Diagnostic Performance of ChatGPT‐4o and DeepSeek‐3 Differential Diagnosis of Complex Oral Lesions: A Multimodal Imaging and Case Difficulty Analysis
Oral Diseases · 2025-12 · 31 (12) · 3361-3371
DOI: 10.1111/odi.70007 ↗Kurz gefasst
Fachartikel in der Zeitschrift Oral Diseases, online erschienen am 01.07.2025. Verfasst von Fatma E A Hassanein und 6 weiteren Autor:innen. Beteiligte Einrichtung: Oral Medicine, Periodontology, and Oral Diagnosis, Faculty of Dentistry, King Salman International University, El Tur, Egypt. Hauptthemen laut MEDLINE: Mouth Diseases, Multimodal Imaging. Die Arbeit wurde bisher 30-mal zitiert (Crossref) und nennt 34 Literaturangaben.
Abstract (Originalsprache)
Background
AI models like ChatGPT‐4o and DeepSeek‐3 show diagnostic promise, but their reliability in complex, image‐based oral lesions remains unclear. This study aimed to evaluate and compare the diagnostic accuracy of ChatGPT‐4o and DeepSeek‐3 despite their differing modalities against oral medicine (OM) experts across varied lesion types and case difficulty levels.
Methods
Eighty standardized clinical vignettes derived from real‐world oral disease cases, including clinical images/radiographs, were evaluated. Differential diagnoses were generated by ChatGPT‐4o, DeepSeek‐3, and four board‐certified OM specialists, with accuracy assessed at Top‐1, Top‐3, and Top‐5 levels.
Results
OM specialists consistently achieved the highest diagnostic accuracy. However, DeepSeek‐3 significantly outperformed ChatGPT‐4o at the Top‐3 level ( p = 0.0153) and showed greater robustness in high‐difficulty and inflammatory cases despite its text‐only modality. Multimodal imaging enhanced diagnostic accuracy. Regression analysis indicated lesion type and imaging modality as positive predictors, while diagnostic difficulty negatively impacted Top‐1 performance.
Conclusions
Remarkably, the text‐only DeepSeek‐3 model exceeded the diagnostic performance of the multimodal ChatGPT‐4o model for complex oral lesions, highlighting its structured reasoning capabilities and reduced hallucination rate. These findings underscore the potential of non‐vision LLMs in diagnostic support, emphasizing the critical need for expert oversight in complex scenarios.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Oral Medicine, Periodontology, and Oral Diagnosis, Faculty of Dentistry, King Salman International University, El Tur, Egypt.
- Oral Medicine and Periodontology, Faculty of Dentistry, Cairo University, Giza, Egypt.
- Oral Medicine and Periodontology, Faculty of Dentistry, Galala University, Suez, Egypt.
- Oral Medicine and Periodontology, Ain Shams University, Cairo, Egypt.
- Prosthodontics Dentistry, Faculty of Dentistry, King Salman International University, El Tur, Egypt.
ORCID-Profile: Fatma E A Hassanein
Zitieren
APA
Hassanein, F. E. A., El Barbary, A., Hussein, R. R., Ahmed, Y., El-Guindy, J., Sarhan, S., & Abou-Bakr, A. (2025). Diagnostic Performance of ChatGPT‐4o and DeepSeek‐3 Differential Diagnosis of Complex Oral Lesions: A Multimodal Imaging and Case Difficulty Analysis. Oral Diseases, 31(12), 3361-3371. https://doi.org/10.1111/odi.70007
Vancouver
Hassanein FEA, El Barbary A, Hussein RR, Ahmed Y, El-Guindy J, Sarhan S, et al. Diagnostic Performance of ChatGPT‐4o and DeepSeek‐3 Differential Diagnosis of Complex Oral Lesions: A Multimodal Imaging and Case Difficulty Analysis. Oral Diseases. 2025;31(12):3361-3371. doi:10.1111/odi.70007
BibTeX
@article{Hassanein2025,
author = {Hassanein, Fatma E A and El Barbary, Ahmed and Hussein, Radwa R and Ahmed, Yousra and El-Guindy, Jylan and Sarhan, Susan and Abou-Bakr, Asmaa},
title = {Diagnostic Performance of ChatGPT‐4o and DeepSeek‐3 Differential Diagnosis of Complex Oral Lesions: A Multimodal Imaging and Case Difficulty Analysis},
journal = {Oral Diseases},
year = {2025},
volume = {31},
number = {12},
pages = {3361--3371},
doi = {10.1111/odi.70007}
}
EUVIMED