Abstract
This study aimed to identify predictors of intrasinusidal alveolar antral artery (AAA) localization at the upper left first molar (UL6) region and develop a clinical risk prediction model. A retrospective analysis of 243 eligible cases with cone-beam computed tomography (CBCT) imaging was conducted, categorizing AAA into intrasinusidal and non-intrasinusidal groups. Multivariable logistic regression revealed three independent predictors: anatomical variations (OR = 0.270, 95% CI 0.102-0.716), lateral Maxillary sinus wall width (OR = 0.583, 95% CI 0.371-0.915), and sinus cavity width (OR = 1.176, 95% CI 1.063-1.302). The nomogram-based prediction model demonstrated moderate discriminative capacity (C-index = 0.758) with satisfactory calibration alignment between predicted and observed outcomes. Receiver operating characteristic analysis yielded an area under the curve of 0.758. The sensitivity and specificity of the model were 0.789 and 0.579, respectively, with an accuracy of 69.0%.The findings demonstrate that when AAA is not visualized on CBCT scans, application of this model to assess the risk of intrasinusidal AAA occurrence in the surgical area enables preoperative planning of the surgical approach, thereby enhancing procedural safety.