Comparison between dual-angle periapical technique and cone-beam computed tomography in the identification of a second canal in mandibular incisors in the presence of a zirconia implant: An in vitro study Authors: Maria Fernanda Silva Andrade-Bortoletto, Fernanda Rodrigues de Oliveira, Thaísa Pinheiro Silva, Marcela Tarosso Réa, Renata Maira de Souza Leal, Débora Costa Ruiz, Deborah Queiroz Freitas Published: 2026 ABSTRACT AbstractPurposeTo compare the accuracy of the dual-angle periapical technique (DPT) and CBCT acquired with two protocols in identifying lower incisors with a second canal adjacent to a zirconia implant.Materials and MethodsTwenty single-rooted lower incisors were used, twelve with one canal and eight with two canals, confirmed with micro-CT. These roots and a zirconia implant were positioned in the alveolus of the mandibular right and left central incisors of a dry human mandible, respectively. Periapical radiographs were obtained using a Focus X-ray unit and a photostimulated phosphor plate (VistaScan system) with parameters of 70 kVp, 7 mA, and 0.125 seconds of exposure time. For each root, one standard and one with the tube shifted distally in 30°, with and without implant, totaling 80 radiographs. In addition, 80 CBCT scans were obtained with the OP300 Maxio (FOV 5×5 cm, 90 kVp), with and without implants, for each root. The variable parameters were: low-resolution (LR) (5 mA, 0.2 mm voxel) and high-resolution (HR) (10 mA, 0.085 mm voxel). Five evaluators assessed all images using a 5-point confidence scale. Two-way ANOVA with Sidak’s post hoc test was used (P<0.05). Intra- and inter-examiner agreement were assessed with weighted kappa coefficients.ResultsDPT and both CBCT protocols(LR and HR) showed comparable AUC, sensitivity, and specificity (P>0.05). DPT and both CBCT protocols presented AUC, sensitivity, and specificity unaffected by the implant(P>0.05).ConclusionDPT and both CBCT protocols showed comparable diagnostic accuracy of the second canal regardless of zirconia implant presence.Keywords: Cone-Beam Computed Tomography; Diagnostic Imaging; Incisor; Radiography, Dental Introduction The adequate identification of internal root morphology is essential before endodontic treatment, as major causes of endodontic treatment failure include root dilacerations and accessory canals, which may compromise the instrumentation, disinfection, and filling of the root canal system.1 This risk is especially relevant in mandibular incisors, which often present anatomical variations, including the presence of two canals with different anatomical configurations.2,3 According to Vertucci’s classification, type I, characterized by a single root canal extending from the pulp chamber to the apex, is the most prevalent canal configuration in mandibular incisors, with an incidence ranging from 59.5% to 99.6%. Type II, characterized by two separate canals leaving the pulp chamber and merging into a single canal before reaching the apical foramen, presents an incidence ranging from 0% to 5%. Type III, characterized by the division and subsequent confluence of the root canal, is considered the second most prevalent configuration in these teeth, with an incidence ranging from 0.6% to 33.0%, depending on the studied population. The prevalence of Vertucci types IV to VIII is considered negligible, generally close to 0%.3,4,5,6,7,8 From a clinical standpoint, undetected or uninstrumented canals may harbor residual necrotic tissue or microorganisms, leading to persistent periapical inflammation and treatment failure.9 Therefore, accurate preoperative assessment using reliable imaging modalities is essential to prevent iatrogenic errors and to ensure adequate cleaning, shaping, and obturation of the entire canal system. Periapical radiography, which is often used to perform the initial evaluation of endodontic treatments, may not contribute to the visualization of accessory canals due to the overlap of anatomical structures.10 In mandibular incisors, this occurs because the bifurcation of the canals results in one branch extending buccally and the other lingually, leading to their superimposition in orthogonal imaging.11 Therefore, dental surgeons often request cone-beam computed tomography (CBCT) prior to treatment to perform this assessment, as it has proven to be a promising tool in identifying anatomical variations, especially in cases of oval or multiple canals, since it provides multiplanar visualization of anatomical structures without overlaps or distortions.2,12,13 However, within periapical radiographic methods, the Clark technique, based on modifying the X-ray beam angulation to dissociate superimposed structures, presents potential for improving the