Evaluation of Dental Panoramic Radiographs by Artificial Intelligence Compared to Human Reference: A Diagnostic...

Area of Science:

• Dentistry
• Artificial Intelligence
• Radiology

Background:

• Artificial intelligence (AI) is increasingly vital in dentistry for diagnosis and treatment planning.
• Dental panoramic radiographs (DPRs) are commonly used diagnostic tools.

Purpose of the Study:

• To evaluate the performance of AI software in analyzing DPRs.
• To assess sensitivity, specificity, accuracy, and precision of AI in identifying various dental features.

Main Methods:

• Analysis of 600 dental panoramic radiographs (DPRs).
• Comparison of AI software analysis with human researcher analysis.
• Assessment of eight variables: missing tooth, root canal filling, endodontic lesion, implant, abutment, pontic, crown, and sound tooth.

Main Results:

• AI demonstrated high performance (>90%) for detecting missing teeth, root canal fillings, and implant abutment crowns.
• Moderate sensitivity and precision were observed for identifying endodontic lesions.
• Lowest precision (65.30%) was recorded for detecting crowns.

Conclusions:

• AI software is a potentially valuable tool for dental diagnosis and treatment planning.
• Clinical verification by dentists is recommended, particularly for endodontic lesions and crowns.
• Further research is needed due to study limitations.