Original article

Standardization of quantitative single photon emission computed tomography in control individuals and in patients with condylar hyperplasia Abedallatif A. AlSharifa, Emad S. Tarawnehb, Yazan I. AlKawaleetc, Ashraf E. Abukarakyd, Hazem T. AlAhmadd, Ziad A. Malkawid and Malik E. Juweida Objectives The aim of the study was to evaluate the diagnostic accuracy of various single photon emission computed tomography (SPECT) quantitative methods in patients with condylar hyperplasia (CH) and to investigate whether normal condylar activity changes with age. Patients and methods We analyzed the SPECT images of 33 patients with CH and those of 16 control individuals. Regions of interest (ROIs) were drawn on whole condyle, or fixed-size ROIs were drawn on both condyles and the clivus on the slice with higher activity [a two-dimensional (2D) approach] and on the summation of five adjacent transaxial slices [a three-dimensional (3D) approach]. A percentage difference between both condyles of above 10% or a cutoff value of 1.44 or 1.88 for abnormal condyle/clivus ratio was considered abnormal.

in condyle/clivus ratio was evident between patients with active and those with inactive CH. Conclusion Use of 2D maximum fixed-size ROI and percentual difference in condylar activity offers optimal diagnostic accuracy in patients with CH and should be encouraged in future studies. The condyle/clivus ratio offers suboptimal results and cannot, therefore, be recommended. No effect of age on normal condylar activity was demonstrated. Nucl Med Commun c 2014 Wolters Kluwer Health | Lippincott 35:1268–1276  Williams & Wilkins. Nuclear Medicine Communications 2014, 35:1268–1276 Keywords: age, clivus, condylar hyperplasia, quantification, SPECT a

Radiology and Nuclear Medicine Department, Nuclear Medicine Section, Radiology and Nuclear Medicine Department, General Radiology Section, Faculty of Medicine and dDepartment of Maxillofacial Surgery, Faculty of Dentistry, University of Jordan, Amman, Jordan b

Results Seventeen patients with active CH, 16 with inactive CH, and 16 control individuals were evaluated. The highest sensitivity and highest specificity were observed for the whole-condyle approach (88 and 87%, respectively), followed by the percentage 2D maximum condyle/total (82.4 and 81.3%, respectively). The condyle/clivus ratio yielded low sensitivity for both 2D and 3D approaches. No effect of age on condylar activity was demonstrated. No statistically significant difference

Introduction Hyperplasia of the mandibular condyle [condylar hyperplasia (CH)] is a rare, distinct pathology affecting the mandibular condyle unilaterally and resulting in mandibular overgrowth and consequently in facial asymmetry, occlusal changes, and often pain or dysfunction of the temporomandibular joint (TMJ) [1–3]. The disorder is usually self-limiting, but as long as CH remains active the asymmetry progresses. CH is diagnosed on the basis of history and clinical findings, followed by radiological evaluation [4]. The question as to whether condylar growth is active or has ceased cannot be addressed by conventional radiography and poses a critical problem in selecting the appropriate treatment procedure. When growth is still ongoing, condylectomy is often performed to avoid secondary adaptive deformation of the maxilla. In contrast, if needed, corrective orthoganthic surgery is advocated when condylar growth has ceased in order to c 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins 0143-3636 

c

Correspondence to Abedallatif A. AlSharif, MD, Radiology and Nuclear Medicine Department, Nuclear Medicine Section, Faculty of Medicine, Jordan University Hospital, University of Jordan, Queen Rania Street, Amman 11942, Jordan Fax: + 962 6535 3388; e-mail: [email protected] Received 18 August 2013 Revised 10 December 2013 Accepted 14 December 2013

avoid relapse [5,6]. Serial clinical examinations remain the method of choice in assessing the active growth of the condyle; however, clinical assessment remains subjective and may not detect subtle changes [4–6]. Serial cephalometric radiological evaluation performed over months to years is needed to evaluate disease progression and patients quite often are not willing to wait that long [4–6]. Activity of CH is effectively demonstrated by bone scintigraphy, which is considered an instant method assessing relative growth rates at the time of imaging [7,8]. Single photon emission computed tomography (SPECT) has been demonstrated to have higher sensitivity and specificity than planar bone scintigraphy in detecting active CH, and therefore SPECT has become an efficient tool for the management of CH [8–16]. SPECT is frequently requested at the initial stages of patient assessment [9–16]. Different quantitative measurements of condylar 99m Tc-methylene diphosphonate (MDP) uptake using DOI: 10.1097/MNM.0000000000000074