detection of accessory canals.14 Furthermore, for endodontic purposes, CBCT scans are usually acquired using high-resolution (HR) protocols to allow detailed visualization of fine root canal structures.15 In such cases, the effective dose from a small FOV (5 cm×5 cm) CBCT scan of the lower incisors is approximately 9 to 125 times higher than that of a periapical radiograph (1.4 µSv). This was demonstrated using both low-resolution (LR) (12 µSv) and HR (175 µSv) protocols.16 Additionally, despite technological advances in CBCT, materials with high atomic number and physical density, such as dental implants, can cause beam hardening artifacts, degrading image quality. These artifacts appear as alternating hypodense bands and hyperdense streaks that do not represent the actual anatomy, resulting from reconstruction errors.17 Consequently, they may hinder image interpretation and compromise diagnoses of adjacent teeth, including root fractures and external cervical resorption.18,19 Given this, this study aimed to compare the diagnostic accuracy between the dual-angle periapical technique (DPT) and CBCT (using HR and LR protocols), in identifying second canals in mandibular incisors adjacent to a zirconia implant. Materials and Methods This study was approved by the Local Research Ethics Committee. Besides, this research complied with the World Medical Association’s Declaration of Helsinki. Sample selection and preparation A dry human mandible of approximately 60 years of age, with preserved alveoli and 20 mandibular incisors extracted for reasons unrelated to the study, was selected to compose the phantom. Teeth that had undergone previous endodontic treatment, presented coronal and root restorative materials, deep caries, internal or external root resorption, root fracture, supernumerary roots, obliterated root canals, and pulp calcifications were excluded from the sample. The teeth were verified through visual, radiographic, and micro-computed tomographic imaging to confirm that they met the ideal criteria for use in the study. In total, 12 teeth presented a single canal, and 8 teeth presented two canals. According to the Vertucci classification, the teeth with two canals corresponded to type III configurations. The sample size was defined based on previously published diagnostic studies with similar experimental designs, including both healthy and affected groups.19,20,21 The number of canals was confirmed with micro-computed tomographic (Skyscan 1174; Bruker, Kontich, Belgium) acquired with the following acquisition parameters (50 kV, 800 µA, voxel size of 29.1 µm, filter of 0.5 mm aluminum, rotation steps 0.5°, 1 frame averaging, and arch rotation of 180°), which served as a reference standard (Fig. 1). Representative micro-computed tomographic (micro-CT) images of the root canal system in two mandibular incisors. Micro-CT imaging served as the reference standard and confirmed the number of canals of the entire sample. A. Axial reconstruction of a root with a single canal. B. Sagittal reconstruction of a root with a single canal. C. Axial reconstruction of a root with two canals. D. Sagittal reconstruction of a root with two canals. The crowns of the 20 mandibular incisors were removed at the cementoenamel junction to standardize the specimens and minimize potential observer bias. Maintaining the natural crowns could allow examiners to recognize individual teeth based on anatomical characteristics, potentially leading to memorization and artificially increased diagnostic performance. A diamond disc (Extec Corp, Enfield, USA) was used at a speed of 400 rpm on Isomet 1000 cutting machine (Buhler, Lake Bluff, USA). The 20 roots were positioned one at a time in the alveolus of the right mandibular central incisor. To evaluate the effect of the presence of materials with high atomic number and physical density on the diagnosis of incisors with two canals, a zirconia implant (Z-Look 3; Z-Systems, Oensingen, Switzerland) was inserted into the alveolus of the left mandibular central incisor. Extracted human teeth were selected and positioned in the alveoli corresponding to the mandibular canines and lateral incisors of a dry human mandible, simulating a clinical condition in which patients present teeth in the anterior region. Image acquisition For DPT, the dry human mandible was fixed on an acrylic support using a locator device to standardize acquisitions. The standardized geometry included a focus-receptor distance of 30 cm, object-receptor distance of 1.5 cm, and 0° vertical angulation. Furthermore, a condensation silicone was used to fabricate a custom support for the image receptor to ensure its parallelism with the teeth of the anterior region, and to guarantee that