Comparison of panoramic radiography and cone-beam computed tomography for qualitative and quantitative measurements regarding localization of permanent impacted maxillary canines

OBJECTIVE: The purpose of this retrospective study was to compare the correlation between digital panoramic radiography (DPR) and cone-beam computed tomography (CBCT) evaluations for localization of impacted permanent maxillary canines (IPMCs) and for other qualitative and quantitative parameters. MATERIALS AND METHOD: DPR and CBCT images of 60 patients (17 men and 43 women) were examined independently by two observers. Correlations between DPR and CBCT images were evaluated regarding qualitative (bucco-palatal positioning of IPMCs, morphology and presence of root resorption of adjacent permanent lateral incisors, and contact relationship between IPMCs and adjacent permanent lateral incisors) and quantitative (angle measurements) variables. All evaluations were repeated 1 month later by each observer. Chi-square and t-tests were used for statistical analysis. Kappa statistics were used to assess intraand interobserver agreement (Cohen’s κ). RESULTS: No correlation was observed for determination of bucco-palatal positioning of IPMCs between DPR and CBCT images (p>0.05). Correlations were observed for other qualitative variables (p<0.05). Differences between DPR and CBCT images were seen for all examined quantitative variables (p<0.01). Intraand interobserver agreements were substantial to almost-perfect. CONCLUSION: No significant correlation was found between DPR and CBCT images for determination of bucco-palatal positioning of IPMCs. All quantitative measurements performed on DPR and CBCT images significantly differed from each other.


INTRODUCTION
Permanent maxillary canines are the second most frequently impacted teeth after the third molars, with an impaction prevalence of 1-3%. 1,2Impacted permanent maxillary canines (IPMCs) are found twice as often in women than in men. 3 These impacted teeth can cause aesthetic concerns because they appear in the smile line.They can cause functional problems as a result of various pathologies, such as cyst formation and root resorption of adjacent teeth. 3Surgical interventions for the rehabilitation of aesthetic and functional problems are very difficult, and orthodontic treatment is difficult and time consuming. 4The proper localization and early detection of IPMCs are very important to prevent possible complications in adjacent teeth, ankyloses, and cysts. 5 clinical practice, panoramic radiography is the preferred primary radiographic imaging technique for impacted teeth. 4When determining the actual position of the impacted tooth, two-dimensional images obtained by occlusal and periapical radiographs can be used in combination.However, these images have many disadvantages, such as blurring, superposition, and distortion due to projection errors.Therefore, three-dimensional (3D) imaging is necessary to determine the actual position of the IPMC. 4 In recent years, cone-beam computed tomography (CBCT) systems for acquiring 3D images of oral structures have been preferred due to their relatively low cost and low radiation dose. 6Several studies have examined the localization of IPMCs for various populations, such as Italian, 7 Swedish, 5 Chinese, 8,9 German, 10 Belgian, 6 North American, 11 Korean, 12 Swiss, 13 and Polish. 14To the best of our knowledge, however, no such studies have been conducted for the Turkish population.
The purpose of this retrospective study was to compare whether there was a correlation between digital panoramic radiographic (DPR) and CBCT evaluations in terms of the localization of IPMCs, morphology of the adjacent permanent lateral incisors (PLIs), contact relationship between IPMCs and PLIs, presence of root resorption in the adjacent PLIs, and various angle measurements.DPR and CBCT images of patients with following criteria were included in this study: patients were over 15 years old, had clear maxillary radiographic images with unilateral or bilateral IPMCs, had no pathology in the maxillary region, and had no artifacts in the maxillary region that would affect image quality.A total of 69 IPMCs of 60 patients (17 men and 43 women) met the study criteria and were examined.Sample size was determined to be adequate by statistical power analysis (power value = 0.94).DPR images were obtained using a Morita Veraviewepocs 2D (Morita, Kyoto, Japan) with parameters of 60-80 kVp, 1-10 mA, 0.5 mm focal spot, and exposure time of 7.4 seconds.CBCT images were obtained with a Promax 3D® (Planmeca, Helsinki, Finland) with 8 × 8 cm, 5 × 8 cm, and 5 × 4 cm fields of view and parameters of 90 kVp, 12 mA, scanning time of 13.85 seconds, and voxel size of 0.4 × 0.4 × 0.4 mm.DPR and CBCT scans were performed with the patient rest-ing in the supine position.The head of the patient was positioned using two light-beam markers, with a vertical positioning light being aligned with the mid-sagittal line of the patient.[3][4][5][6][7][8][9][10][11][12][13][14][15] The amount of magnification, which was determined by the manufacturer (×1.3), was taken into account in the measurements of DPR images.Morphology of the adjacent PLI, contact relationship between the IPMC and the adjacent PLI, presence of root resorption in the adjacent PLI, and angle measurements were evaluated in DPR images and in the coronal and sagittal sections of CBCT images (Table 1).

