Professional Documents
Culture Documents
Comparasion de 4 métodos para asesoría cuantitativa
Comparasion de 4 métodos para asesoría cuantitativa
Comparasion de 4 métodos para asesoría cuantitativa
2019;38(2):72–79
Original Article
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.
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
Introduction Table 1
Definition of quantitative techniques.
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.
Results
Statistical analysis
Reproducibility (inter-operator variability)
The average and standard deviation was obtained for each
measure. Normality of data distribution was evaluated by the Table 2 summarizes the results for inter-operator differences.
Shapiro–Wilk test. The four techniques were compared by one-way Taking into account the 95% CI value, the method of reading in
ANOVA. Intra-observer repeatability was estimated by the intra- 5 images a fixed ROI provides less variability between operators,
class correlation coefficient (ICC) and its 95% confidence interval followed by the method using 1 fixed image. The methods based
(95% CI). Inter-observer reproducibility was estimated by Dahlberg on variable ROI had greater variability. There are no differences
coefficient and ICC. Bland–Altman graphs were also obtained to related to the mandibular deviation side. The Dahlberg S coeffi-
compare data variability. The diagnostic agreement for detection cient in fixed-size ROI, the method of reading in 5 images is almost
of uptake percentage condylar differences ≥10% was calculated as equal using total counts (0.90) or normalized counts (0.88), but
percentage and Kappa coefficient. The statistical analyses were car- significantly lower than in variable-size ROIs, indicating better
ried out using the program Stata 13 (Stata Corp., College Station, TX, inter-operator reproducibility.
USA). The level of significance was 0.05.
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.
Table 3
Repeatability (intra-operator reproducibility) using total and normalized counts.
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.
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
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 77
Table 4
Uptake percentage using total and normalized counts.
Table 5
Agreement of positive cases between total and normalized counts within each technique.
Table 6
Diagnostic agreement between pairs of techniques.
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
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 absorption between condyles, but also with the absorption ratio in
88% of sensitivity and 87.5% specificity. But when the variable-size order to identify bilateral cases.
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 Conclusions
condyle. Lin et al.,33 in a 3D evaluation of the condyle in patients
with mandibular asymmetry reported that in fact the condylar size The comparison of four techniques to interpret SPECT results
was increased in the asymmetric side. Therefore, the uptake of for the diagnostic of UCH suggests that the more reliable option is
radionuclide may be influenced by the condylar size, but as well to use the method of 5 trans-axial slices, fixed-size ROI and total
may be due to the activity in the growth nucleus, articular incon- counts.
gruence and the zone of photon scattering within the tissues. In
condyles of higher size, the likelihood of false positive diagnostics Funding sources
could be increased.
Regarding intra-operator and inter-operator reproducibility, the This research did not receive any specific grant from funding
technique offering best reproducibility was the 5 image fixed-size agencies in the public, commercial, or non-profit sectors.
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 Conflict of interest
images demonstrated a strong correlation between operators and
a 100% of agreement when total counts were used. But in that The authors have no conflicts of interest to declare.
study intra-operator repeatability was not measured and differ-
ent quantitative techniques were not compared as in the present Acknowledgment
study.
In variable-size ROI data, the diagnostic difference obtained is The authors appreciate the Research Institute of Centro Médico
notorious when total counts vs normalized counts are used. Alto- Imbanaco for its support during the development of this project.
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 References
the diagnostic of UCH.
Using 5 images instead of only one, as is common in previous 1. Obwegeser HL, Makek MS. Hemimandibular hyperplasia – hemimandibular
studies, is justified by the likelihood that the selected image may elongation. J. Maxillofac. Surg. 1986;14:183–208.
2. AlSharif AA, Tarawneh ES, AlKawaleet YI, Abukaraky AE, AlAhmad HT, Malkawi
not be representative of the pathology. In SPECT images, reorienta- ZA, et al. Standardization of quantitative single photon emission computed
tion of the volumetric image before selection of the slice is a critical tomography in control individuals and in patients with condylar hyperplasia.
issue. The use of 5 images reduces errors related to a wrong selec- Nucl. Med. Commun. 2014;35:1268–76.