the radiation beam was perpendicular to the teeth and image receptor for the standard periapical radiographs. A 30° horizontal tube-shift angulation was selected based on previous literature and preliminary assessments performed before the experimental phase, in which this angulation provided greater separation of superimposed root canals.22,23 To acquire the radiographs with horizontal tube shift, a protractor was used to measure the 30° horizontal angulation for the radiographs obtained with the X-ray tube shifted distally. The protractor was positioned on the acrylic platform supporting the mandible, with its 90° reference aligned parallel with the mandibular midline. Subsequently, the 30° angulation was marked on the acrylic platform supporting the mandible. During image acquisition, the cylinder was aligned with the previously marked 30° line on the platform, ensuring a centered position relative to the reference mark and standardized angulation across all acquisitions. In addition, to simulate the effects of radiation attenuation and scattering caused by soft tissues in intraoral radiographs, a 2.5 cm thick acrylic resin plate was placed between the mandible and the X-ray machine. Periapical radiographs were obtained with the Focus X-ray machine (Instrumentarium Dental, Tuusula, Finland), using photostimulated phosphor (PSP) plates from the VistaScan system (Dürr Dental, Bietigheim-Bissingen, Germany). Periodic calibration procedures are routinely performed for the X-ray equipment used in our institution, and the calibration of the device was up to date during the experimental phase. Images of the 20 roots were acquired using the following exposure parameters: 70 kVp, 7mA, and 0.125 seconds of exposure time. Two acquisition protocols were performed per root as follows: one acquisition without any metallic material in the alveolus of the left mandibular central incisor (control); one acquisition with the zirconia implant positioned in the alveolus of the left mandibular central incisor. In total, 80 periapical radiographs were obtained [20 roots×2 acquisition protocols×2 periapicals (standard and tube shifted distally)] (Fig. 2). Periapical radiographs acquired with dual-angle periapical technique (standard and tube shifted distally) of a mandibular incisor with one canal, and another with two canals, according to the different conditions (with and without implant). Yellow arrows indicate the location of the root canal(s). The OP300 Maxio unit (Instrumentarium Dental, Tuusula, Finland) was used to perform the cone-beam computed tomography acquisitions. The device underwent routine periodic calibration procedures at the present institution, and its calibration status was verified and up to date during the experimental phase. The anterior region of the phantom was positioned in the center of the field of view (FOV), and its position was standardized using the equipment’s guide lights as a reference. The phantom was fixed with wax inside a plastic container filled with water to simulate the attenuation of the X-ray beam by soft tissues. The CBCT images were acquired with the fixed exposure parameters of 90 kVp, 5 cm×5 cm FOV. The variable exposure parameters were: LR protocol - 5 mA and 0.2 mm voxel, and HR protocol - 10 mA and 0.085mm voxel. Four acquisition protocols were performed per root as follows: one acquisition without any metallic material in the alveolus of the left mandibular central incisor (control) using the LR protocol; one acquisition without metallic material in the same alveolus using the HR protocol (control). Additionally, two acquisitions were obtained with a zirconia implant positioned in the alveolus of the left mandibular central incisor, using the LR and HR protocols. This way, 80 CBCT scans were obtained (20 roots×4 acquisition protocols)(Fig. 3). Axial reconstructions of a mandibular incisor with one canal and another with two canals according to the different acquisition protocols (low- and high-resolution) and conditions (without and with implant). Yellow arrows indicate the location of the root canal(s). CBCT: cone-beam computed tomography. Image evaluation Five oral radiologists with more than 5 years of experience in the interpretation of CBCT scans and periapical radiographs were previously calibrated for this evaluation using CBCT scans and periapical radiographs that did not belong to the present sample. The evaluators worked independently and were instructed to first assess the periapical radiographs and, upon completion, to proceed with the evaluation of the CBCT scans, always in a quiet, dimly lit room. Each evaluator used their own personal computer, all of which were considered suitable for dental image evaluation. This