Abstract:

This consensus was developed to standardize artificial intelligence-assisted imaging detection, data annotation, model validation, and clinical application for alveolar bone fenestration and dehiscence. A multidisciplinary expert panel systematically reviewed the relevant literature, assessed the quality of evidence using the GRADE approach, and formulated recommendations through three rounds of a modified Delphi process and external review. Standardized recommendations were proposed for thin alveolar bone assessment, metallic artifact management, selection of two-dimensional and three-dimensional models, multimodal fusion, external validation, clinical decision support, and data governance, thereby establishing a framework for image acquisition, hierarchical annotation, algorithm evaluation, and clinical application. The consensus rate for the core recommendations was 91.3%. Artificial intelligence may assist in the detection and risk assessment of bone fenestration and dehiscence but cannot replace specialist judgment or direct surgical examination. This consensus provides guidance for dataset development, model development, product evaluation, and clinical application.

Key words: artificial intelligence, cone-beam computed tomography(CBCT), bone fenestration, bone dehiscence, alveolar bone, expert consensus