YOLO11 deep learning model estimates gonial angle on panoramic X-rays with high accuracy

In a retrospective study of 1000 dental panoramic X-rays, a YOLO11 deep learning model estimated gonial angle with mean absolute error 2.37° and RMSE 2.58°, achieving pose F1-score 0.998.

• Retrospective study of 1000 dental panoramic radiographs (2023–2024), split 80/10/10 for training/validation/testing; annotations by one oral radiologist.
• YOLO11 achieved pose F1-score 0.9975, mAP@50 0.995, and mAP@50:95 0.992; angular MAE 2.37°, RMSE 2.58°. Outperformed YOLOv8 (exact figures not reported).
• Limitation: single-center, no external validation, single annotator.

RadPigeon summaries are original and for information only. They are not clinical advice.