approach was considered acceptable, as previous studies have shown comparable diagnostic performance across different display devices.24,25 In the CBCT assessments, the evaluators analyzed the images throughout the entire root length using multiplanar reconstructions, including axial, coronal, and sagittal views. They were blinded regarding the proportion of roots with one and two canals, as well as the CBCT protocols used. The DPT consists of evaluating a pair of periapical radiographs: one standard and one acquired with the X-ray tube shifted distally or mesially. Therefore, the evaluators received 40 randomized and coded folders containing both images (standard and with tube shifted distally) in TIFF format for each case. For this evaluation, the JPEG View version 1.0.37 imaging software (public domain software) was used. The evaluators were instructed to evaluate a maximum of 20 cases per day, with a 24-hour interval between evaluations. They could use zoom, brightness, and contrast tools to evaluate the images. The presence or absence of two canals was evaluated using a 5-point scale: 1 (definitely absent), 2 (probably absent), 3 (uncertain), 4 (probably present), and 5 (definitely present). The CBCT images were exported in DICOM format. Afterward, the images were randomized and coded for subsequent evaluation by the same five oral radiologists. They were instructed to evaluate a maximum of 20 volumes per day, with a 24-hour interval between evaluations. It was recommended to evaluate in a quiet environment, using the OnDemand 3D software (Cybermed, Seoul, Korea). They were allowed to use zoom, brightness, and contrast tools to evaluate the images. The evaluators indicated the presence of two canals based on a 5-point scale: 1 (definitely absent), 2 (probably absent), 3 (uncertain), 4 (probably present), and 5 (definitely present). Twenty days after the end of the evaluation, 20% of the sample was reevaluated under the same conditions to verify intra-examiner agreement. Statistical analysis All statistical analyses were performed using SPSS version 25.0 (IBM, Armonk, NY) with a significance level of 5%. The diagnostic values of area under the ROC curve, sensitivity, and specificity were calculated for each examiner by comparing their responses with the standard reference information. Two-way ANOVA and Sidak’s post-hoc test were used to verify the influence of study factors on the diagnosis of the two canals. Intra- and inter-examiner agreement was assessed using weighted Kappa. A post-calculation demonstrated that the sample size of the present study ensured a power of 80% for the area under the receiver operating characteristic curve (AUC) and specificity, and 85% for sensitivity. The null hypothesis considered that there was no difference in the diagnostic accuracy of the second canal in mandibular incisors would be observed when comparing the examination modalities (DPT and CBCT - LR and HR). Results Table 1 shows the AUC, sensitivity, and specificity values for the diagnosis of the second canal in mandibular incisors using the dual-angle periapical technique (DPT) and the LR and HR CBCT protocols. No significant differences were observed among the imaging modalities or between implant conditions for any of the diagnostic performance measures (P>0.05). Table 1Diagnostic performance of the dual-angle periapical technique and low-(LR) and high-resolution (HR) cone-beam computed tomography (CBCT) protocols for the detection of a second canal in mandibular incisorsAUC: area under the receiver operating characteristic curve. No statistically significant differences were observed between imaging modalities or implant conditions for AUC, sensitivity, or specificity (P>0.05). Diagnostic performance of the dual-angle periapical technique and low-(LR) and high-resolution (HR) cone-beam computed tomography (CBCT) protocols for the detection of a second canal in mandibular incisors AUC: area under the receiver operating characteristic curve. No statistically significant differences were observed between imaging modalities or implant conditions for AUC, sensitivity, or specificity (P>0.05). According to a guideline, the sum of the sensitivity and specificity of a diagnostic test must be at least 1.5 to consider that it has the potential to adequately distinguish between abnormal and normal results, and this was achieved by DPT and both HR and LR CBCT protocols, with and without the zirconia implant, ranging from 1.69 to 1.94.26 The AUC values reinforce the idea that the DPT and both the HR and LR CBCT protocols are excellent tools for discriminating between one and two canals.27 Inter-examiner agreement for DPT ranged from fair to substantial (0.38 to 0.80), and