MATERIALS AND METHOD
The Canine-Incisor Index (CII) was used to localize the bucco-palatal position of the IPMC in DPR images.The CII was calculated by dividing the widest mesiodistal size of the IPMC by the widest mesiodistal size of the permanent central incisor on the same side.If the resulting number was smaller than 1.15, then the position was classified as "buccally located"; if the number was greater than 1.15, then the position was classified as "palatally located". 9,15,16To determine the bucco-palatal position of the IPMC in the CBCT image, distances from the IPMC crown to the buccal and palatal cortical bones were measured.When the distance from the IPMC crown to the buccal bone was shorter than the distance to the palatal bone, the tooth was categorized as "buccally located"; otherwise, the tooth was categorized as "palatally located". 17

Data analysis
Data were analyzed statistically using SPSS program version 21.0 (SPSS Inc., Chicago, IL, USA).Fisher's exact test was performed.Cramer's V correlation coefficients were calculated for statistical analysis of the correlation between DPR and CBCT images for the following qualitative variables: bucco-palatal positioning of the IPMC, contact relationship between the IPMC and the adjacent PLI, and morphology of the PLI.Differences between DPR and CBCT images were statistically analyzed by using the t-test for quantitative variables, including angle measurements.Kappa statistics were used to assess intra-and interobserver agreements.Interpretation of the correlation coefficient obtained from the Kappa statistics was evaluated as suggested by Landis and Koch. 18Statistical analyses were performed at significance levels of 95% and 99% for qualitative and quantitative variables, respectively.

Intra-and interobserver agreements
When we examined intraobserver agreement for the DPR method for quantitative variables, we obtained Cohen's κ values of 0.78-0.98 and 0.74-0.95for the first and second rater, respectively (substantial to almost-perfect agreement).For the CBCT method, intraobserver Cohen's κ values varied 0.95-0.99 for both raters (almost-perfect agreement).When interobserver agreements in both methods for the quantitative variables were examined, Cohen's κ values were 0.84-0.98(almost-perfect agreement).

Qualitative variables
No correlation was found between DPR and CBCT images for bucco-palatal positioning of IPMCs (Table 2; p>0.05).The percentage of teeth observed in the buccal region on both DPR and CBCT images was 33.8%.Correlations between DPR and CBCT images were found for the morphology of the adjacent PLI, the contact relationship, and the root resorption (Tables 3-5; p<0.05).These parameters were determined identically on DPR and CBCT images in 85.5%, 84.1%, and 55% of cases, respectively.

Quantitative variables
Differences were found between DPR and CBCT images for all examined quantitative variables (Table 6; p<0.01).

DISCUSSION
To the best of our knowledge, only five studies in the literature have compared panoramic radiography and CBCT in terms of the localization of IPMCs. 6,10,17,19,20In these studies, orthodontists, oral surgeons, and/or dental practitioners worked as observers.Although radiology was an important part of these studies, none of them involved dentomaxillofacial radiologists.In contrast, all of the evaluations in this study were performed by specialists in dentomaxillofacial radiology with at least 2 years of experience.
Previous studies used magnification methods 6,20,21 and CII 9,15,16 calculations to determine the bucco-palatal positions of IPMCs in panoramic radiographs.Chaushu et al. 16 confirmed that localization of the bucco-palatal the same study, root resorption was observed in 13% of permanent central incisors on panoramic radiography and 15.1% on CBCT. 6In the present study, the presence of root resorption in adjacent PLIs was detected in 23.2% of DPR images and 62.3% of CBCT images.Identical readings in DPR and CBCT were found in 55% of cases.These findings are similar to the findings of previous studies. 6,11gles of IPMCs to the midline, occlusal plane, and PLIs are useful for estimation of the possibility of root resorption in adjacent PLIs, and also for the localization of the bucco-palatal position of the IPMC. 14,25Our findings regarding these quantitative variables are in accordance with a previous study, in which the angle measurements in panoramic radiographs and CBCT images were inconsistent. 9