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Quantitative SPECT in condylar hyperplasia AlSharif et al. 1269

SPECT have been suggested [8,17,18]. Typically, the abnormal condylar activity is compared with the apparently clinically normal contralateral condyle or with a neutral reference point such as the clivus. However, the normal condyle in CH patients is subject to altered joint mechanics and stress distribution, whereas activity in the clivus can be variable and may change with age [14,19,20]. In addition, different SPECT quantitative approaches have been used: in some a variable-sized region of interest (ROI) is drawn around the whole normal and abnormal condyle, whereas in others increased activity in the abnormal condyle is related to its larger size and therefore the use of fixed-size ROIs is preferred. ROIs were also drawn on the single slice with the highest condylar activity [a twodimensional (2D) approach] or on the summation of three to five transaxial slices [a three-dimensional (3D) approach]. Total ROI count and mean or maximum count were all used in different reports [7–18]. In this study, we evaluated the diagnostic accuracy of the existing quantitative approaches of SPECT in CH patients compared with a group of normal control individuals. We also investigated whether normal condylar activity changes with age and whether abnormal joint mechanics lead to altered normal condylar activity in patients with facial asymmetry.

Patients and methods Patients

We retrospectively reviewed all scintigraphic studies of consecutive patients referred to Jordan University Hospital with clinically diagnosed unilateral CH during the period from March 2007 until October 2012. Inclusion criteria were vertical, transverse, or vertical and transverse lower facial plane asymmetry, as demonstrated clinically and radiologically by experienced maxillofacial surgeons. Patients with previous trauma or surgery to the TMJ or with previous mandibular fractures were excluded from the evaluation. Bone SPECT was performed routinely as a part of the workup to confirm the diagnosis and to assess growth activity. Patients who subsequently underwent condylectomy or demonstrated progressive facial asymmetry on clinical and radiological exams were considered as having active CH [8]. Patients who did not demonstrate progressive facial asymmetry and therefore did not undergo condylectomy were considered as having inactive CH [8]. SPECT was also performed on control individuals aged between 17 and 40 years. These individuals did not have facial asymmetry or known TMJ disorders or metabolic bone disease. They were referred for bone scintigraphy for unrelated indications, such as suspected osteomyelitis outside the head region or complex regional pain syndrome, or as part of their metastatic workup. This study was approved by the institutional review board at Jordan University Hospital.

Scintigraphy

Bone SPECTwas performed 3 h after intravenous injection of 740 MBq of 99mTc-MDP. Images were acquired by a singlehead gamma camera (Meridian; Philips, USA) or a dual-head gamma camera (E.CAM; Siemens Medical Solutions, Malvern, Pennsylvania, USA). Both gamma cameras are equipped with high-resolution parallel hole collimators. The peak energy was set at 140 keV with a 20% window width. A total of 128 projection images of 20 s each were acquired over 3601 using a 64  64 matrix. Transaxial, coronal, and sagittal tomograms were reconstructed using standard Esoft software (Siemens Medical Solutions) with an iterative reconstruction algorithm (OSEM, 10 iterations and four subsets). Images were filtered using a 14th order Butterworth filter with a cutoff frequency of 0.5 cycle/Nyquist. Quantitative analysis of single photon emission computed tomography images