for CBCT from moderate to substantial (0.47 to 0.79). The intra-examiner agreement for DPT ranged from substantial to almost perfect (0.66 to 0.91), and for CBCT ranged from substantial to almost perfect (0.77 to 1).28 Discussion Mandibular incisors represent one of the most anatomically complex tooth groups due to the high prevalence of root canal morphological variations.29 The frequent presence of undetected canals is a major clinical concern, as uninstrumented areas may retain pulp tissue or bacteria, ultimately leading to endodontic failure.30 Therefore, assessing accurate diagnostic methods to identify the presence of two canals in mandibular incisors is of significant clinical importance. In the present study, neither LR nor HR CBCT protocols demonstrated significantly higher diagnostic accuracy for detecting the second canal, regardless of implant presence. Based on these results, we failed to reject the null hypothesis. Considering that CBCT involves a higher radiation dose than two periapical radiographs, these results suggest that a reliable diagnosis can be achieved with minimal radiation exposure using conventional periapical radiography.16 Beyond radiation safety, the DPT offers other practical advantages. It is a simple procedure that can be performed by clinicians, including undergraduate students, using standard intraoral radiographic equipment.31 Additionally, the financial cost of acquiring two periapical radiographs is considerably lower than that of a CBCT exam, further emphasizing the effectiveness and clinical applicability of this conventional approach. However, the presence of high-density materials interferes with the automatic exposure compensation (AEC) system of digital sensors, leading to alterations in the grey values of dental tissues such as pulp and dentin, which may compromise the diagnostic assessment of these structures, such as the presence of a second canal.32 The authors initially hypothesized that CBCT might outperform the DPT in the absence of implants due to its ability to visualize anatomical structures without superimposition. Conversely, they expected that the DPT would perform better in the presence of implants, even knowing that they could influence the AEC (Automatic Exposure Compensation), given that these materials can cause beam hardening artifacts that compromise CBCT image quality.17 Prior research has demonstrated the adverse effects of high-density materials such as gutta-percha, metallic posts, and orthodontic brackets on diagnostic accuracy.18,33,34,35 The artifacts caused by these materials may either obscure or simulate root fractures, root resorptions, and caries lesions, thereby hindering diagnosis and compromising prognosis.18,36,37 Interestingly, the current findings indicate that the presence of implants did not affect CBCT diagnostic accuracy for detecting the second canal in mandibular incisors. A plausible explanation for this observation is that, although the secondary canals are typically narrow and may exhibit distinct anatomical configurations, they are generally less subtle and easier to visualize than root fractures, particularly when the fracture has not led to fragment separation and manifests only as a fine fissure. Under these conditions, even when beam hardening artifacts are produced by nearby implants, they are unlikely to obscure the secondary canal sufficiently to compromise detection. Consequently, the visualization of the second canal remained unaffected despite the presence of artifacts. Since all tested imaging modalities yielded similar accuracy, the authors recommend the use of two periapical radiographs. Few studies have compared CBCT and periapical radiographic techniques using vertical tube-shift angulations for identifying the relationship between impacted third molars and the mandibular canal;22,38,39 however, their results remain controversial. Some studies demonstrated good agreement with CBCT,38,39 while others concluded that the technique was not reliable for predicting this anatomical relationship.22 In endodontics, a previous study demonstrated that a 20° horizontal mesial angulation improved the detection and visualization of the fourth canal in mandibular first molars, although without comparison to CBCT.40 Similarly, in the present study, the use of dual-angle periapical radiographs promoted greater separation of superimposed root canals, which may explain the comparable diagnostic performance observed between DPT and CBCT for the identification of second canals in mandibular incisors adjacent to zirconia implants. The present recommendation aligns with the ALADAIP principle (As Low As Diagnostically Acceptable, being Indication-oriented and Patient-specific), which