Related Experiment Video
Updated: Jul 16, 2026
07:32
Measuring Maxillary Posterior Tooth Movement: A Model Assessment using Palatal and Dental Superimposition
Published on: February 23, 2024
Comparative Evaluation of Large Language Models for Reporting Jaw Lesions on Panoramic Radiographs.
Duygu Çelik Özen1, Okan Özen2, Utku Tuğberk Göktürk1
1Department of Oral and Maxillofacial Radiology, Faculty of Dentistry, Inonu University, 44280 Malatya, Turkey.
Diagnostics (Basel, Switzerland)
|July 15, 2026
Summary
This study evaluated artificial intelligence chatbots for diagnosing jaw lesions on panoramic radiographs. Gemini and ChatGPT showed potential, while Copilot performed less effectively, indicating AI can be a supportive tool with further validation.
Area of Science:
- Dentistry
- Artificial Intelligence
- Radiology
Background:
- Panoramic radiography is crucial for diagnosing jaw lesions.
- Large language model (LLM)-based AI chatbots are emerging tools in medical diagnostics.
- Assessing AI diagnostic capabilities in radiology is essential for potential clinical integration.
Purpose of the Study:
- To evaluate the diagnostic performance of ChatGPT 4.0, Gemini 2.5, and Microsoft Copilot.
- To compare AI chatbot accuracy in identifying jaw lesions with varying radiographic densities (radiolucent, radiopaque, mixed).
- To determine the potential utility of LLM-based AI as supportive tools in radiographic interpretation.
Main Methods:
- 120 panoramic radiographs of jaw lesions were analyzed by three AI chatbots.
- A standardized scoring framework assessed lesion characteristics and overall diagnostic scores.
- Statistical analysis (Kruskal-Wallis test) compared LLM performance.
Main Results:
- Significant differences in diagnostic performance were found among the AI chatbots (p < 0.05).
- Gemini excelled in radiolucent and mixed lesions; ChatGPT performed better in radiopaque lesions.
- Microsoft Copilot showed the lowest overall performance across all lesion types.
Conclusions:
- LLM-based AI chatbots exhibit variable diagnostic capabilities for jaw lesions on panoramic images.
- Gemini and ChatGPT show promise as potential supportive tools in radiographic interpretation.
- Further validation studies are necessary before widespread clinical adoption of AI in this field.
