Cone-beam computed tomography (CBCT) is prone to several inherent and introduced artifacts which may cause difficulty in the interpretation of images. “Beam hardening” and “Scatter” artifacts occur when metallic objects exist in the field of imaging. X-ray photons that are diffracted from their original path after interaction with the metallic object cause the scattering phenomenon. As an X-ray beam passes through an object, lower energy photons are absorbed in preference to higher energy photons. Beam hardening leads to two types of artifacts: (1) Distortion of metallic structures as a result of differential absorption, known as a cupping artifact, and (2) Streaks and dark bands, which occur between two dense objects, create extinction or missing value artifacts. ¹

Amalgam restorations can be responsible for such artifacts when CBCT images of the dentition and adjacent structures are obtained. These artifacts can interfere with diagnostic evaluations such as the diagnosis of dental caries. ² , ³

Utilizing digital radiographic modalities, including CBCT leads to better diagnosis in various fields of dentistry such as endodontics and restorative dentistry ⁴ , ⁵ , ⁶ , ⁷ and in accordance with other advancements in the field of dentistry, reduces the complications of dental treatments. ⁸ , ⁹ Seker et al. study ¹⁰ and Kayipmaz et al. ¹¹ found that the accuracy of CBCT imaging in occlusal, proximal, and recurrent dental caries is better than the plain radiographs and digital radiographs (phosphor storage plates [PSPs]). On the other hand, Zhang et al. ¹² have reported that the accuracy of diagnosis of three-dimensional (3D) systems such as ProMax and Kodak 9000 has no significant differences with conventional intraoral PSPs in noncavitated carious lesions detection. Therefore, they do not suggest using CBCT in caries detection. Pauwels et al. ¹³ study concluded that regions in the vicinity of metal rods were moderately or severely affected by artifacts, particularly in the area between the rods.

Esmaeili et al. ¹⁴ and Isman et al. ¹⁵ studies noted that artifacts produced by metallic objects such as orthodontic brackets, which occur due to beam-hardening phenomenon, occur in all computed tomography (CT) and CBCT imaging systems and lead to issues in the diagnostic process.

Furthermore, when a patient has a CBCT image for other purposes if CBCT can be also useful in diagnosing caries, the individual no longer requires plain radiographs, and therefore, this can be useful for reducing patients' radiation according to as low as reasonably achievable principle, which states that a minimum amount of radiation must be given to the patient. Moreover, due to the controversial results of previous studies about using CBCT for detecting caries and the effect of metal artifacts on diagnosing process, one important question arises: Are CBCT images reliable for caries detection, especially when they are next to a metal restoration?

This study aimed to evaluate the effect of amalgam restorations on adjacent teeth for detecting proximal caries using CBCT.

In this in vitro study, 102 extracted human permanent posterior teeth (46 premolars and 52 molars) with or without apparent caries were randomly selected. The teeth were extracted due to orthodontics or periodontal problems. All procedures followed were under the Ethical Standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008. All procedures performed in the present study were approved by the Ethical Committee of Isfahan University of Medical Sciences (#IR.MUI.RESEARCH.REC.1395.3.247). The present study was funded by Isfahan University of Medical Sciences (395,247) and was performed as a partial requirement for obtaining DDS degree. Any debris and calculus were removed and all teeth were disinfected with 2% hypochlorite solution for 20 min and then stored in normal saline. Six molar teeth were selected and mesio-occluso-distal (MOD) cavities were prepared on them. Then, they were restored with amalgam. The gingival floor of the cavity was placed 2 mm coronal to the cementoenamel junction (CEJ) and the crown of the tooth was sectioned to separate the root and crown 2 mm apical to the CEJ.

The 96 remaining teeth were divided into six groups using the block technique with six blocks with a size of 16 by utilizing the random allocation software 2.0 (Mahmood Saghaei, Isfahan, Iran). Each group included both mandibular and maxillary posterior teeth. The crowns of the teeth were placed in red wax (Polywax, Tehran, Iran) so that the proximal contacts were reestablished. Then, the teeth were mounted on six dry human skulls in their respective dental socket. The teeth were mounted in a way that each group of teeth was placed in their sockets and the appropriate occlusion was established between jaws by MS. Furthermore, the MS glued and fixed the cranium and the mandible to maintain this position Figure 1. Figure 1

Frequency percentage of diagnosis of caries level by histopathological findings.

The histopathologic findings of the proximal surface of the 96 teeth are presented in Figure 1.