CONCLUSION
DPR and CBCT images yielded similar results for some of the qualitative parameters, including morphology of the adjacent PLI, contact relationship with adjacent PLIs, and root resorption.However, discrete findings were obtained for the bucco-palatal position of IPMCs and for all quantitative variables in DPR and CBCT techniques.
positioning of IPMCs can be determined reliably using the CII in panoramic radiography.Similar results were reported in other studies. 6,20Haney et al. 19 reported significant correlations between CBCT and DPR images for the bucco-palatal localization of IPMCs.However, no other study to date has analyzed IPMC localization by comparing DPR (using CII) and CBCT.In this study, only 33.8% of IPMCs were in an identical bucco-palatal position in both DPR and CBCT images.
3][24] Lai et al. 13 investigated the morphology of PLIs on CBCT images of IPMCs.Permanent lateral incisors were absent in 2.9%, had a normal morphology in 70.9%, and had a peg-shaped morphology in 26.1% of images.In the present study, on DPR images, pegshaped PLIs were identified in 21.7% of cases, whereas CBCT revealed this rate as 10.1%.However, the two methods were statistically consistent with each other for the "morphology of the adjacent PLI" variable, with identical readings in 85.5% of cases.
Consistent with previous studies of CBCT images, IPMCs were mostly impacted in the palatal side in patients with peg-shaped PLIs. 23,24Ericson and Kurol 5 emphasized that resorption in maxillary PLIs in patients with IPMCs was due to the pressure from the permanent canine during eruption or from contact between these teeth.They also found that the IPMC crown was in contact with the adjacent PLI in 67% of cases, and was in contact with the permanent central incisor in 57% of cases. 5Lai et al. 13 classified the contact relationship between the IPMC and the PLI as follows: cervical (10.4%), middle (43.2%), or apical third (21.6%) of the root.Alqerban et al. 6 stated that there was a contact relationship between respective teeth in 73.9% of DPR images and 89% of CBCT images.In the present study, 50.7% of PLIs were in contact with the IPMC in the cervical third and 49.3% were in contact in the apical third, as evaluated on DPR images.No tooth without contact was found.This study showed that DPR and CBCT readings were substantially consistent, with a rate of identical readings of 84.1%.
Root resorption caused by IPMC occurs most often in PLIs. 2,5,8Ericson and Kurol 5 reported that root resorption in permanent incisors due to ectopically positioned IPMCs was approximately 50% and four times more prevalent in women than in men.Preda et al. 7 observed root resorption in PLIs at a rate of 27.6% on spiral CT images.In studies conducted by CBCT, root resorption in adjacent PLIs was reported at rates of 13.4%, 17 25.4%, 13 30.1%, 1243%, 20 and 59.6%. 11Liu et al. 8 reported that root resorption was present in PLIs on CBCT at a rate of 27.2%, whereas the rate for the permanent central incisors was 23.4%.Alqerban et al. 6 compared panoramic radiography with CBCT and found root resorption in PLIs at rates of 29.4-30.7% on panoramic radiography and 50.9-53.9%on CBCT.In

Figure 1 .
Figure 1.Assessment of bucco-palatal position of the IPMC.(A) Bucco-palatal position of the IPMC on DPR image according to CII. (B) CBCT image for measurements of distances from IPMC crown to buccal and palatinal cortical plates in the sagittal plane.

Figure 3 .
Figure 3. Contact relationship of IPMC with adjacent PLI.(A&B) Contact relationship in the cervical third of the root on DPR image and CBCT image in the sagittal plane.(C&D) Contact relationship in the apical third of the root on DPR image and CBCT image in the sagittal plane.

Figure 4 .
Figure 4. Absence of root resorption on the adjacent PLI: (A) on DPR image and (B) on CBCT image in the sagittal plane.

Table 1 .
Evaluated variables and categories of variables in the study

Table 4 .
Correlation between DPR and CBCT images for contact relationship of the IPMC and the adjacent PLI

Table 5 .
Correlation between DPR and CBCT images for root resorption caused by IPMC in adjacent PLI a Fisher's exact test; * statistically significant (p<0.05)

Table 6 .
Comparison of DPR and CBCT images for quantitative variables a t-test; * statistically significant (p<0.05)