In vivo detection of external apical root resorption induced by apical periodontitis using periapical radiography and cone-beam computed tomography

Objectives: To investigate sensitivity, specicity, predictive values and accuracy of periapical radiography (PR) and Cone beam computed tomography (CBCT) for detection of external apical root resorption (EARR). Materials and Methods: Dog’s teeth with experimentally induced root resorption underwent or not root canal treatment (n = 62 roots). True positives (TP), false positives (FP), true negatives (TN) and false negatives (FN) in PR and CBCT diagnoses were determined using histopathologic ndings as gold standard. Sensitivity, specicity, positive predictive value (PPV), negative predictive value (NPV) and diagnostic accuracy (TP + TN) in the diagnosis of EARR were calculated. Data was compared using chi-squared test ( (cid:0) = 0.05). Results: We found that PR detected EARR in 35% of roots and CBCT, in 47%. EARR was microscopically diagnosed in 50% (p = 0.03 comparison between PR and microscopy; p = 0.67 comparison between CBCT and microscopy). Overall, CBCT produced more accurate diagnoses than PR (0.93 for CBCT versus 0.70 for PR; p = 0.008). Interestingly, when data was stratied into small and large resorptions, PR and CBCT allowed identication of large resorption in 100% of the cases and showed the same accuracy. However, for small resorptions, PR showed an accuracy of 0.83, whereas CBCT showed an accuracy of 0.96 (p = 0.003). Conclusions: We demonstrated that CBCT showed higher accuracy to detect EARR. These ndings shed light on the use of CBCT for detection of initial root resorption. Clinical relevance: Early identication of resorption allows a prompt treatment and reduces the risk of dental structure loss.


Introduction
External apical root resorption (EARR) might be induced by endodontic pathogens in a pathologic process, consequence of a persistent stimulation of in ammatory mediators that triggers a resorptive mechanism, which gradually eliminates cementoblasts, cementum and the dentin area on the external surface of the dental root (1,2). Current technology has given access to new adjunct diagnostic tools to perform an adequate diagnosis.
Though periapical radiography (PR) remains as one of the most used diagnostic aids in dentistry, Cone beam computed tomography (CBCT) are now available for the practitioner to deliver a better diagnostic and treatment plan. CBCT is a well-established method that compensate for the PR drawbacks such as distortion, anatomic superimposition and magni cation, allowing the operator to completely explore the region of interest in a CBCT scan using a computer program (3,4).
Small eld of view (FOV) cone beam computed tomography (CBCT) has increasingly replacing several radiological procedures (5). In cases in which lower dose radiography does not provide adequate or satisfactory diagnostic information, small eld of view CBCT has been recommended for assessment and/or management of root resorption when imaging would potentially change or enhance treatment plan (6).
Previously we demonstrated that cone beam computed tomography (CBCT) allows detection of small size apical periodontitis (AP) and that healing rates following root canal treatment were lower than previous radiographic studies have predicted (7)(8)(9). Speci cally, for investigation of external root resorption, CBCT has a high accuracy in detecting ex-vivo arti cial dental root surface mineral loss (10)(11)(12)(13). Even though previous demonstrated that CBCT portrays an attractive and valuable option to assess periodontal and periapical tissues, its accuracy to detect external root resorption in vivo has not been investigated. Therefore, this study aims to investigate sensitivity, speci city, predictive values and accuracy of periapical radiography (PR) and Cone Beam Computed Tomography (CBCT) in diagnosing external apical root resorption.