3. Nitzan DW, Katsnelson A, Bermanis I, Brin I, Casap N. The clinical characteristics
tion of the area or its reorientation. On the other hand, variable ROIs
of condylar hyperplasia: experience with 61 patients. J. Oral Maxillofac. Surg.
are more dependent on the operator selection and interpretation, 2008;66:312–8.
and therefore tend to be more variable than fixed-size ROIs. 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
Finally, the use of normalized counts, assigning a constant value
Radiol. Endod. 2003;96:711–7.
for each pixel within each ROI, avoids the bias of selection when 5. Raijmakers PG, Karssemakers LHE, Tuinzing DB. Female predominance and
the area of highest uptake is selected. However, the use of normal- effect of gender on unilateral condylar hyperplasia: a review and meta-analysis.
ized counts underestimates the zone of highest uptake because it J. Oral Maxillofac. Surg. 2012;70:e72–6.
6. Lopez DF, Corral CM. Comparison of planar bone scintigraphy and single photon
redistributes the counts in pixels of less uptake and its value is not emission computed tomography for diagnosis of active condylar hyperplasia. J.
well defined because there are established normal values. On the Craniomaxillofac. Surg. 2016;44:70–4.
other hand, the use of maximum total counts is less variable and 7. Bishara SE, Burkey PS, Kharouf JG. Dental and facial asymmetries: a review. Angle
Orthod. 1994;64:89–98.
more reproducible than normal counts, but its high sensibility may 8. Sora BC, Jaramillo VPM. Diagnóstico de las asimetrías faciales y dentales. Rev Fac
cause more false positive results. For this reason, it is important Odontol Univ Antioquia. 2005;16:15–25.
that the diagnosis of CH should be made from an adequate cor- 9. Wolford LM, Movahed R, Perez DE. A classification system for conditions causing
condylar hyperplasia. J. Oral Maxillofac. Surg. 2014;72:567–95.
relation between facial and intraoral clinical characteristics with 10. Gray RJ, Sloan P, Quayle AA, Carter DH. Histopathological and scintigraphic fea-
radiographic and/or tomographic findings. Being SPECT the only tures of condylar hyperplasia. Int. J. Oral Maxillofac. Surg. 1990;19:65–71.
test providing information about the cellular metabolic activity of 11. Cisneros GJ, Kaban LB. Computerized skeletal scintigraphy for assessment of
mandibular asymmetry. J. Oral Maxillofac. Surg. 1984;42:513–20.
the suspected condyle, in order to distinguish between an active or
12. Rodrigues DB, Castro V. Condylar hyperplasia of the temporomandibular joint:
passive CH and/or another entity. types, treatment, and surgical implications. Oral Maxillofac Surg Clin North Am.
A limitation of the present study, common to almost all the 2015;27:155–67.
13. Grummons DC, Kappeyne van de Coppello MA. A frontal asymmetry analysis. J.
previous studies, is that it does not provide data to correlate the
Clin. Orthod. 1987;21:448–65.
condylar metabolic activity with morphologic condylar changes, in 14. Kaban LB, Cisneros GJ, Heyman S, Treves S. Assessment of mandibular growth
order to improve specificity. Another limitation is that real values by skeletal scintigraphy. J. Oral Maxillofac. Surg. 1982;40:18–22.
of sensitivity and specificity could not be supported by histologic 15. Saridin CP, Raijmakers P, Becking AG. Quantitative analysis of planar bone
scintigraphy in patients with unilateral condylar hyperplasia. Oral Surg. Oral
post-surgical data. In similar studies sensitivity, specificity and pre- Med. Oral Pathol. Oral Radiol. Endod. 2007;104:259–63.
dictive values were obtained by clinical follow-up of the patients, 16. Villanueva-Alcojol L, Monje F, Gonzalez-Garcia R. Hyperplasia of the mandibular
in absence of a histopathological gold standard.2 condyle: clinical, histopathologic, and treatment considerations in a series of 36
patients. J. Oral Maxillofac. Surg. 2011;69:447–55.
On the other hand, it is well established that increasing sensi- 17. Pripatnanont P, Vittayakittipong P, Markmanee U, Thongmak S, Yipintsoi T. The
tivity, the specificity is reduced. Therefore, it is necessary to accept use of SPECT to evaluate growth cessation of the mandible in unilateral condylar
an intermediate compromise between the likelihood of obtaining hyperplasia. Int. J. Oral Maxillofac. Surg. 2005;34:364–8.
18. Saridin CP, Raijmakers PGHM, Tuinzing DB, Becking AG. Comparison of pla-
false positive results and false negative results, in a situation when nar bone scintigraphy and single photon emission computed tomography in
both may have deleterious effects. In this study the use of normal- patients suspected of having unilateral condylar hyperactivity. Oral Surg. Oral
ized counts detected less cases considered as positive for CH but Med. Oral Pathol. Oral Radiol. Endod. 2008;106:426–32.