underscores the importance of selecting the most appropriate imaging modality for each diagnostic task.41 This principle advocates not only for optimization of exposure parameters such as milliampere and kilovoltage but also for choosing the least invasive imaging method capable of providing diagnostically sufficient information.41 Historically, the DPT was introduced by Clark in 1910 to determine the position of unerupted teeth by altering projection angles.14 Decades later, Goerig and Neaverth expanded its use in endodontics, highlighting its value for identifying root resorptions, fractures, and perforations, as well as for pre-surgical localization of roots or apices.42 Despite the rapid advancement of three-dimensional imaging technologies, the DPT remains a simple, low-cost, and diagnostically valuable method. Its clinical relevance continues to be demonstrated, for example, as a reference standard when evaluating panoramic radiographs for localizing impacted maxillary canines.43 In the present in vitro design, 20 roots were radiographed twice (standard and tube shifted distally) and scanned twice with CBCT under low- and high-resolution protocols. This setup ensured standardized imaging conditions without exposing patients to unnecessary radiation. To better simulate in vivo conditions, a 2.5 cm thick acrylic resin plate was placed between the mandible and the X-ray machine during the acquisition of the periapical radiographs.44 For the CBCT scans, the mandible was immersed in water to reproduce the soft-tissue equivalent attenuation.44 Moreover, the findings are reliable, given that the intra-examiner agreement ranged from substantial to almost perfect, demonstrating that the methodology provided consistent and reproducible assessments. Despite the variability in inter-examiner agreement, probably caused by the variety of response options of the 5-point scale, the resulting high diagnostic values confirm that the DPT and both CBCT protocols reliably differentiated between one and two canals.27 Based on the findings, further investigations comparing DPT and CBCT are warranted. Additionally, although image interpretation in the present study was performed by experienced oral radiologists, who are specifically trained in diagnostic imaging assessment, the absence of endodontists among the observers may represent a limitation, as clinical decision-making and canal interpretation could differ according to professional background. Future studies including both radiologists and endodontists may provide complementary perspectives and enhance the clinical applicability of the results. Furthermore, future investigations could also explore diagnostic performance under more clinically realistic restorative conditions, such as the presence of full-coverage crowns or intracoronal restorations, which may generate artifact patterns different from those produced by adjacent implants. Additionally, evaluating the effect of tools such as Metal Artifact Reduction (MAR) algorithms on low-dose CBCT exams could provide insights into optimizing diagnostic performance while minimizing radiation exposure. Some studies have reported improvements in vertical root fracture detection near artifact-generating materials when MAR was activated, although similar benefits were not observed for horizontal fractures.34,45 Exploring whether MAR activation can produce significant differences between high- and low-dose CBCT protocols would also be clinically meaningful. Finally, since the present study was limited to one CBCT unit and one PSP receptor system, additional studies involving different imaging equipment and systems are recommended to evaluate the potential variability in diagnostic outcomes across systems. In conclusion, the findings indicate that both DPT and CBCT, whether using LR and HR protocols, performed well and similarly in terms of diagnostic accuracy regardless of the presence of the zirconia implant. Both imaging modalities’ diagnostic values were unaffected by the presence of the zirconia implant. Since DPT and CBCT demonstrated comparable diagnostic performance irrespective of the presence of a zirconia implant, DPT may be considered a suitable alternative for detecting a second canal in this clinical scenario, as it aligns with the principles of radiological protection and dose optimization. CBCT should be reserved for cases where DPT is inconclusive and multiplanar information is strictly necessary. Acknowledgments The authors used ChatGPT (OpenAI) solely for English language editing and grammar improvement during some parts of this manuscript. The authors take full responsibility for the content of the manuscript. Footnotes Conflicts of Interest: None References