Materials And Methods
This is a secondary study based on experimental protocols and material generated by the study of Paula-Silva et al. (7,8), previously approved by the Animal Research Ethical Committee of Ribeirão Preto Dental School of the University of São Paulo (process number 07. 1.192.53.6). Brie y, an animal experiment with 12 mongrel dogs (12 months of age, body weight from 10 to 15 kg) was performed and, for that, premolar teeth were used to induce AP (62 dental roots of dogs' teeth). The animals received medication Roots with AP were treated either by one-visit, two-visit therapy or left untreated. The animals were followed throughout the experimental period to observe changes in eating habits, development of in ammatory processes or suppuration of tissues, among other abnormalities. After 6 months, the animals were euthanized with a lethal intravenous overdose of sodium pentobarbital. The mandibles were dissected and sectioned to obtain individual roots that were xed in 10% buffered formalin for 72 hours, demineralized in EDTA, and embedded in para n. The specimens were serially sectioned.
Longitudinal sections of 5-mm thickness were stained with hematoxylin and eosin (HE).
Periapical radiographs used in this study was performed according to the parallelism technique, using a device for standardizing radiographic takes in dogs (14). Ultraspeed® size 2 periapical lms (Eastman Kodak Company, Rochester, USA), Heliodent® dental X-ray machine (Siemens, New York, USA), with 60 kVp and 10 mA and exposure time of 1 second were used. The radiographs were developed manually using the time / temperature method and led in plastic cards. The images were digitalized using an optical scanner (Scanjet 7450C, Hewlett-Packard; Palo Alto, CA) with a resolution of 1,200 dpi.
CBCT scans were performed using a NewTom 3G cone beam computed tomography equipment (QR Srl, Verona, Italy). For this study, 120 kv and 3.6 mA were used for standardization of the technique, with 9inch FOV ( eld of view) and exposure time of 36 seconds. The NewTom 3G device captures 360 images with an interval of 1 between them, in 36 seconds, with an image reconstruction resolution of 512 × 512 pixels and 12 bits per pixel (4096 gray scale).
To investigate the presence of EARR in the current report, microscopic slides from 62 roots were collected and examined under uorescence microscopy using HE stained cuts (excitation at 460-500 nm and emission at 512-542 nm) at 20 × magni cation. A skilled and trained observer classi ed EARR as small external apical root resorption when mineralized tissue loss involved cementum only or large external root resorption when mineralized tissue loss involved cementum and dentin.
Three calibrated examiners (κ = 0.94) evaluated presence of external root resorption in radiographic and tomographic images. For radiographic analysis, images were assessed using the Image J 1.28 Software (National Institutes of Health, Bethesda, MD) in which brightness, contrast and magni cation tools could be used. For CBCT evaluation, the NewTom 3G equipment software was used. In order to allow a direct comparison with the periapical radiography, the use of sections of 1 millimeter thick with 0.5 millimeter spacing between sections was standardized, in sagittal view.
Based on the presence of microscopically detected small external root resorption (involving cementum only) or large external root resorption (involving cementum and dentin), true-positives (TPs), falsepositives (FPs), true-negatives (TNs), and false-negatives (FNs) in PR and CBCT diagnoses were determined using microscopy ndings as the gold standard. Sensitivity, speci city, positive predictive value (PPV), negative predictive value (NPV), and diagnostic accuracy (true positives + true negatives) in the diagnosis of AP were calculated. Data was compared using chi-squared test ( = 0.05).

Results
After microscopic analyses, we found that external root resorption was present in 50% of the specimens.
PR allowed detection of EARR in 35.5% of roots and CBCT in 46.8%. (p = 0.03 comparison between PR and microscopy; p = 0.67 comparison between CBCT and microscopy) ( Table 1). False negative diagnoses were most common after PR evaluation (14.5% of the roots) than CBCT (3.2% of the roots).
Both tools presented adequate speci city, yet CBCT presented higher sensitivity (0.93). Results of TP, FP, TN, FN, sensitivity, speci city, PPV, NPV, and diagnostic accuracy for diagnosis of EARR are summarized in Tables 2 and 3. CBCT produced more accurate diagnoses than PR (p = 0.008).
When data was strati ed into small and large resorptions, we found that for large resorptions, in which microscopically cementum and dentin were damaged, PR and CBCT allowed identi cation of resorption in 100% of the cases and showed the same accuracy. For small resorptions, in which only cementum was damaged, PR showed an accuracy of 0.83 whereas CBCT showed an accuracy of 0.96 (p = 0.003). PR had a lower sensitivity (0.62) in nding small resorptions, yet its positive predictive value was 100%. Results of TP, FP, TN, FN, sensitivity, speci city, PPV, NPV, and diagnostic accuracy for diagnosis of small EARR and large resorptions are summarized in Tables 2 and 3.

