Rev Esp Med Nucl Imagen Mol.

2019;38(2):72–79

Original Article

Comparison of four methods for quantitative assessment of

99m
Tc-MDP SPECT in patients with suspected condylar hyperplasia夽
Diego F. López Buitrago a,∗ , Juan M. Muñoz Acosta b , Rodrigo A. Cárdenas-Perilla b
a
Orthodontics Graduate Program, Universidad del Valle, Cali, Colombia
b
Nuclear Medicine Department, Centro Médico Imbanaco, Cali, Colombia

a r t i c l e i n f o a b s t r a c t

Article history: Objective: To compare variability, reproducibility and repeatability of four quantitative evaluation meth-
Received 12 June 2018 ods to interpret the 99m Tc-MDP SPECT reports in patients with clinically suspected unilateral condylar
Accepted 23 July 2018 hyperplasia (UCH).
Available online 1 November 2018
Method: This was a descriptive observational study carried out with SPECT images of 38 patients with
clinical and radiographic signs of UCH, and interpreted using four quantitative methods: (1) one image,
Keywords: variable-size region of interest (ROI); (2) one image, fixed-size ROI (1.76 cm2 ), (3) five image variable-
Condylar hyperplasia
size ROI; (4) five image, fixed-size ROI (1.76 cm2 ). Each of the images were report simultaneously but
Facial asymmetry
99m
Tc-MDP
in an independent way by the two nuclear medicine experts in both total radioactive counts as well
Bone SPECT as normalized counts to quantify the reproducibility (inter-operator variability) and the repeatability
(intra-operator variability).
Results: Higher reproducibility and repeatability were obtained in 5-image fixed-size ROI method (intra-
class correlation coefficient: 0.979 [0.959; 0.989]). A high grade of diagnostic agreement (97.4%) was also
attained in fixed methods (Kappa 0.940, p value: 0.000) from either total or normalized counts. There
was no difference between fixed-size 1 vs 5 image methods. The methods based on variable-size ROI
had a low grade of agreement (Kappa < 0.20). More positive cases were identified using one image, ROI
variable total counts (27 cases), but when the counts were normalized, they presented a lower number
(5 cases).
Conclusion: Five-image fixed-size ROI provides the best intra-operator and inter-operator reliability for
the diagnosis of unilateral condylar hyperplasia. In the four methods using normalized counts fewer
positive cases were detected (≥10%), unlike with total counts when more positive cases were found.
© 2018 Sociedad
Española de Medicina Nuclear e Imagen Molecular. Published by Elsevier España, S.L.U. All rights reserved.

Comparación de cuatro métodos de valoración cuantitativa del SPECT con

99m
Tc-MDP en pacientes con sospecha clínica de hiperplasia condilar

r e s u m e n

Palabras clave: Objetivo: Comparar la variabilidad, reproducibilidad y repetitividad de cuatro métodos de evaluación
Hiperplasia condilar cuantitativa del SPECT con 99m Tc-MDP en pacientes con sospecha clínica de hiperplasia condilar unilateral
Asimetría facial (HCU).
99m
Tc-MDP
Métodos: Estudio observacional descriptivo, realizado en 38 imágenes SPECT de pacientes con signos
SPECT óseo
clínicos y radiográficos de HCU. Estos fueron interpretados utilizando cuatro métodos cuantitativos de la
prueba SPECT: (1) Región objeto de interés (ROI) variable de 1 imagen, (2) ROI fija de 1,76cm2 en 1 imagen,
(3) ROI variable de 5 imágenes y (4) ROI fija de 1,76cm2 en 5 imágenes. Cada una de las imágenes fue
leída simultáneamente de forma independiente por dos médicos nucleares tanto en sus cuentas totales
como en las cuentas normalizadas para cuantificar la reproducibilidad (variabilidad inter-observador) de
los métodos y la repetitividad (variabilidad intra-observador).
Resultados: Mayor reproducibilidad y repetitividad fue obtenida con el método de ROI fija de 1.76cm2 en
5 imágenes (coeficiente de correlación intra-clase de 0,979 [0,959; 0,989]). De igual forma se obtuvo un
mayor grado de correspondencia diagnóstica (97,4%) en los métodos fijos (Kappa 0,940, valor p: 0.000)
tanto en cuentas totales como normalizadas. No se encontraron diferencias entre ellos. Los métodos
basados en el análisis de ROIs variables presentaron un menor grado de correspondencia (Kappa < 0,20).
El mayor número de casos positivos fue identificado utilizando una ROI variable de 1 imagen en cuentas
totales (27 casos), pero al normalizar las cuentas presentaron el menor número (5 casos).

夽 Please cite this article as: López Buitrago DF, Muñoz Acosta JM, Cárdenas Perilla RA. Comparación de cuatro métodos de valoración cuantitativa de 99mTC-MDP SPECT
en pacientes con sospecha clínica de hiperplasia condilar. Rev Esp Med Nucl Imagen Mol. 2019;38:72–79.
∗ Corresponding author.
E-mail address: dr.diegolopez10@gmail.com (D.F. López Buitrago).

2253-8089/© 2018 Sociedad Española de Medicina Nuclear e Imagen Molecular. Published by Elsevier España, S.L.U. All rights reserved.

