INTRODUCTION

Radiographs were first used by Pacini1 in the study of facial anthropometrics in 1922. Following publications by Broadbent2 and Hofrath3 on cephalometric procedures in 1931, lateral and frontal conventional cephalometry became widely used for diagnosis and treatment planning of orthodontic problems and for determination of growth and treatment effects.

Frontal radiographs (FRs) provide crucial information regarding asymmetry and width of the jaws, and thus can be extremely beneficial for treatment planning of surgical cases. However, many orthodontic treatment procedures are geared toward resolving conditions that cannot be appraised adequately with the use of conventional two-dimensional radiographs. For example, many relationships of the craniofacial complex, such as the position of the mandibular condyles in the temporomandibular fossa with respect to the occlusal scheme, and the association of airway abnormalities with craniofacial morphology, cannot be evaluated with conventional imaging approaches. Thus, multislice computed tomography (MSCT) has become a facilitating technique in dental practice in recent decades.4 Although this method provides an obvious advantage over conventional FR in providing three-dimensional (3D) scans, it was replaced by cone beam computed tomography (CBCT), which exposes the patient to much less radiation.5 The therapeutic benefits of CBCT include measurement of the dentition and tooth volume, planning and evaluation of temporary anchorage devices, monitoring of alveolar bone density and height in interdisciplinary periodontal and orthodontic treatment, and 3D treatment simulation.6–8 Most of these activities can be performed during a single imaging session with a radiation dose that is in the same range as that of conventional images, which do not provide the same quantity of data.

To our knowledge, after a bibliographic search in Medline using PubMed and the keywords “frontal radiographs,” “cephalometrics,” “cone beam computed tomography,” and “human skull,” no study has compared conventional FRs vs CBCT-constructed FRs obtained from the same patient. Most studies investigating this topic have focused on human skulls. However, images from dry skulls do not suffer from distortion caused by soft tissues, and we have to use radiographs obtained from patients with orthodontic anomalies for their treatment planning. Therefore, the aim of this study was to find out whether measurements obtained on conventional FRs are comparable with the same measurements obtained on CBCTconstructed FRs from the same patients for use in possible longitudinal researches where only two-dimensional records from the past were available.

MATERIALS AND METHODS

The sample size for the study was calculated on the basis of a significance level of .05 and a power of 80% to detect a clinically meaningful difference of 1.00 ± 1.35 mm for the moR-moL parameter. Power analysis showed that 30 patients were required. Thus, the study included the records of 30 patients (16 females and 14 males) between the ages of 18 and 23 years (mean age, 21 ± 1.1 years) who were randomly selected from the archive of the oral diagnosis and orthodontic departments at Karadeniz Technical University. Approval from the ethics committee was not required for this retrospective study. Inclusion criteria included the presence of (1) permanent upper and lower incisors; (2) the first permanent upper and lower molars; (3) a reproducible, stable occlusion; and (4) both conventional FR (Veraviewepocs, Morita, Kyoto, Japan) (Figure 1A) and CBCT scan (Kodak 9500 Cone Beam 3D System, Kodak Dental Systems, Carestream Health, Rochester, NY), which became available within a 2-month period (average, 35 days).

View Figure

• Download as Powerpoint
• Download Figure

Cephalometry

Upon cephalometric analysis, 23 conventional hard tissue cephalometric landmarks were identified (Table 1 and Figure 2). Twenty-one widely used cephalometric variables (14 linear distances, 4 angles, and 3 ratios) were calculated (Table 2). Conventional FRs were traced with a 0.3-mm 2H pencil on matte acetate paper and were measured with a digital caliper and protractor. CBCT-constructed FRs were imported into Kodak software (Kodak Dental Imaging software, version 6.11.7.0-B, Carestream Health). Relevant landmarks were identified, and measurements were calculated with the use of software. All cephalometric procedures were performed by a single examiner (MN) in random order.

View Figure

• Download as Powerpoint
• Download Figure

Table 1 Cephalometric Landmarks Used in the Present Study

View Fullscreen

Table 2 Linear and Angular Parameters and Ratios Measured on Posterior-Anterior Radiographs

View Fullscreen

RESULTS

Intraobserver reliability for conventional and CBCT-constructed FRs was acceptable. Corresponding intraobserver correlation coefficients for both radiographs were greater than r²  =  0.78. The correlation coefficient between first and second measurements ranged from 0.78 to 0.96 (average, 0.87) for conventional FRs and from 0.85 to 0.99 (average, 0.93) for CBCT-constructed FRs (Table 3).

