Gumussoy Ismail, Haylaz Emre, Duman Suayip Burak, Kalabalık Fahrettin, Eren Muhammet Can, Say Seyda, Celik Ozer, Bayrakdar Ibrahim Sevki
Department of Oral and Maxillofacial Radiology, Faculty of Dentistry, Sakarya University, Sakarya 54050, Turkey.
Department of Oral and Maxillofacial Radiology, Faculty of Dentistry, Inönü University, Malatya 44000, Turkey.
Diagnostics (Basel). 2025 Jul 4;15(13):1713. doi: 10.3390/diagnostics15131713.
The infraorbital canal (IOC) is a critical anatomical structure that passes through the anterior surface of the maxilla and opens at the infraorbital foramen, containing the infraorbital nerve, artery, and vein. Accurate localization of this canal in maxillofacial, dental implant, and orbital surgeries is of great importance to preventing nerve damage, reducing complications, and enabling successful surgical planning. The aim of this study is to perform automatic segmentation of the infraorbital canal in cone-beam computed tomography (CBCT) images using an artificial intelligence (AI)-based model. A total of 220 CBCT images of the IOC from 110 patients were labeled using the 3D Slicer software (version 4.10.2; MIT, Cambridge, MA, USA). The dataset was split into training, validation, and test sets at a ratio of 8:1:1. The nnU-Net v2 architecture was applied to the training and test datasets to predict and generate appropriate algorithm weight factors. The confusion matrix was used to check the accuracy and performance of the model. As a result of the test, the Dice Coefficient (DC), Intersection over the Union (IoU), F1-score, and 95% Hausdorff distance (95% HD) metrics were calculated. By testing the model, the DC, IoU, F1-score, and 95% HD metric values were found to be 0.7792, 0.6402, 0.787, and 0.7661, respectively. According to the data obtained, the receiver operating characteristic (ROC) curve was drawn, and the AUC value under the curve was determined to be 0.91. Accurate identification and preservation of the IOC during surgical procedures are of critical importance to maintaining a patient's functional and sensory integrity. The findings of this study demonstrated that the IOC can be detected with high precision and accuracy using an AI-based automatic segmentation method in CBCT images. This approach has significant potential to reduce surgical risks and to enhance the safety of critical anatomical structures.
眶下管（IOC）是一个重要的解剖结构，它穿过上颌骨的前表面并在眶下孔开口，包含眶下神经、动脉和静脉。在颌面、牙种植和眼眶手术中准确确定该管的位置对于预防神经损伤、减少并发症以及实现成功的手术规划至关重要。本研究的目的是使用基于人工智能（AI）的模型对锥束计算机断层扫描（CBCT）图像中的眶下管进行自动分割。使用3D Slicer软件（版本4.10.2；美国马萨诸塞州剑桥市麻省理工学院）对110例患者的220张眶下管CBCT图像进行了标注。数据集按照8:1:1的比例分为训练集、验证集和测试集。将nnU-Net v2架构应用于训练和测试数据集，以预测并生成合适的算法权重因子。使用混淆矩阵检查模型的准确性和性能。测试结果计算了骰子系数（DC）、交并比（IoU）、F1分数和95%豪斯多夫距离（95% HD）指标。通过对模型进行测试，发现DC、IoU、F1分数和95% HD指标值分别为0.7792、0.6402、0.787和0.7661。根据获得的数据绘制了受试者工作特征（ROC）曲线，曲线下面积（AUC）值确定为0.91。在手术过程中准确识别和保留眶下管对于维持患者的功能和感觉完整性至关重要。本研究结果表明，使用基于AI的自动分割方法可以在CBCT图像中高精度、准确地检测眶下管。这种方法具有显著的潜力来降低手术风险并提高关键解剖结构的安全性。