Reconstructed images were processed using different methods by a nuclear medicine physician (A.A.): (1) ROIs were drawn around the whole condyle in the summed transaxial images where the condyle was identified and total condylar counts were computed (the whole-condyle approach) (Fig. 1a). (2) A fixed-size ROI (1.14 cm2) was drawn at the center of both condyles and on the clivus on a single transaxial slice with the highest condylar activity. Total ROI counts and average and maximum pixel counts were then computed (the 2D method) (Fig. 1b). (3) A fixed-size ROI (1.14 cm2) was drawn on both condyles and on the clivus on a summed image of five transaxial slices. Total counts and maximum pixel counts were then computed (the 3D method) (Fig. 1c). (4) Skull reference area: a 4 cm2 ROI was drawn at the center of a summed image of the 128 projection images to obtain the least count in the summed projection image (Fig. 1d) [18]. The following parameters were then calculated: (1) Percentage activity in the affected condyle for the whole-condyle approach: Abnormal condylar count 100 % : Abnormalþnormal condyle counts (2) Percentage total activity in the fixed-size ROI in the affected condyle for the 2D and 3D methods: Abnormal total condylar count 100 % : Abnormalþnormal condyle total counts (3) Ratio of fixed-size ROI condyle count/fixed-size ROI total clivus count for the 2D and 3D methods. (4) Percentage maximum activity in the affected condyle for the 2D and 3D methods (fixed-size ROI): Abnormal condylar maximun count 100 % : Abnormalþnormal condyle maximum counts

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

1270 Nuclear Medicine Communications 2014, Vol 35 No 12

Fig. 1

(a)

(b)

Lt

Rt

clivus RT

(c)

LT

(d)

clivus Rt

Lt

(a) Region of interest (ROI) is drawn around the whole abnormal left condyle and around the normal right condyle. (b) Fixed-size ROI (1.14 cm2) is drawn at both condyles and at the clivus on a single transaxial slice. (c) Fixed-size ROI (1.14 cm2) is drawn at both condyles and at the clivus on the sum of five transaxial slices (three dimensional). (d) Fixed-size 4 cm2 ROI is drawn at the area of least count on the summation of 128 projection images.

(5) Ratio of the fixed-size ROI condyle maximum count/ fixed-size ROI maximum clivus count. (6) The TMJ index calculated as mean condylar count in a 2D fixed-size ROI/mean count in the skull reference area [18].

Statistical analysis

Data are presented as mean±SD. The T-test was used to examine whether the examined parameters were significantly different among all patient groups. Pearson’s correlation coefficient was used to test the correlation between reference area and age. Data were analyzed using SPSS software for windows (version 11.5; SPSS Inc., Chicago, Illinois, USA). A P value less than 0.05 was considered statistically significant.

Results Thirty-three patients and 16 control individuals were included in the study (Table 1). The mean age of the patients was 22.12±6.36 years and the mean age of the control individuals was 24.94±95 years. Seventeen patients were diagnosed as having active CH, of whom 15 underwent condylectomy and two elected for a wait-and-see approach. The average duration between bone scan and condylectomy was 30.9±43.0 days (range 3–161 days). Ten patients who underwent condylectomy underwent subsequent corrective orthoganthic surgery (one bimaxillary osteotomy, two bimaxillary osteotomy with genoplasty, two bisagittal split osteotomy, two genoplasty, and three unilateral mandibular osteotomy). Five patients

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Quantitative SPECT in condylar hyperplasia AlSharif et al.

Table 1

1271

Characteristics of patients and control participants Number of participants

All patients Active CH Inactive CH Control individuals

33 17 16 16

Female patients [n (%)] 25 13 12 5

Mean age±SD

Age range (years)

Mean follow-up±SD

Follow-up (months)

22.12±6.36 23.12±7.25 21.12±5.35 24.94±6.95

15–42 15–42 15–36 17–38

43±20.29 41.37±21.19 45.53±19.93 –

8–75 8–72 9–75 –

(76) (76.5) (75.0) (31.25)

CH, condylar hyperplasia.

