Enhanced Cone-Beam Computed Tomography Imaging through Deep Learning Model Reconstruction: Noise Reduction and Image Quality Optimization in Dental Diagnostics

Abstract

Abstract To assess the impact of a vendor-agnostic deep learning model (DLM) on image quality parameters and noise reduction in dental cone-beam computed tomography (CBCT) reconstructions. A retrospective study was conducted on CBCT scans of patients from a single center, using the inclusion criteria of standard radiation dose protocol images. Objective image quality was assessed through contrast-to-noise ratio (CNR) measurements. Subjective quality was evaluated by two experienced readers using a five-point scale. The inter-reader reliability and repeatability were calculated. Thirty-seven patients were included in the study. The CNR levels in DLM reconstructions were significantly higher than in native reconstructions, and the mean CNR in ROI1 − 3 in DLM images was 11.12 ± 9.29, while for native reconstructions, it was 8.42 ± 5.89 (p < 0.05). However, there were no significant differences in the mean signal and noise levels between the two reconstruction methods. Subjective image quality assessment showed no statistically significant differences between native and DLM reconstructions. The use of deep learning-based image reconstruction algorithms in CBCT imaging of the oral tissues can improve the image quality by enhancing the contrast-to-noise ratio. This study underscores the potential of DLMs in improving dental diagnostic imaging and calls for further research on their clinical impact.