Investigation of the Use of Artificial Intelligence in Anterior Loop Detection: A Panoramic Radiography Study
Diagnostics, cilt.16, sa.14, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 16 Sayı: 14
- Basım Tarihi: 2026
- Doi Numarası: 10.3390/diagnostics16142213
- Dergi Adı: Diagnostics
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, EMBASE, Directory of Open Access Journals, Academic Search Ultimate (EBSCO), Biomedical Reference Collection: Corporate Edition (EBSCO)
- Anahtar Kelimeler: anterior loop, artificial intelligence, deep learning, mandibular canal, panoramic radiography
- Ankara Üniversitesi Adresli: Evet
Özet
Background/Objectives: The accurate detection of the anterior loop (AL) of the inferior alveolar nerve is critical to avoid neurosensory complications during surgical procedures in the interforaminal region, and panoramic radiography continues to be widely used in routine dental diagnostics due to its accessibility and cost-effectiveness. This study aimed to evaluate the performance of a deep learning approach in automatic detection of the AL in panoramic radiographs. Methods: A total of 305 anonymised panoramic radiographs containing 413 annotated ALs were used to train a YOLOv8x-based model for automatic AL detection. The dataset was divided into training, validation, and test sets consisting of 245 images (332 AL annotations), 30 images (40 AL annotations), and 30 images (41 AL annotations). Labelling was carried out by using the polygonal annotation method. The model’s performance in identifying the AL region was measured using precision, recall, F1 score, and mean average precision (mAP@0.5). Results: The model achieved a precision of 0.75, a recall of 0.6585, and a F1 score of 0.7013. The average precision at an intersection over union (IoU) threshold of 0.5 (mAP@0.5) was 0.739. Conclusions: This study demonstrates the feasibility of using a YOLOv8x-based detection model to detect ALs in panoramic radiographs. Although further improvements are needed to enhance model sensitivity and generalisability, the findings demonstrate the potential to support clinical decision-making.