Discussion
Our study showed that CBCT has a higher sensitivity in detecting external apical root resorption than PR. Nevertheless, both techniques presented a similar speci city. Our results are consistent with previous studies that reported CBCT as the most accurate method in nding arti cial external root resorption when compared to panoramic or periapical radiography (13,15,16). Even though, we observed that PR has been found to be less sensitive, a high speci city of PR was obtained.
In order to obtain detailed information about the degree of mineralised tissue resoprtion, we classi ed large and small resorption by analyzing the affected mineralized tissue after histological processing.
Large resorption, in which cementum and dentin were damaged, were detected using both radiographic techniques. For small resorption detection, however, where only cementum was affected, PR had a lower sensitivity than CBCT. These results could be explained by a previous study that stated that mineral bone loss could not be detected by the human eye in periapical radiograph until reaching 7,1% of loss (17), indicating that small resorption will not be perceive using conventional radiographic techniques, whereas CBCT will provide a better sensitivity and accuracy in detecting apical external root resorption.
In some cases, cone beam computed tomography (CBCT) has increasingly replacing conventional radiological procedures due to the possibility of arbitrary reconstructions and views free of superimposition (5). Previous studies have reported that CBCT has a high accuracy in detecting EER, however these studies were performed under different methodologies, in ex vivo models with arti cial or natural EER, that may impair their results after comparison with other radiographic techniques (10-13).
Our analysis was performed in teeth that were treated endodontically or not in vivo, in which soft tissue thickness as well as other anatomic features that may in uence the periapical image are considered.
Furthermore, using the microscopic analysis as gold standard assures an accurate diagnosis, as was previously reported by our research group (7). Imaginologic evaluation is a diagnostic tool of irrefutable importance when needing further and careful assessment; however, previous detailed clinical evaluation should be performed in order to prescribe radiation-based imaginologic tools (18). Several international groups have published statements aiming to provide a recommendation about CBCT indication (19)(20)(21). The bene ts of using CBCT are widely known and recognized, though the amount of radiation used each scanning time is still a matter of controversy. When in need of a detailed image, optimization of radiation exposure may be performed. In cases of impacted and supplementary teeth, dentoalveolar trauma, orofacial defects dental anomalies and bone pathologies, where there is a justi ed reason for CBCT prescription, effective dose should be managed (21). Reduction of radiation risk could be achieved by managing X-Ray tube voltage and current (22), exposure time, FOV, number of projections and patient's shielding devices (19). Considering that the amount of radiation is directly related to the period of image taken and the eld of the vision (FOV) (23), diagnostic modalities with smaller eld are currently used. Limited cone-beam computed tomography (LCBCT) offers a small FOV (1.6 -3.1 inch/4 -8cm) with a high-resolution image with less radiation than a common CT (24,25). In this study, a 9-inch FOV was used to acquire images of the entire low jaw of the animal because several teeth were used for investigation. Clinically, a smaller FOV could be used if we consider that a single tooth should be examined.
This study has compared CBCT and PR's accuracy in diagnosing EARR in an animal model. Even though an animal model is a more adequate subject of study than in vitro specimens, some limitations may arise such as species' difference and sample size. The canine specie has more similarities in perioral and oral tissues to humans than other species such as mice and guinea pigs, however, still exists differences such as teeth morphology that should be considered prior endodontic treatment. Furthermore, radiographs should be taken under complete anesthesia and with a pre-design standardization device to obtain a valuable image in all subjects. Another possible limitation could be our small sample size. Since this research was performed in animals, the sample was the minimum necessary that were needed to prove our hypothesis, following the technical regulations of the International Organization of Standardization. However, this limitation was considered in statistics analysis.
We found that CBCT was more accurate and sensitive when compared to PR in detecting external apical root resorption. These ndings shed light on the use of CBCT for detection of initial root resorption that might occurs as a result of a dental trauma, pulp exposition or unmanaged orthodontic forces. Early identi cation of resorption allows a prompt treatment and reduces the risk of dental structure loss, nevertheless, CBCT should be prescribed under situations of strictly need and following effective dose optimization.

Conclusions
We demonstrated that CBCT showed higher accuracy to detect EARR. These ndings shed light on the use of CBCT for detection of initial root resorption. Early identi cation of resorption allows a prompt treatment and reduces the risk of dental structure loss.

Declarations
Con ict of Interest: Author Stephanie Diaz Huamán declares that she has no con ict of interest. Author Maya Fernanda Manfrin Arnez declares that she has no con ict of interest. Author Fernanda Maria Machado Pereira Cabral de Oliveira declares that she has no con ict of interest. Author Andiara De Rossi declares that she has no con ict of interest. Author Léa Assed Bezerra Silva declares that she has no con ict of interest. Author Francisco Wanderley Garcia Paula-Silva declares that he has no con ict of interest.