Downloaded for Vanessa Ledesma (vl576@mynsu.nova.edu) at Nova Southeastern University from ClinicalKey.com by Elsevier on
September 27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 D.F. López Buitrago et al. / Rev Esp Med Nucl Imagen Mol. 2019;38(2):72–79 73

Conclusiones: De acuerdo a la variación intra e interoperadores, la técnica de mejor repetitividad y repro-

ducibilidad para el diagnóstico de la hiperplasia condilar unilateral fue la ROI fija de 5 imágenes. En los
cuatro métodos, al usar cuentas normalizadas, se detectaron menos casos positivos (≥10%) y con cuentas
totales mas casos positivos.
© 2018 Sociedad Española de
Medicina Nuclear e Imagen Molecular. Publicado por Elsevier España, S.L.U. Todos los derechos reservados.

Introduction Table 1
Definition of quantitative techniques.

Condylar hyperplasia (CH) is a pathology that affects facial Technique Definition

esthetics, masticatory function, occlusal stability and temporo- Variable-size ROI In a single transaxial slice with the highest
mandibular joint health as a consequence of the progressive facial 1 image condylar activity, a variable-size ROI for each
asymmetry due to overgrowth of one mandibular condyle.1,2 CH is condyle keeping the trace of the contour to
a self-limiting condition related to unbalanced growth and causing include total uptake was manually drawn
(Fig. 1)
facial deformations, more frequently before growth ends, but may
Fixed-size ROI 1 image In a single transaxial slice with the highest
be present after that.3,4 condylar activity, a fixed-size ROI (1.76 cm2 )
CH is frequently unilateral and more likely to occur in women on each condylar head was positioned (Fig. 2)
than in men according to some reports.5,6 Its etiology is not def- Variable-size ROI In a image obtain by the summed of five
5 image transaxial slices, a variable-size ROI for each
initely established but it has been related to genetic, traumatic,
condyle keeping the trace of the contour to
functional, tumor or hormonal factors.7,8 CH is classified in three include total uptake was manually draw.
categories: hemimandibular hyperplasia (HH) that is a vertical (Figs. 3 and 4)
growth asymmetry, hemimandibular elongation (HE) that is a hor- Fixed-size ROI 5 image In a image obtain by the summed of five
izontal growth asymmetry, and a combination of the above two transaxial slices with the highest condylar
activity, a fixed-size ROI (1.76 cm2 ) on each
types.1,9
condylar head was positioned. (Figs. 3 and 5)
CH diagnostic is based on extraoral/intraoral clinical findings as
well as radiographic/tomographic findings.3,10 Usually, this diagno- Following the institutional protocol, all the SPECT images
sis is confirmed by referring the patient for a nuclear medicine test were obtained 2 h after intravenous administration of 99m Tc-MDP
to determine if there is an increased osteoblast metabolic activity (555 MBq in patients over 18 year and normalized according to the
in one of the condyles, indicating that the pathology is active.11–16 Pediatric Dosage Card in patients under 18 year). The images were
Historically, the two most used nuclear medicine tests for ®
obtained with a dual-head gamma-camera (GE Infina ) using low
CH are: planar bone scintigraphy and single photon emission energy, high resolution planar collimators, a 128 × 128 matrix, and
computed tomography (SPECT). Both use 99m Tc methylene- zoom 1.3. Over 360◦ , 45 images of 18 s each were acquired for every
diphosphonate (MDP) as radioactive tracer. Many studies found 180◦ .
that SPECT is more efficient than planar bone scintigraphy in The data were reconstructed in a process station Xeleris 3, with
detecting active CH.6,17–21 There is no standard technique for the program “Volumetrix MI Evolution for Bone”, under iterative
the quantitative evaluation of SPECT results to assess a differ- reconstruction OSEM from 4 iterative steps and 8 subsets, applying
ence between condylar uptake of the radiotracer 99m Tc-MDP. It is a Butterworth 0.45 filter and power 12. From this reconstruction,
remarkable that the decision to perform condylectomy is almost trans-axial images were obtained to be used in the next four quanti-
exclusively based on a SPECT positive diagnosis of CH, that is tative techniques: variable-size ROI 1 image, fixed-size ROI 1 image,
defined by a percentage of uptake by the affected condyle equal variable-size ROI 5 image and fixed-size ROI 5 image, each section
or higher than 10% of the uptake detected in the contralateral with a constant width: 4.4 mm. Table 1 and Figs. 1–5 describe each
condyle.12,22–25 Therefore, it is very important to identify a method technique of evaluation.
of quantitative evaluation and interpretation of the test that offers The area to be measured in the fixed- size ROIs was standardized
the best possible CH diagnostic precision and accuracy. by the processor as 1.76 cm2 and for each technique, the maximum
Therefore, the objective of this study was to compare four counts, number of pixels and normalized counts were extracted.
quantitative techniques to interpret SPECT results in patients with Two nuclear medicine experts independently read twice all 38
clinically suspected CH active in terms of variability, reproducibility reports. Additionally, 15 SPECT images were evaluated by the same
and repeatability. The four methods are based on the combination observer 15 days after the first reading, to assess intra-observer
of condylar regions of interest (ROI) with a fixed-size of 1.76 cm2 or reproducibility.
variable-size ROI measured in either one or five images, measuring From the total counts and normalized counts, according to the
total counts or normalized counts. number of pixels, it was calculated the uptake for each condyle,
using the following equations:
Material and methods
Condyle counts
Number of normalized counts =
The group of patients in the study included 21 women (55%) and Number of pixels
17 men (45%) with a range of age between 11 and 44 (average: 21
year), referred for a SPECT test by their treating specialist to the % condylar uptake
Nuclear Medicine Department, between 2015 and 2017 due to a
Condyle counts
preliminary clinical diagnosis of facial asymmetry and clinically and = ∗ 100%
radiographic suspected UCH. Patients with a history of temporo- Condyle counts + Contralateral condyle counts
mandibular joint (TMJ) surgery, dentofacial syndromic anomalies,
craniofacial trauma or present pregnancy were excluded. Normalized counts are the total counts obtained for the selected
The study was approved by the Institutional Ethics Committee ROI divided by the number of pixels to provide a constant value for
(Approval code CEI-293). each pixel.