Table 3 Intraobserver Reliability of Measurements Obtained From Conventional and Cone Beam Computed Tomography–Constructed Frontal Radiographs

View Fullscreen

Table 4 shows mean and standard deviation values and the results of paired t-tests comparing measurements obtained from conventional and CBCT-constructed FRs. Statistical analysis showed a statistically significant difference between the two radiographic methods for all linear measurements (P < .01), except for the ans-cr parameter, which was a vertical linear measurement (P > .05). On the other hand, no statistically significant differences were noted for ratio (R me-ag, R mx-ans, R cr-lo) and angular measurements (me-agR-cr, me-agL-cr, agL-agR-loR).

Table 4 Mean, Standard Deviation, and Difference Values of Measurements Obtained From Conventional and Cone Beam Computed Tomography–Constructed Frontal Radiographs

View Fullscreen

DISCUSSION

In this study, 21 widely used cephalometric measurements on conventional FRs and on CBCT-constructed frontal cephalometric radiographs of patients randomly selected from the archive of the dental faculty were compared. Data from the archive were used in the present study because it is not ethically acceptable to expose patients to radiation through the use of both conventional radiography and CBCT.

Most of the studies9–12 conducted to compare measurements on conventional FRs and CBCT-constructed frontal cephalometric radiographs used CBCT-constructed frontal cephalometric radiographs taken from dry human skulls. Dry human skulls do not suffer from distortion caused by soft tissues, thus the chance of error in landmark identification is reduced. However, in the clinic, CBCT-constructed frontal cephalometric radiographs are used, and these radiographs may be affected by distortion caused by soft tissues. To our knowledge, no study has compared conventional vs CBCT-constructed FRs obtained from the same patient.

In the studies of van Vlijmen et al.,10,11 measurements were taken by a single observer. The authors indicate that systematic errors in the identification of landmarks were the same for both types of radiographs; thus it was justified to have a single observer for this type of study. For this reason, all measurements on both radiographs in the present study were performed by a single observer.

For all linear measurements, except one (ans-cr), statistically significant differences were found between conventional and CBCT-constructed FRs. Differences ranged from 1.13 to 15.06 mm, maximum for the eurR-eurL parameter and minimum for the moR-moL parameter. It is a known fact in conventional cephalometry that magnification increases as the distance from anatomic points to the film increases, and that magnification decreases as the film gets closer (as the distance diminishes).13,14 Therefore, the difference between conventional and CBCT-constructed FRs for the eurR-eurL parameter might be maximum. In a recently published paper,11 the mean differences for FRs were much greater than for lateral radiographs. The reason for this was that the landmarks for FRs are located in a greater number of tomographic planes compared with those for lateral cephalometry.

In recent studies,9,10 the reproducibility of measurements obtained from CBCT scans was found to be greater than the reproducibility of those obtained from conventional FRs. In accordance with this finding, the reproducibility of measurements has ranged from 0.85 to 0.99 (average, 0.93) for CBCT-constructed FRs, and from 0.78 to 0.96 (average, 0.87) for conventional FRs.

On the other hand, it has been shown that positioning a patient while FRs are made is very critical, because rotation of the head results in measurement differences.10,12,15 Because a CBCT device has no cephalostat, extra care should be taken when positioning the patient in the device. In addition, van Vlijmen et al.10,11 recommended using extra light beams in doing so, to prevent unwanted tilt of the head. In the present study, to eliminate the odds of error during measurement in synthesized views, a correction was made to the orientation of the computed tomography (CT) volume through tasks of iterative adjustment and reassessment, and the head position was reoriented with the use of Kodak software, in keeping with the head positioning rules of conventional cephalometric projections.

Another important finding of the present study was that there was no statistically significant difference between angular and ratio measurements performed on conventional FRs, compared with CBCT-constructed FRs. Kumar et al.16 compared measurements from synthesized CBCT cephalograms using orthogonal and perspective projections vs those derived from conventional cephalometric radiographs, and showed that angular measurements were not statistically different for any modality. However, van Vlijmen et al.11 concluded that there was a clinically relevant difference between angular measurements obtained from conventional FRs and those obtained from CBCT-constructed 3D models. They stated that tracings of 3D models therefore were not suitable for longitudinal research in cases where only two-dimensional records from the past were available. This might be related to patient positioning, as already stated. If a 3D model is made and 3D analysis is performed, it is likely that positioning of the patient has no effect on measurements taken during 3D cephalometric analysis.

CONCLUSIONS

• Differences have been noted between measurements performed on conventional FRs, in comparison with measurements performed on frontal cephalometric radiographs constructed from CBCT scans, particularly for linear measurements.

• However, no statistically significant difference has been observed between angular and ratio measurements performed on conventional FRs compared with CBCT-constructed FRs.

• Results suggest that frontal cephalometric radiographs constructed from CBCT scans should not be used for longitudinal research in cases where only two-dimensional records from the past are available.