AI-driven analysis of jaw-bone alterations in CBCT images associated with systemic diseases: A systematic review

Objective: This systematic review aims to evaluate the diagnostic performance of artificial intelligence (AI) algorithms in detecting jawbone alterations associated with systemic diseases, primarily osteoporosis (OP) and temporomandibular joint osteoarthritis (TMJ OA) using cone-beam computed tomography (CBCT) images. Methods: Following PRISMA guidelines, a comprehensive search was conducted across multiple databasesStudies were included if they applied AI techniques to CBCT scans for detecting jawbone changes related to systemic bone conditions. Data were extracted on study design, AI architecture, population, diagnostic performance, and risk of bias using QUADAS-2 and an AI-specific quality checklist. Results: Seven studies published between 2018 and 2024 met the inclusion criteria. AI models such as convolutional neural networks (CNNs), you only look once (YOLO), artificial neural networks (ANNs), and ensemble methods were employed. These models demonstrated high diagnostic performance, with accuracies ranging from 76.78 % to 98.85 %, sensitivities up to 100 %, and specificities above 94 %. OP-related studies focused on mandibular bone density and cortical morphology, while TMJ OA studies targeted condylar changes such as flattening, erosion, and osteophytes. Most models were trained on single-center datasets with limited external validation. Risk-of-bias analysis revealed common concerns in patient selection and reference standard reporting. Conclusion: AI models, particularly DL-based architectures, show strong potential for identifying osseous changes associated with systemic diseases on CBCT images. These findings highlight AI's potential to support early detection of jawbone changes and enable screening for systemic diseases during routine dental imaging.