Downloaded for Vanessa Ledesma (vl576@mynsu.nova.edu) at Nova Southeastern University from ClinicalKey.com by Elsevier on
September 27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
74 D.F. López Buitrago et al. / Rev Esp Med Nucl Imagen Mol. 2019;38(2):72–79

Fig. 2. Fixed size ROI one image: The more active image was selected in each condyle
and a fixed-size (1.76 cm2 ) ROI was outlined on each condyle head to obtain maxi-
Fig. 1. ROI variable-size, one image: The image showing highest uptake was selected mum counts, pixel number and average counts.
in both condyles. A variable-size ROI was manually outlined, including the total
uptake area. Maximum counts, pixel number and average counts were obtained for
that area.

Statistical analysis
The average and standard deviation was obtained for each
measure. Normality of data distribution was evaluated by the
Shapiro–Wilk test. The four techniques were compared by one-way
ANOVA. Intra-observer repeatability was estimated by the intra-
class correlation coefficient (ICC) and its 95% confidence interval
(95% CI). Inter-observer reproducibility was estimated by Dahlberg
coefficient and ICC. Bland–Altman graphs were also obtained to
compare data variability. The diagnostic agreement for detection
of uptake percentage condylar differences ≥10% was calculated as
percentage and Kappa coefficient. The statistical analyses were car-
ried out using the program Stata 13 (Stata Corp., College Station, TX,
USA). The level of significance was 0.05.
Results
Reproducibility (inter-operator variability)
Table 2 summarizes the results for inter-operator differences.
Taking into account the 95% CI value, the method of reading in
5 images a fixed ROI provides less variability between operators,
followed by the method using 1 fixed image. The methods based
on variable ROI had greater variability. There are no differences
related to the mandibular deviation side. The Dahlberg S coeffi-
cient in fixed-size ROI, the method of reading in 5 images is almost
equal using total counts (0.90) or normalized counts (0.88), but
significantly lower than in variable-size ROIs, indicating better
inter-operator reproducibility.

Fig. 3. (A) The five transaxial slices with the highest condylar activity. (B) The summed image of the five transaxial slices.

Downloaded for Vanessa Ledesma (vl576@mynsu.nova.edu) at Nova Southeastern University from ClinicalKey.com by Elsevier on
September 27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 D.F. López Buitrago et al. / Rev Esp Med Nucl Imagen Mol. 2019;38(2):72–79 75

Table 2
Inter-operator reproducibility using total and normalized counts for each condyle.

Technique Total counts Normalized counts/pixels

Right condyle Left condyle Right condyle Left condyle

Variable-size ROI 1 image

Dahlberg S 2.30 2.30 2.07 2.07
[95% CI] 0.880 [0.959; 0.989] 0.880 [0.778; 0.937] 0.884 [0.788; 0.938] 0.884 [0.788; 0.938]

Fixed-size ROI 1 image

Dahlberg S 3.75 1.50 1.50 1.50
[95% CI] 0.856 [0.738; 0.923] 0.891 [0.797; 0.942] 0.891 [0.797; 0.943] 0.891 [0.797; 0.943]

Variable-size ROI 5 image

Dahlberg S 3.15 3.15 1.34 1.34
[95% CI] 0.810 [0.661; 0.898] 0.810 [0.661; 0.898] 0.837 [0.703; 0.913] 0.837 [0.703; 0.913]

Fixed-size ROI 5 image

Dahlberg S 0.90 0.90 0.88 0.88
[95% CI] 0.979 [0.959; 0.989] 0.979 [0.959; 0.989] 0.975 [0.953; 0.987] 0.975 [0.953; 0.987]

Table 3
Repeatability (intra-operator reproducibility) using total and normalized counts.

