Diagnostic accuracy of three CBCT protocols and Metal Artifact Reduction tool for detecting the second canal of mandibular incisors adjacent to a zirconia implant
Maria Fernanda Silva Andrade-Bortoletto, Thaísa Pinheiro Silva, Marcela Tarosso Réa, Renata Maira de Souza Leal, Débora Costa Ruiz, Deborah Queiroz FreitasAbstract
Objective
To compare the diagnostic accuracy of three cone beam computed tomography (CBCT) acquisition protocols, one low-resolution (LR) and two high-resolution (HR) protocols, using a metal artifact reduction tool (MAR) for detecting a second canal in mandibular incisors adjacent to a zirconia implant.
Material and Methods
Twenty single-rooted mandibular incisors were included, twelve with one canal and eight with two canals confirmed by micro-CT. The roots and a zirconia implant were placed into the alveolar sockets of the right and left mandibular central incisors of a dry human mandible, respectively. In total, 240 CBCT scans were obtained using FOV 5x5 cm, 90 kVp, with and without implant, for each root. The variable parameters included the MAR tool, one LR-protocol (0.2 mm voxel, 5 mA), and two HR-protocols (0.085 mm voxel, 4 and 10 mA). Five calibrated examiners independently evaluated the scans. An ordinal cumulative link mixed model (CLMM) with crossed random intercepts for roots and examiners was used for data analysis (α = 5%).
Results
The CLMM revealed a significant main effect for implant presence (p = 0.049) and canal status (p < 0.001), but not for protocol (p = 0.135) or MAR (p = 0.392). Significant protocol x canal status (p < 0.001) and implant x canal status (p < 0.001) interactions were found, whereas MAR interactions were not significant (p > 0.05). In two-canal roots, the zirconia implant significantly induced false-negative shifts under HR-4 (p < 0.0001) and LR (p < 0.0001) protocols, whereas HR-10 preserved sensitivity (p = 0.7760). In single-canal roots, implants induced false-positive shifts in HR-10 (p = 0.0011) and HR-4 (p = 0.0011).
Conclusion
Zirconia implants impair second canal detection via false-negative and false-positive errors. High-resolution CBCT (HR-10) should be prioritized to preserve diagnostic sensitivity, whereas MAR offers no clinical benefit and cannot compensate for reduced resolution or exposure.
Advances in knowledge
CBCT diagnostic performance near zirconia implants relies on acquisition protocol rather than MAR activation. High-resolution CBCT (HR-10) mitigates implant-induced false negatives, whereas lower-resolution or lower-milliamperage protocols compromise diagnostic accuracy.