padwa BL, Zim-
merman RE, Zurakowski D, Nissenbaum M,
Kaban LB, Treves ST. Use of (99m)Tc-MDP
SPECT for assessment of mandibular
growth: development of normal values.
Eur J Nucl Med Mol Imaging 2010;37:
1002–10.
16. Lo J, Yau YY, Yeung DWC, Cheung LK.
Planar scintigraphy in assessment of man-
dibular asymmetry: Unilateral condylar
hyperplasia vs asymmetric mandibular hy-
perplasia. J Biomed Sci Eng 2012;5: 836–44.
17. Kaban LB, Cisneros GJ, Heyman S, Treves
S. Assessment of mandibular growth by
skeletal scintigraphy. J Oral Maxillofac Surg
1982;40:18–22.
18. Melsen B. The cranial base: the postnatal
development of the cranial base studied his-
tologically on human autopsy material. Acta
Odontol Scand 1974;32(Supp):62.
19. Saridin CP, Raijmakers PG, Al Shamma S,
Tuinzing DB, Becking AG. Comparison of
different analytical methods used for analyz-
ing SPECT scans of patients with unilateral

condylar hyperactivity. Int J Oral Maxillofac
Surg 2009;38:942–6.
20. Kim JH, Kim YK, Kim SG, Yun PY, Kim JD,
Min JH. Effectiveness of bone scans in the
diagnosis of osteoarthritis of the temporo-
mandibular joint. Dentomaxillofac Radiol
2012;41:224–9.
21. Epstein JB, Rea A, Chahal O. The use of
bone scintigraphy in temporomandibular
joint disorders. Oral Dis 2002;8:47–53.
22. Almeida LE, Zacharias J, Pierce S. Condylar
hyperplasia: an updated review of the liter-
ature. Korean J Orthod 2015;45:333–40.
23. Ahn BC, Kim HJ, Lee SW, Yoo J, Choi JK,
Lee J. New quantitative method for bone
tracer uptake of temporomandibular joint
using Tc-99m MDP skull SPECT. Ann Nucl
Med 2009;23:651–6.
24. Saridin CP, Raijmakers PG, Tuinzing DB,
Becking AG. Bone scintigraphy as a diag-
nostic method in unilateral hyperactivity of
the mandibular condyles: a review and meta-
analysis of the literature. Int J Oral Max-
illofac Surg 2011;40:11–7.
25. Pogrel MA, Kopf J, Dodson TB, Hattner R,
Kaban LB. A comparison of single-photon
emission computed tomography and planar
Descargado para Anonymous User (n/a) en Pontifical Xavierian University de ClinicalKey.es por Elsevier en febrero 06, 2021.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2021. Elsevier Inc. Todos los derechos reservados.
imaging for quantitative skeletal scintigra-
phy of the mandibular condyle. Oral Surg
Oral Med Oral Pathol Oral Radiol Endod
1995;80:226–31.
26. Robinson PD, Harris K, Coghlan KC, Alt-
man K. Bone scans and the timing of treat-
ment for condylar hyperplasia. Int J Oral
Maxillofac Surg 1990;19:243–6.
27. Cisneros GJ, Kaban LB. Computerized skel-
etal scintigraphy for assessment of mandib-
ular asymmetry. J Oral Maxillofac Surg
1984;42:513–20.
28. Wolford LM, Morales-Ryan CA, Garcia-
Morales P, Perez D. Surgical management
of mandibular condylar hyperplasia type 1.
Proc (Bayl Univ Med Cent) 2009;22:321–9.
29. Lippold C, Kruse-Losler B, Danesh G, Joos
U, Meyer U. Treatment of hemimandibular
hyperplasia: the biological basis of condy-
lectomy. Br J Oral Maxillofac Surg
2007;45:353–60.
30. Wall BF, Hart D. Revised radiation doses for
typical x-ray examinations. Report on a re-
cent review of doses to patients from medical
X-ray examinations in the UK by NRPB.
National Radiological Protection Board. Br J
Radiol 1997;70:437–9.
Diagnostic accuracy of SPECT 479
31. Granlund C, Thilander-Klang A, Ylhan B,
Lofthag-Hansen S, Ekestubbe A. Absorbed
organ and effective doses from digital intra-
oral and panoramic radiography applying
the ICRP 103 recommendations for effec-
tive dose estimations. Br J Radiol
2016;89:20151052.
32. Wu TH, Lin WC, Chen WK, Chang YC,
Hwang JJ. Predicting cancer risks from den-
tal computed tomography. J Dent Res
2015;94:27–35.
Address:
Yiu Yan Leung
Discipline of Oral and Maxillofacial Surgery
Faculty of Dentistry
The University of Hong Kong
Prince Philip Dental Hospital
34 Hospital Road
Sai Ying Pun
Hong Kong
Tel.: +852 2859 0511
Fax: +852 2857 5570
E-mail: mikeyyleung@hku.hk