Technique Total counts – right condyle Normalized counts/pixel – left condyle

Operator 1 Operator 2 Operator 1 Operator 2

Variable-size ROI 1 image

Dahlberg S 2.24 3.32 1.55 3.77
[95% CI] 0.899 [0.728; 0.965] 0.859 [0.627; 0.952] 0.892 [0.717; 0.962] 0.814 [0.517; 0.936]

Fixed-size ROI 1 image

Dahlberg S 3.94 3.47 1.65 1.58
[95% CI] 0.984 [0.944; 0.995] 0.929 [0.713; 0.979] 0.982 [0.949; 0.994] 0.899 [0.723; 0.966]

Variable-size ROI 5 image

Dahlberg S 2.01 3.60 0.91 0.91
[95% CI] 0.957 [0.872; 0.986] 0.666 [0.247; 0.878] 0.903 [0.741; 0.966] 0.832 [0.582; 0.939]

Fixed-size ROI 5 image

Dahlberg S 0.32 0.32 0.60 0.52
[95% CI] 0.991 [0.945; 0.994] 0.998 [0.994; 0.999] 0.982 [0.945; 0.994] 0.999 [0.996; 0.999]

Fig. 4. Variable-size ROI image obtained from 5 trans-axial slices visualizing the Fig. 5. Fixed-size (1.76 cm2 ) ROI image from 5 trans-axial slices visualizing the
condyles. condyles.

Repeatability (intra-operator variability)

ROI using total counts by both operators. The techniques based on

Table 3 shows the Dahlberg S value and 95% CI for total and reading a fixed ROI area of 1.76 cm2 provided excellent repeata-
normalized counts for each condyle. The lowest Dahlberg S value bility (ICC > 0.90), although readings from operator 2 were more
(0.32) was obtained using the method of reading in 5 images a fixed variable.

Downloaded for Vanessa Ledesma (vl576@mynsu.nova.edu) at Nova Southeastern University from ClinicalKey.com by Elsevier on
September 27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
76 D.F. López Buitrago et al. / Rev Esp Med Nucl Imagen Mol. 2019;38(2):72–79
Comparison of techniques
No significant differences were detected in the percentage of
uptake measured by any method when sides were compared (right
vs left), or when the results from total counts were compared to
normalized counts (Table 4) (p > 0.05).
Regarding the variability of the method, the lowest standard
deviation, and therefore the lowest variance was obtained in the
method using 1 image, variable-size ROI method, with normalized
values (SD: 3.9) followed by the method using 5 images, fixed-size
ROI, normalized values (SD: 5.5).
The number of SPECT positive cases (uptake difference between
sides ≥10%, indicative of unilateral active growth) detected was
higher when total counts were used, compared to normalized
counts, particularly in variable-size ROI techniques. The agreement
between results based on total counts vs normalized counts was
higher in fixed-size ROI techniques than in variable-size techniques
(97.4% of agreement, Kappa > 0.94) (Table 5).
The number of SPECT positive cases, was also compared among
methods. The method using 1 image, variable-size ROI total counts,
detected the highest number of positive cases (n = 27). When nor-
malized values were used in this technique, only n = 5 cases were
identified as positive. This tendency was also observed in the
method using 5 images, variable method. In the method using
1 image fixed-size ROI, there is no difference between total and
normalized counts, concerning this apparent diagnostic sensitivity
(kappa and p value presented in Table 5).
The diagnostic agreement compared between pairs of methods
indicates higher concordance in fixed-size ROI techniques when
1 image is compared to 5 images (92.1% of agreement). On the
other hand, the agreement is poor when variable-size ROI tech-
niques were compared (Table 6). In general, agreement between
techniques is better when normalized counts are used.
Fig. 6 illustrates the comparative variability between fixed and
variable ROI methods. The method of 1 image variable ROI appar-
ently tends to overestimate the percentage of uptake, specifically
when total counts per condyle are higher than 20%. Limits of agree-
ment (LA) = −7.91; [−3.27 to 16.9]. On the other hand, the results
for the fixed ROI methods reported an LA = −1.04; [−8.97 to 6.88]
what might suggest that both methods could be clinically equiv-
alent as the points oscillate around 0 and the LA falls within the
clinically acceptable range.
Discussion
The first report of planar bone gammagraphy in 21 patients with
facial asymmetry was published by Cisneros in 1984.11 Posteriorly
Israel et al. in 198826 and Front et al. in 198927 , studied the ability
of SPECT to detect deep bone lesions that planar image methods
could not detect. Saridin et al. in 200818 compared the two kinds
of gammagraphic tests in 56 patients with suspected UCH. They
found more sensitivity (93%) and specificity (96%) for SPECT com-
pared to planar bone scintigraphy, concluding that SPECT is more
reliable as a diagnostic tool identifying the radiotracer uptake in
each condyle. More recently, López and Corral in 20166 published
their study of 61 patients, that shows as well that SPECT is supe-
rior to planar image to identify condylar uptake of the radionuclide
(32 positive results detected by SPECT (52.46%) against 8 (13.1%)
detected by planar scintigraphy). The meta-analysis published by
Saridin et al. in 201119 reports better sensitivity and specificity for
SPECT (90–95%) compared to planar scintigraphy (71–92%).
Additionally Wen et al. in 201428 evaluated the diagnostic
value of 99m Tc-MDP SPECT in patients with clinically sus-
pected CH, based on the higher condylar activity in the affected
side (59%) against the contralateral (41%), expressed by a
Downloaded for Vanessa Ledesma (vl576@mynsu.nova.edu) at Nova Southeastern University from ClinicalKey.com by Elsevier on
September 27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
differential uptake of the radionuclide. This report is in agreement
with Hooder et al. (2000),24 Kajan et al. (2006)25 and Fahey et al.
(2009)22 and reports that differences of more than 10% between
condyles is indicative of UCH. Considering the above summarized
findings, different therapeutic algorithms have been suggested by
Hooder et al. in 200024 and Rushinek et al. in 2016.29 The sur-
gical approach, condylectomy of the affected side, is suggested
when the difference between condylar uptake of 99m Tc-MDP is
more than 10%. The results of the present study show a dra-
matic change in the number of positive cases, when fixed ROI
are compared to variable ROI and when total counts are com-
pared to normalized counts in variable ROIs as well. This may
be related to the fact that variable ROIs detect the total condylar
uptake that is related to the grade of activity in the ossification
nucleus and/or to condylar size. On the other hand, a fixed ROI
detects uptake in the area considered for both condyles. Hence,
the method applied may significantly alter the therapeutic surgical
decision.
López et al. in 201730 analyzed 51 SPECT images from patients
with facial asymmetry and compared the value of qualitative
reports, based upon the subjective interpretation of a nuclear
medicine expert, versus quantitative reports obtained measuring
radionuclide uptake in a region of interest (ROI) of each condyle.
The conclusion was that qualitative reports are less reliable than
quantitative data (percentage of agreement: 39.2%). The authors
considered that the main limitation of quantitative reports is their
inability to detect bilateral CH cases, because the diagnosis is based
on a 10% difference between both sides in condylar uptake.
Therefore, it is generally accepted that SPECT quantitative
results are the best option available to detect hyperactivity in a
condylar tissue. But there is no consensus about the standardiza-
tion of SPECT quantitative protocols, to select the more reliable
technical approach, in terms of reproducibility and variability. Pri-
patnanont et al., 200517 studied 29 UCH patients to determine the
finalization of mandibular growth, using three methods for quan-
titative interpretation of SPECT results. From a ROI located in the
head of the condyle, they measured the total radioactive counts,
the number of pixels and the normalized counts per pixel. They
concluded that the total counts are sufficient to establish the per-
centage of uptake in each condyle and distinguish active/passive
condylar growth stages.
Fahey et al., 22 evaluated radionuclide uptake in healthy sub-
jects with no symptoms of CH, using two techniques to interpret
the data; the first technique measured counts in a fixed-size ROI
of 1.13 cm2 , considering the maximum values and the second tech-
nique measured activity in fixed ROIs obtaining the average of total
counts in three transaxial sections. Alsharif et al., 2 evaluated the
diagnostic value of three quantitative techniques for SPECT inter-
pretation in 33 patients with CH: the first technique used a whole
condyle approach, measuring a variable ROI equal to the total vol-
ume of each condyle; the second technique measured total counts
in a fixed-size ROI (1.14 cm2 ) drawn at the center of each condyle,
on a single transaxial image; the third technique considered a fixed-
size ROI (1.14 cm2 ), but on five transaxial images. They concluded
that fixed-size ROI maximum counts and the percentage difference
in condylar activity of 10% as a cut-off provide the best combination
of sensitivity, specificity, positive and negative predictive values.
Similar findings were obtained in the present study as no signifi-
cant differences in the number of positive cases were detected using
total counts vs normalized counts in fixed-size ROIs.
Additionally, Rushinek et al.29 evaluated three approaches to
SPECT analysis for the diagnostic of UCH, in terms of sensitivity
and specificity: fixed-size ROI, average volume of interest (VOI),
and maximum VOI. The specificity of these three approaches was
respectively: 77%, 65%, and 70% with the same sensitivity (88%) for
all of them.
 D.F. López Buitrago et al. / Rev Esp Med Nucl Imagen Mol. 2019;38(2):72–79 77

