To systematically evaluate the accuracy and generalizability of artificial intelligence (AI)–based segmentation of mandibular canal and related anatomical structures on cone-beam computed tomography (CBCT) and to examine methodological factors influencing reported performance. A systematic search of PubMed/MEDLINE, Scopus, Web of Science, Cochrane Library, and Wiley Online Library was conducted through April 10, 2026. Studies evaluating AI-based segmentation of the mandibular canal or related anatomical structures on CBCT were included. Risk of bias was assessed using QUADAS-2 tool. Random-effects meta-analysis using restricted maximum likelihood estimation was performed to pool segmentation performance, with the Dice Similarity Coefficient (DSC) as the primary outcome. Subgroup analyses, meta-regression, publication bias assessment, and GRADE certainty evaluation were conducted. Fifty-nine studies were included qualitatively, and 46 studies were eligible for quantitative synthesis. Across 40 effect sizes, the pooled DSC was 0.806 (95
To review the development pipeline of artificial intelligence (AI) in dental imaging, focusing on dataset design, annotation quality, model development, validation, and clinical translation. Published literature on AI applications in dental imaging, including machine learning, deep learning, and foundation-model-based approaches. Peer-reviewed studies addressing imaging modalities, dataset construction, annotation protocols, model architectures, performance evaluation, and clinical translation. Studies involving panoramic radiography, intraoral radiography, cone-beam computed tomography (CBCT), and intraoral scanning were reviewed, with emphasis on dataset quality, annotation methodology, model evaluation, interpretability, and clinical implementation. AI has demonstrated considerable potential for automated classification, detection, segmentation, and decision support in dental imaging. However, challenges including dataset heterogeneity, annotation inconsistency, domain shift, information leakage, interpretability, and regulatory requirements continue to limit clinical adoption. Emerging approaches such as multimodal learning and foundation models may improve generalizability and clinical applicability. This review highlights key methodological and translational considerations beyond algorithm performance, providing guidance for developing reliable and clinically meaningful AI systems in dental imaging.
To determine whether longitudinal skeletal muscle assessment improves prognostic evaluation for head and neck cancer-specific survival (HNC-Surv) and recurrence outcomes in radiotherapy-treated patients, and secondarily to compare prognostic performance of C3- versus L3-derived skeletal muscle indices. This prognostic study included 206 patients with HNC treated with curative-intent radiotherapy. Skeletal muscle index (SMI) was assessed at baseline and post-treatment imaging, and longitudinal metrics (percentage change and change rate) were calculated. Fine-Gray competing-risk regression was used for HNC-Surv, and Cox regression for locoregional recurrence-free survival (LRFS) and distant recurrence-free survival (DRFS). Discrimination was evaluated using time-dependent ROC analysis and compared using the DeLong test. Over a median follow-up of 5.6 years, 68 deaths occurred, including 52 HNC-related deaths; 59 patients (29
The aim of this study was performed to investigate the salivary gland SPECT/CT of parotid glands in patients with lymphoepithelial sialadenitis including preliminary study on clinical and pathological features. The study was performed in 5 patients with lymphoepithelial sialadenitis who underwent salivary gland SPECT/CT. Correlation between maximum standardized uptake value (SUV) at ratio of pre- to post-stimulation and clinicopathological parameters in lymphoepithelial sialadenitis were performed by Pearson correlation coefficient. The maximum SUV of 10 parotid glands in 5 patients with lymphoepithelial sialadenitis using salivary gland SPECT/CT was 23.8 ± 16.9 at pre-stimulation, 10.5 ± 5.1 at post-stimulation, and 2.06 ± 0.66 at ratio of pre- to post-stimulation. Furthermore, the maximum SUV at ratio of pre- to post-stimulation was significantly correlated with Saxon test score (r = 0.637, p = 0.048). The salivary gland SPECT/CT can be useful in clinical practice for the quantitative management of parotids in patients with lymphoepithelial sialadenitis.
To develop and validate a logistic regression model that integrates clinical and fluorodeoxyglucose positron emission tomography/computed tomography (FDG-PET/CT)-derived radiomic features to predict cervical lymph node metastasis in tongue squamous cell carcinoma. This retrospective study included 316 patients who underwent preoperative FDG-PET/CT between December 2013 and December 2023, classified into analysis (n = 250) and validation (n = 66) cohorts. Primary tumours were segmented using standardised uptake value (SUV)-based thresholds (≥ 40
