Objectives: To develop and validate a modified deep learning (DL) model based on nnU-Net for classifying and segmenting five-class jaw lesions using cone-beam CT (CBCT). Methods: A total of 368 CBCT scans (37 168 slices) were used to train a multi-class segmentation model. The data underwent manual annotation by two oral and maxillofacial surgeons (OMSs) to serve as ground truth. Sensitivity, specificity, precision, F1-score, and accuracy were used to evaluate the classification ability of the model and doctors, with or without artificial intelligence assistance. The dice similarity coefficient (DSC), average symmetric surface distance (ASSD), and segmentation time were used to evaluate the segmentation effect of the model. Results: The model achieved the dual task of classifying and segmenting jaw lesions in CBCT. For classification, the sensitivity, specificity, precision, and accuracy of the model were 0.871, 0.974, 0.874, and 0.891, respectively, surpassing oral and maxillofacial radiologists (OMFRs) and OMSs, approaching the specialist. With the model's assistance, the classification performance of OMFRs and OMSs improved, particularly for odontogenic keratocyst (OKC) and ameloblastoma (AM), with F1-score improvements ranging from 6.2% to 12.7%. For segmentation, the DSC was 87.2% and the ASSD was 1.359 mm. The model's average segmentation time was 40 ± 9.9 s, contrasting with 25 ± 7.2 min for OMSs. Conclusions: The proposed DL model accurately and efficiently classified and segmented five classes of jaw lesions using CBCT. In addition, it could assist doctors in improving classification accuracy and segmentation efficiency, particularly in distinguishing confusing lesions (eg, AM and OKC).

Background: Medication-related osteonecrosis of the jaw (MRONJ) is a well-known severe adverse reaction of antiresorptive, antiangiogenic or targeted therapies, and usually occurs after tooth extraction. This review is aimed at determining the efficacy of any intervention of tooth extraction to reduce the risk of MRONJ in patients taking antiresorptive drugs, and present the distribution of evidence in these clinical questions. Methods: Primary studies and reviews were searched from nine databases (Medline, EMBase, Cochrane Library, Scopus, WOSCC, Inspec, KCI-KJD, SciELO and GIM) and two registers (ICTRP and ClinicalTrials.gov) to 30 November 2022. The risk of bias was assessed with the ROBIS tool in reviews, and the RoB 2 tool and ROBINS-I tool in primary studies. Data were extracted and then a meta-analysis was undertaken between primary studies where appropriate. Results: Fifteen primary studies and five reviews were included in this evidence mapping. One review was at low risk of bias, and one randomized controlled trial was at moderate risk, while the other eighteen studies were at high, serious or critical risk. Results of syntheses: (1) there was no significant risk difference found between drug holiday and drug continuation except for a subgroup in which drug continuation was supported in the reduced incidence proportion of MRONJ for over a 3-month follow-up; (2) the efficacy of the application of autologous platelet concentrates in tooth extraction was uncertain; (3) there was no significant difference found between different surgical techniques in any subgroup analysis; and (4) the risk difference with antibacterial prophylaxis versus control was −0.57, 95% CI −0.85 to −0.29. Conclusions: There is limited evidence to demonstrate that a drug holiday is unnecessary (and may in fact be potentially harmful) in dental practice. Primary closure and antibacterial prophylaxis are recommended despite limited evidences. All evidence have been graded as either of a low or very low quality, and thus further high-quality randomized controlled trials are needed to answer this clinical question.