Table 4
Uptake percentage using total and normalized counts.

Technique Total counts Normalized counts/pixel

Right condyle Left condyle Right condyle Left condyle

Variable-size ROI 1 image

Average ± SD 50.5 ± 14.5 49.5 ± 14.5 49.5 ± 3.9 50.5 ± 3.9
95% CI 4.63 4.63 1.26 1.26

Fixed-size ROI 1 image

Average ± SD 49.3 ± 6.9 50.7 ± 6.9 49.3 ± 6.2 50.7 ± 6.2
95% CI 2.20 2.20 2.00 2.00

Variable-size ROI 5 image

Average ± SD 49.9 ± 8.5 50.1 ± 8.5 51.3 ± 7.8 48.7 ± 7.8
95% CI 2.75 2.75 2.49 2.49

Fixed-size ROI 5 image

Average ± SD 49.0 ± 6.1 51.0 ± 6.1 49.2 ± 5.5 50.8 ± 5.5
95% CI 1.97 1.96 1.77 1.77

Table 5
Agreement of positive cases between total and normalized counts within each technique.

Technique Total counts Normalized counts/pixel % of agreement Kappa (p value)

Variable-size ROI 5 image 17 5 63.2% 0.201 (0.044)

Fixed-size ROI 5 image 13 12 97.4% 0.940 (0.000)
Variable-size ROI 1 image 27 5 42.1% 0.116 (0.062)
Fixed-size ROI 1 image 14 13 97.4% 0.943 (0.000)

Table 6
Diagnostic agreement between pairs of techniques.