Perfect intra-observer agreement was found for the first and second observer (κ values ranging from 0.895 to 0.936), as well as inter-observer agreement (κ values ranging from 0.857 to 0.931).

Due to strong intra-observer agreement, further calculations were based on the first readings of each observer.

The placement of amalgam fillings caused a decrease in sensitivity, specificity, PPV, and NPV. However, the FPR was significantly increased after the placement of amalgam restorations. The FNR was not significantly different between the two sets of images Table 1.{Table 1}

The frequency percentage of diagnosis of caries level before and after amalgam filling by CBCT and histopathological findings are shown in Figure 2and Figure 3. Figure 2

Frequency percentage of diagnosis of caries level by histopathological findings.

In the present study, no significant difference existed between pathologic diagnosis and radiographic diagnosis of dental caries before amalgam restoration placement. However, placement of MOD amalgam restorations significantly decreased sensitivity, specificity, and PPV and NPV for caries detection. This finding confirmed that CBCT imaging is less accurate in the diagnosis of proximal caries adjacent to amalgam restorations compared to nonrestored teeth.

Different studies show different results regarding the accuracy of CBCT images for caries detection.

In one study, Zhang et al. ¹² evaluated the accuracy of proximal caries detection by CBCT, film, and phosphor plates. They used ProMa × 3D, Kodak 9000 3D imaging, phosphor plates, and film, and the gold standard was the histologic examination. The results of their study showed that to detect subtle noncavitated proximal caries, the detection accuracy with the CBCT images was a little better than chance performance and was similar to that with phosphor plate – and film-based intraoral images. Their findings are according to this study. Haiter-Neto et al. ¹⁶ reported that the sensitivity of Accuitomo and NewTom was 21% and 13%–18%, respectively, in field of view of 6″, 9″, and 12″. The low value in this study may be explained by subtle caries in teeth samples besides the high percentage of sound to carious surfaces. The low sensitivity value in the present study also would be explained by this reason. Furthermore, the results of their study showed that the NewTom 3G CBCT had a lower diagnostic accuracy for the detection of caries lesions than intraoral modalities and the 3DX Accuitomo CBCT. In an investigation which had been done by Kamburoğlu et al., ¹⁷ no differences had been found between different methods of caries detection such as intraoral radiographs, intraoral digital radiographs charged coupled device (CCD), and 3D imaging modalities (Accuitomo and Iluma CBCT imaging systems).

In an in vitro study, Tarım Ertas et al. ¹⁸ compared the diagnostic accuracy of different methods in detecting occlusal caries lesions. In their study, CBCT exhibited better performance in detecting deep occlusal caries than other imaging modalities. Another study performed by Sansare et al. ¹⁹ showed that CBCT is a more accurate method in the diagnosis of cavitated proximal caries compared to bitewing radiography. Moreover, it is expressed in this study that CBCTs taken for other purposes can also be used for cavitated caries.

In a study by Qu et al., ²⁰ AZ-values for five types of CBCT systems were very close together (about 0.5), indicating that CBCT is not a suitable tool for diagnosis of primary proximal caries, that these results are in agreement with the result of this study.

Nabha et al. ²¹ showed that metallic substances such as amalgam create an artifact in 3D imaging which, in turn, cause problems for dental diagnostic purposes. This study also showed that MOD amalgam restorations create the most artifacts in CBCT images.

In a recent study by Kulczyk et al., ²² it was shown that specificity, sensitivity, PPV, and NPV for the diagnosis of caries adjacent to amalgam restorations are 0.64, 47.27, 0.0, and 33.56, respectively. These figures indicate inadequate accuracy of CBCT in the diagnosis of caries which complies with the results of the present study.

Performing the imaging procedure in an optimum in vitro condition without motion blur and surrounding soft tissue, which makes the imaging and diagnosis process difficult and misleading in caries detection. In reality, this was the limitation of this study. The authors suggest performing further studies in a clinical environment with a higher number of samples.

CBCT imaging is not efficient enough to detect proximal caries adjacent to amalgam restorations.

Acknowledgments

The present study was funded by Isfahan University of Medical Sciences (395,247) and was performed as a partial requirement for obtaining DDS degree. Furthermore, the dry human skulls were provided by the Maxillofacial Department of Isfahan Medical University.

Financial support and sponsorship

The present study was funded by Isfahan University of Medical Sciences (395,247) and was performed as a partial requirement for obtaining DDS degree.

Conflicts of interest

The authors of this manuscript declare that they have no conflicts of interest, real or perceived, financial or non-financial in this article.