Automated detection of zygomatic fractures on spiral computed tomography using a deep learning model

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Radiology

Background:

• Zygomatic fractures are common facial bone injuries.
• Accurate detection is crucial for effective treatment.
• Manual interpretation of CT scans can be time-consuming.

Purpose of the Study:

• To evaluate the performance of the YOLOv8 deep learning model for detecting zygomatic fractures.
• To assess the model's accuracy across different fracture categories.

Main Methods:

• A dataset of 13,988 axial and 14,107 coronal CT slices with zygomatic fractures was used.
• Fracture lines were annotated across seven categories.
• The YOLOv8 model was trained and validated using a 6:2:2 data split.

Main Results:

• The YOLOv8 model achieved high accuracy (94.2-97.9%) in detecting zygomatic fractures.
• Recall exceeded 90% for all fracture categories.
• The model demonstrated the highest performance for zygomaticomaxillary suture and zygomatic arch fractures.

Conclusions:

• The YOLOv8 deep learning model shows significant promise for automated zygomatic fracture detection.
• This technology could enhance diagnostic efficiency and accuracy in radiology.