Techniques compared 95% CI n % of agreement Kappa (p value)

Total Variable ROI 5 image vs fixed ROI 5 image 0.593 [0.409; 0.777] 22 57.9% 0.129 (0.208)
counts Variable ROI 5 image vs variable ROI 1 image 0.301 [0.077; 0.524] 26 68.4% 0.395 (0.002)
Variable ROI 5 image vs fixed ROI 1 image 0.535 [0.314; 0.757] 21 55.3% 0.079 (0.309)
Fixed ROI 5 image vs variable ROI 1 image 0.277 [0.120; 0.434] 24 63.2% 0.349 (0.002)
Fixed ROI 5 image vs fixed ROI 1 image 0.889 [0.825; 0.954] 35 92.1% 0.828 (0.000)
Variable ROI 1 image vs fixed ROI 1 image 0.307 [0.129; 0.486] 23 60.5% 0.289 (0.012)

Normalized Variable ROI 5 image vs fixed ROI 5 image 0.300 [0.064; 0.536] 31 81.6% 0.494 (0.000)
counts/pixel Variable ROI 5 image vs variable ROI 1 image 0.142 [−0.082; 0.367] 34 89.5% 0.539 (0.000)
Variable ROI 5 image vs fixed ROI 1 image 0.282 [0.020; 0.545] 30 78.9% 0.451 (0.000)
Fixed ROI 5 image vs variable ROI 1 image 0.729 [0.597; 0.861] 29 76.3% 0.349 (0.006)
Fixed ROI 5 image vs fixed ROI 1 image 0.916 [0.868; 0.965] 35 92.1% 0.821 (0.000)
Variable ROI 1 image vs fixed ROI 1 image 0.724 [0.606; 0.841] 30 78.9% 0.451 (0.000)

30
A 30
B
Differeces of variables-size ROI

Differeces of fixed-size ROI

20 +1.96 Sd 16.9 20
(5 images-1 image)
(5 images-1 image)

10
10 +1.96 Sd 6.88

0
0 Mean -1.04
Mean -7.91
-10
-10 -1.96 Sd -8.97
-20
-20
-30
-1.96 Sd -32.7
-30
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Differences total count in condyles Differences total count in condyles

Fig. 6. Bland–Altman graphs showing average differences and dispersion in the uptake percentage. (A) Differences between variable-size ROI from using 5-images and
1-image methods. (B) Differences between fixed-size ROI from using 5-images and 1-image methods.

All the above-mentioned investigations coincide with the contralateral side in patients with facial asymmetry. As these differ-
present study about the advantage of fixed-size ROIs compared to ences affect the condylar shape, articular eminence height and the
variable ROIs. In the present study, the agreement between total projection of condyle within the articular fossa, the use of only one
and normalized counts is high for fixed-size ROIs (97.4% of agree- transaxial slice might not be as representative as the summatory
ment, Kappa 0.940 and p 0.000). The diagnostic agreement between of a few transaxial tomographic slices, to detect uptake differences
paired techniques is also favorable to fixed-size ROIs (92.1% of between condyles. Therefore, anatomic differences should be con-
agreement, Kappa 0.820 and p 0.000), but the difference between sidered in patients with facial asymmetry when the SPECT is used
using 1 or 5 images is not relevant. and interpreted.
However, Kim et al., 31 and Velásquez et al., 32 provided evi- According to the results obtained by Alsharif et al.,2 the tech-
dence of anatomic differences between the deviated side and the nique using variable ROI and one image with normalized values