Table 2

Calculated parameters for control participants

Average SD Lowest Highest Above cutoffa

%2D

2D/C

%2Dmax

2Dmax/C

%3D

3D/C

%3Dmax

3Dmax/C

TMJ index

50 3 43 57 1

1.20 0.31 0.77 1.96 2

50 2 46 55 0

1.28 0.34 0.72 2.00 2

50 3 44 56 1.00

1.15 0.32 0.65 2.04 1

50 3 44 56 1

1.14 0.32 0.67 2.06 1

0.16 0.07 0.06 0.35 1

%2D and %2Dmax refer to percentual condylar activity in total and maximum counts. %3D and %3Dmax refer to percentual condylar activity in total and maximum counts. 2D/C and 3D/C refer to condyle/clivus ratio in 2D and 3D, respectively. 2Dmax/C and 3Dmax/C refer to condyle/clivus maximum in 2D and 3D, respectively. 2D, two dimensional; 3D, three dimensional; TMJ, temporomandibular joint. a 55 for %condyle/total, 1.88 for condyle/clivus ratio, and 30% for TMJ index.

who underwent condylectomy decided not to undergo subsequent corrective surgery and did not demonstrate further mandibular growth on follow-up. In contrast, the two patients who did not undergo condylectomy developed further mandibular asymmetry evident clinically and radiologically. Sixteen patients were considered as having inactive CH. Of them, nine elected not to undergo any surgery and were referred for orthodontic treatment. Two patients underwent bilateral sagittal split osteotomy, two underwent bimaxillary osteotomy, two underwent genoplasty, and one underwent bimaxillary osteotomy with genoplasty. The average duration between the scan and surgery was 1.6±2.1 months (range 0.03–5.8 months). None of the 16 patients demonstrated progression of facial asymmetry on an average follow-up of 44.12±19.39 months (range 8.23–74.83 months). Table 2 summarizes the computed parameters of the control individuals. No statistically significant correlation was found between the normal condylar activity measured (as a ratio to the clivus) and age (r = – 0.16, P = 0.21, for fixed-size 2D; r = – 0.26, P = 0.06, for fixedsize 2D maximum; r = – 0.154, P = 0.22, for fixed-size 3D; and r = – 0.23, P = 0.07, for fixed-size 3D maximum), nor between normal condylar activity measured as a ratio to skull reference area and age (r = – 0.2, P = 0.06). In addition, no statistically significant correlation was found between 2D fixed-size mean clivus activity (measured as a ratio to skull reference area) and age (r = 0.13, P = 0.1). No statistically significant difference was evident between the contralateral condyle (normal)/clivus ratios in active CH patients compared with patients with inactive

CH or control individuals for any of the calculated parameters (P > 0.05). The normal condyle/clivus ratio was markedly variable with a range of 0.538–3.670 in all participants. The TMJ index of the normal condyle was not significantly different among the three groups. The condyle/clivus ratio in patients with active CH was statistically higher than the condyle/clivus ratio in control individuals using fixed-size 2D ROI (P < 0.03), fixed-size 3D (P = 0.03), and fixed-size 3D maximum approaches (P < 0.05) and was only marginally significant for 2D maximum (P = 0.08). The condyle/clivus ratio between patients with active CH and those with inactive CH was only marginally significant using fixed-size 2D ROI (P = 0.092) and fixedsize 3D (P = 0.068) and was insignificant for fixed-size 2D maximum (P = 0.132) and fixed-size 3D maximum (P = 0.16). No significant difference was observed for the condyle/ clivus ratio between patients with inactive CH and control individuals using any of the approaches (P value was 0.17 for fixed-size 2D ROI, 0.35 for fixed-size 2D maximum, 0.32 for fixed-size 3D, and 0.13 for fixed-size 3D maximum) (Fig. 2). The percentage total activity and percentage maximum activity of abnormal condyle in active CH patients were statistically higher than the corresponding activity in control individuals (P < 0.01) and that in patients with inactive disease (P < 0.01) on using all approaches (Fig. 3). In addition, in patients with inactive CH the percentage total activity and percentage maximum activity of abnormal condyle were also significantly higher than the corresponding condyle activity in the control group (P value was