19. Saridin CP, Raijmakers PGHM, Tuinzing DB, Becking AG. Bone scintigraphy as
total counts detected more positive cases.
a diagnostic method in unilateral hyperactivity of the mandibular condyles:
The correct quantitative assessment of the SPECT test will allow a review and meta-analysis of the literature. Int. J. Oral Maxillofac. Surg.
the development of studies related not only to the percentage of 2011;40:11–7.
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 79
20. Pogrel MA, Kopf J, Dodson TB, Hattner R, Kaban LB. A comparison of single- 28. Wen B, Shen Y, Wang C-Y. Clinical value of 99Tc m-MDP SPECT bone scintig-
photon emission computed tomography and planar imaging for quantitative raphy in the diagnosis of unilateral condylar hyperplasia. Sci World J. 2014:
skeletal scintigraphy of the mandibular condyle. Oral Surg. Oral Med. Oral Pathol. 256–62.
Oral Radiol. Endod. 1995;80:226–31. 29. Rushinek H, Tabib R, Fleissig Y, Klein M, Tshori S. Evaluation of three analy-
21. Slootweg PJ, Muller H. Condylar hyperplasia. A clinico-pathological analysis of sis methods for 99mTc MDP SPECT scintigraphy in the diagnosis of unilateral
22 cases. J. Maxillofac. Surg. 1986;14:209–14. condylar hyperplasia. Int. J. Oral Maxillofac. Surg. 2016;45:1607–13.
22. Fahey FH, Abramson ZR, Padwa BL, Zimmerman RE, Zurakowski D, Nis- 30. Lopez DF, Ruiz J, Corral CM, Carmona AR, Sabogal A. Comparison of 99mTc-MDP
senbaum M, et al. Use of (99m)Tc-MDP SPECT for assessment of mandibular SPECT qualitative vs quantitative results in patients with suspected condylar
growth: development of normal values. Eur. J. Nucl. Med. Mol. Imaging. hyperplasia. Rev Esp Med Nucl Imagen Mol. 2017;36:207–11.
2010;37:1002–10. 31. Kim J-Y, Kim B-J, Park K-H, Huh J-K. Comparison of volume and position of the
23. Motamedi MH. Treatment of condylar hyperplasia of the mandible using uni- temporomandibular joint structures in patients with mandibular asymmetry.
lateral ramus osteotomies. J. Oral Maxillofac. Surg. 1996;54:1161–70. Oral Surg Oral Med Oral Pathol Oral Radiol. 2016;122:772–80.
24. Hodder SC, Rees JI, Oliver TB, Facey PE, Sugar AW. SPECT bone scintigraphy in 32. Velasquez RL, Coro JC, Londono A, McGorray SP, Wheeler TT, Sato S.
the diagnosis and management of mandibular condylar hyperplasia. Br. J. Oral Three-dimensional morphological characterization of malocclusions with
Maxillofac. Surg. 2000;38:87–93. mandibular lateral displacement using cone-beam computed tomography.
25. Kajan ZD, Motevasseli S, Nasab NK, Ghanepour H, Abbaspur F. Assessment of Cranio. 2017:1–13.
growth activity in the mandibular condyles by single-photon emission com- 33. Lin H, Zhu P, Lin Y, Wan S, Shu X, Xu Y, et al. Mandibular asymmetry: a
puted tomography (SPECT). Aust. Orthod. J. 2006;22:127–30. three-dimensional quantification of bilateral condyles. Head Face Med. 2013;
26. Israel O, Jerushalmi J, Frenkel A, Kuten A, Front D. Normal and abnormal single 9:42.
photon emission computed tomography of the skull: comparison with planar 34. Karssemakers LHE, Raijmakers PG, Nolte JW, Tuinzing DB, Becking AG. Interob-
scintigraphy. J. Nucl. Med. 1988;29:1341–6. server variation of single-photon emission computed tomography bone scans
27. Front D, Israel O, Jerushalmi J, Frenkel A, Iosilevsky G, Feinsod M, et al. Quanti- in patients evaluated for unilateral condylar hyperactivity. Oral Surg Oral Med
tative bone scintigraphy using SPECT. J. Nucl. Med. 1989;30:240–5. 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.