Photon-Counting Detector CT-Based Virtual Monoenergetic Imaging With Metal Artifact Reduction for Endodontic Diagnostics

Introduction: This ex vivo study aimed to determine the optimal energy level for virtual monoenergetic imaging (VMI) using photon-counting detector computed tomography (PCD-CT) and to evaluate the effectiveness of iterative metal artifact reduction (iMAR) in assessing simulated endodontic challenges and complications. Methods: Sixteen extracted third molars were simulated with one of eight distinct endodontic diagnostic challenges and imaged using PCD-CT at radiation doses equivalent to standard-dose cone-beam CT. VMIs were reconstructed from 70-190 keV at 10 keV increments, both with and without iMAR. Diagnostic accuracy, depiction quality of endodontic challenges, artifact severity, and visualization of key endodontic anatomical structures were independently assessed by 2 observers using a 5-point visual analogue scale (1 = least favourable, 5 = most favourable). Descriptive statistics were calculated, and inter-reader agreement was analysed using Krippendorff's alpha coefficient. Results: VMIs achieved excellent diagnostic accuracy (97%) and high-quality visualization of endodontic challenges (median: 5, IQRs: 4-5 or 4.25-5; α = 0.63-1.00) across the entire reconstructed energy spectrum (70-190 keV), with minimal artifacts, particularly at ≥ 110 keV (α = 1.0). Task-specific analysis demonstrated optimal visualization of caries at 70-80 keV and fractured files at 70-100 keV. For other pathologies, VMI at ≥100 keV effectively reduced artifacts without compromising anatomical detail. IMAR did not improve image quality and consistently reduced diagnostic performance. Conclusions: VMI from PCD-CT provides high-quality imaging with minimal artifacts, well-suited for indication-specific endodontic diagnostics. Clinical relevance: PCD-CT supports novel, indication-specific workflows in endodontic imaging, potentially enhancing diagnostic precision and long-term treatment follow-up.