Downloaded for Vanessa Ledesma (vl576@mynsu.nova.edu) at Nova Southeastern University from ClinicalKey.com by Elsevier on
September 27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
78 D.F. López Buitrago et al. / Rev Esp Med Nucl Imagen Mol. 2019;38(2):72–79
presents the lowest variability, as in the present study, and provides
88% of sensitivity and 87.5% specificity. But when the variable-size
ROI is applied to the total volume of condylar head, the high sen-
sitivity and specificity may depend on the variability in size of the
condyle. Lin et al.,33 in a 3D evaluation of the condyle in patients
with mandibular asymmetry reported that in fact the condylar size
was increased in the asymmetric side. Therefore, the uptake of
radionuclide may be influenced by the condylar size, but as well
may be due to the activity in the growth nucleus, articular incon-
gruence and the zone of photon scattering within the tissues. In
condyles of higher size, the likelihood of false positive diagnostics
could be increased.
Regarding intra-operator and inter-operator reproducibility, the
technique offering best reproducibility was the 5 image fixed-size
ROI, followed by the 1 image fixed-size ROI. This finding is con-
cordant with the study of Karssemakers et al.,34 that in 67 SPECT
images demonstrated a strong correlation between operators and
a 100% of agreement when total counts were used. But in that
study intra-operator repeatability was not measured and differ-
ent quantitative techniques were not compared as in the present
study.
In variable-size ROI data, the diagnostic difference obtained is
notorious when total counts vs normalized counts are used. Alto-
gether, our results suggest that the best technical option is the use
of 5 image, fixed-size ROI total counts when SPECT is applied for
the diagnostic of UCH.
Using 5 images instead of only one, as is common in previous
studies, is justified by the likelihood that the selected image may
not be representative of the pathology. In SPECT images, reorienta-
tion of the volumetric image before selection of the slice is a critical
issue. The use of 5 images reduces errors related to a wrong selec-
tion of the area or its reorientation. On the other hand, variable ROIs
are more dependent on the operator selection and interpretation,
and therefore tend to be more variable than fixed-size ROIs.
Finally, the use of normalized counts, assigning a constant value
for each pixel within each ROI, avoids the bias of selection when
the area of highest uptake is selected. However, the use of normal-
ized counts underestimates the zone of highest uptake because it
redistributes the counts in pixels of less uptake and its value is not
well defined because there are established normal values. On the
other hand, the use of maximum total counts is less variable and
more reproducible than normal counts, but its high sensibility may
cause more false positive results. For this reason, it is important
that the diagnosis of CH should be made from an adequate cor-
relation between facial and intraoral clinical characteristics with
radiographic and/or tomographic findings. Being SPECT the only
test providing information about the cellular metabolic activity of
the suspected condyle, in order to distinguish between an active or
passive CH and/or another entity.
A limitation of the present study, common to almost all the
previous studies, is that it does not provide data to correlate the
condylar metabolic activity with morphologic condylar changes, in
order to improve specificity. Another limitation is that real values
of sensitivity and specificity could not be supported by histologic
post-surgical data. In similar studies sensitivity, specificity and pre-
dictive values were obtained by clinical follow-up of the patients,
in absence of a histopathological gold standard.2
On the other hand, it is well established that increasing sensi-
tivity, the specificity is reduced. Therefore, it is necessary to accept
an intermediate compromise between the likelihood of obtaining
false positive results and false negative results, in a situation when
both may have deleterious effects. In this study the use of normal-
ized counts detected less cases considered as positive for CH but
total counts detected more positive cases.
The correct quantitative assessment of the SPECT test will allow
the development of studies related not only to the percentage of
Downloaded for Vanessa Ledesma (vl576@mynsu.nova.edu) at Nova Southeastern University from ClinicalKey.com by Elsevier on
September 27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
absorption between condyles, but also with the absorption ratio in
order to identify bilateral cases.
Conclusions
The comparison of four techniques to interpret SPECT results
for the diagnostic of UCH suggests that the more reliable option is
to use the method of 5 trans-axial slices, fixed-size ROI and total
counts.
Funding sources
This research did not receive any specific grant from funding
agencies in the public, commercial, or non-profit sectors.
Conflict of interest
The authors have no conflicts of interest to declare.
Acknowledgment
The authors appreciate the Research Institute of Centro Médico
Imbanaco for its support during the development of this project.
References
1. Obwegeser HL, Makek MS. Hemimandibular hyperplasia – hemimandibular
elongation. J. Maxillofac. Surg. 1986;14:183–208.
2. AlSharif AA, Tarawneh ES, AlKawaleet YI, Abukaraky AE, AlAhmad HT, Malkawi
ZA, et al. Standardization of quantitative single photon emission computed
tomography in control individuals and in patients with condylar hyperplasia.
Nucl. Med. Commun. 2014;35:1268–76.
3. 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.
4. Eslami B, Behnia H, Javadi H, Khiabani KS, Saffar AS. Histopathologic compari-
son of normal and hyperplastic condyles. Oral Surg. Oral Med. Oral Pathol. Oral
Radiol. Endod. 2003;96:711–7.
5. Raijmakers PG, Karssemakers LHE, Tuinzing DB. Female predominance and
effect of gender on unilateral condylar hyperplasia: a review and meta-analysis.
J. Oral Maxillofac. Surg. 2012;70:e72–6.
6. Lopez DF, Corral CM. Comparison of planar bone scintigraphy and single photon
emission computed tomography for diagnosis of active condylar hyperplasia. J.
Craniomaxillofac. Surg. 2016;44:70–4.
7. Bishara SE, Burkey PS, Kharouf JG. Dental and facial asymmetries: a review. Angle
Orthod. 1994;64:89–98.
8. Sora BC, Jaramillo VPM. Diagnóstico de las asimetrías faciales y dentales. Rev Fac
Odontol Univ Antioquia. 2005;16:15–25.
9. Wolford LM, Movahed R, Perez DE. A classification system for conditions causing
condylar hyperplasia. J. Oral Maxillofac. Surg. 2014;72:567–95.
10. Gray RJ, Sloan P, Quayle AA, Carter DH. Histopathological and scintigraphic fea-
tures of condylar hyperplasia. Int. J. Oral Maxillofac. Surg. 1990;19:65–71.
11. Cisneros GJ, Kaban LB. Computerized skeletal scintigraphy for assessment of
mandibular asymmetry. J. Oral Maxillofac. Surg. 1984;42:513–20.
12. Rodrigues DB, Castro V. Condylar hyperplasia of the temporomandibular joint:
types, treatment, and surgical implications. Oral Maxillofac Surg Clin North Am.
2015;27:155–67.
13. Grummons DC, Kappeyne van de Coppello MA. A frontal asymmetry analysis. J.
Clin. Orthod. 1987;21:448–65.
14. Kaban LB, Cisneros GJ, Heyman S, Treves S. Assessment of mandibular growth
by skeletal scintigraphy. J. Oral Maxillofac. Surg. 1982;40:18–22.
15. Saridin CP, Raijmakers P, Becking AG. Quantitative analysis of planar bone
scintigraphy in patients with unilateral condylar hyperplasia. Oral Surg. Oral
Med. Oral Pathol. Oral Radiol. Endod. 2007;104:259–63.
16. Villanueva-Alcojol L, Monje F, Gonzalez-Garcia R. Hyperplasia of the mandibular
condyle: clinical, histopathologic, and treatment considerations in a series of 36
patients. J. Oral Maxillofac. Surg. 2011;69:447–55.
17. Pripatnanont P, Vittayakittipong P, Markmanee U, Thongmak S, Yipintsoi T. The
use of SPECT to evaluate growth cessation of the mandible in unilateral condylar
hyperplasia. Int. J. Oral Maxillofac. Surg. 2005;34:364–8.
18. Saridin CP, Raijmakers PGHM, Tuinzing DB, Becking AG. Comparison of pla-
nar 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.
19. Saridin CP, Raijmakers PGHM, Tuinzing DB, Becking AG. Bone scintigraphy as
a diagnostic method in unilateral hyperactivity of the mandibular condyles:
a review and meta-analysis of the literature. Int. J. Oral Maxillofac. Surg.
2011;40:11–7.
 D.F. López Buitrago et al. / Rev Esp Med Nucl Imagen Mol. 2019;38(2):72–79 79

20. Pogrel MA, Kopf J, Dodson TB, Hattner R, Kaban LB. A comparison of single-
photon emission computed tomography and planar imaging for quantitative
skeletal scintigraphy of the mandibular condyle. Oral Surg. Oral Med. Oral Pathol.
Oral Radiol. Endod. 1995;80:226–31.
21. Slootweg PJ, Muller H. Condylar hyperplasia. A clinico-pathological analysis of
22 cases. J. Maxillofac. Surg. 1986;14:209–14.
22. Fahey FH, Abramson ZR, Padwa BL, Zimmerman RE, Zurakowski D, Nis-
senbaum M, et al. 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.
23. Motamedi MH. Treatment of condylar hyperplasia of the mandible using uni-
lateral ramus osteotomies. J. Oral Maxillofac. Surg. 1996;54:1161–70.
24. 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.
25. Kajan ZD, Motevasseli S, Nasab NK, Ghanepour H, Abbaspur F. Assessment of
growth activity in the mandibular condyles by single-photon emission com-
puted tomography (SPECT). Aust. Orthod. J. 2006;22:127–30.
26. Israel O, Jerushalmi J, Frenkel A, Kuten A, Front D. Normal and abnormal single
photon emission computed tomography of the skull: comparison with planar
scintigraphy. J. Nucl. Med. 1988;29:1341–6.
27. Front D, Israel O, Jerushalmi J, Frenkel A, Iosilevsky G, Feinsod M, et al. Quanti-
tative bone scintigraphy using SPECT. J. Nucl. Med. 1989;30:240–5.
28. Wen B, Shen Y, Wang C-Y. Clinical value of 99Tc m-MDP SPECT bone scintig-
raphy in the diagnosis of unilateral condylar hyperplasia. Sci World J. 2014:
256–62.
29. Rushinek H, Tabib R, Fleissig Y, Klein M, Tshori S. Evaluation of three analy-
sis methods for 99mTc MDP SPECT scintigraphy in the diagnosis of unilateral
condylar hyperplasia. Int. J. Oral Maxillofac. Surg. 2016;45:1607–13.
30. Lopez DF, Ruiz J, Corral CM, Carmona AR, Sabogal A. Comparison of 99mTc-MDP
SPECT qualitative vs quantitative results in patients with suspected condylar
hyperplasia. Rev Esp Med Nucl Imagen Mol. 2017;36:207–11.
31. Kim J-Y, Kim B-J, Park K-H, Huh J-K. Comparison of volume and position of the
temporomandibular joint structures in patients with mandibular asymmetry.
Oral Surg Oral Med Oral Pathol Oral Radiol. 2016;122:772–80.
32. Velasquez RL, Coro JC, Londono A, McGorray SP, Wheeler TT, Sato S.
Three-dimensional morphological characterization of malocclusions with
mandibular lateral displacement using cone-beam computed tomography.
Cranio. 2017:1–13.
33. Lin H, Zhu P, Lin Y, Wan S, Shu X, Xu Y, et al. Mandibular asymmetry: a
three-dimensional quantification of bilateral condyles. Head Face Med. 2013;
9:42.
34. Karssemakers LHE, Raijmakers PG, Nolte JW, Tuinzing DB, Becking AG. Interob-
server variation of single-photon emission computed tomography bone scans
in patients evaluated for unilateral condylar hyperactivity. Oral Surg Oral Med
Oral Pathol Oral Radiol. 2013;115:399–405.

Downloaded for Vanessa Ledesma (vl576@mynsu.nova.edu) at Nova Southeastern University from ClinicalKey.com by Elsevier on
September 27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.