EUVIMEDOpdag EUVIMED
QuerschnittstudieOpen access

Multimodal Large Language Models for Dental Chart Image Interpretation: Cross-Sectional Benchmarking Study With Students and Clinicians

Ah-Young Cho iD, Soo-Heang Eo iD, Mi-Jeong Jeon iD, Jae-Hoon Kim iD, Jungjoon Ihm iD, Deog-Gyu Seo iD

JMIR Medical Education · 2026-09-04 · 12 · e91809-e91809

DOI: 10.2196/91809 ↗

At a glance

Querschnittstudie in JMIR Medical Education, published online on 2026-09-04. Written by Ah-Young Cho and 5 co-authors. Institution: Department of Conservative Dentistry and Dental Research Institute, School of Dentistry, Seoul National University, 101 Daehakno, Jongno-Gu, Seoul, 0…. Major MEDLINE topics: Large Language Models, Benchmarking, Students, Dental, Dental Records. Cited 2 times so far (Crossref); 60 references. The full text is freely available in PubMed Central.

AbstractSubjectsBibliografiske oplysningerCiteRelated studies

Abstract (original language)

Background

Entering clinical training, dental students must learn to read tooth-centered electronic dental records, but limited teaching time in patient-centered clinics can leave gaps in chart-reading literacy. Multimodal large language models (LLMs) that process dental record images may offer scalable educational support, yet their performance has not been benchmarked.

Objective

This study evaluated multimodal ChatGPT models on dental chart image interpretation and compared the best-performing model with dental students and residents.

Methods

A retrospective, cross-sectional benchmark study used deidentified dental chart images from 15 patients at Seoul National University Dental Hospital (2017‐2025). Charts containing Korean and English text were captured as sequential screenshots (154 images). For each patient, 24 Korean-language questions (360 total) spanned four categories: Type A, general factual retrieval; Type B, tooth- or procedure-specific retrieval; Type C, interpretation requiring multientry synthesis; and Type D, absent-information questions assessing abstention. Nine multimodal ChatGPT models (available August 2, 2025‐August 9, 2025) were evaluated under standardized conditions. Outputs were scored against a gold standard using 7 metrics, with sentence bidirectional encoder representations from transformers (SBERT) similarity prespecified as the primary semantic measure. Human baselines included 2 third-year students and 2 first-year residents. Groups were compared with Kruskal-Wallis tests and Dunn post hoc analyses. Gold-standard reliability was assessed by independent senior-expert review and chance-corrected agreement (Gwet AC1) among clinical reference raters, and the model ranking was confirmed by content-based clinical accuracy analysis.

Results

Across 360 items, GPT-5 Thinking achieved the highest median SBERT similarity (0.900, IQR 0.525‐1.000), followed by GPT-5 Pro (0.861, IQR 0.501‐1.000) and OpenAI o3 (0.831, IQR 0.489‐1.000), with significant overall group differences ( P

Subjects

MeSH terms (MEDLINE)

Large Language ModelsBenchmarkingStudents, DentalDental RecordsHumansCross-Sectional StudiesRetrospective StudiesMaleFemale

Author keywords

ChatGPTbenchmarkingclinical competencedental recordseducation, dentalgenerative AIlarge language modelsmultimodal ChatGPTnatural language processingstudents, dental

Authors and institutions

  • Department of Conservative Dentistry and Dental Research Institute, School of Dentistry, Seoul National University, 101 Daehakno, Jongno-Gu, Seoul, 03080, Republic of Korea, 82 2-6256-3182.
  • Graduate School, Department of Urban Big Data Convergence, University of Seoul, Seoul, Republic of Korea.
  • AI Agent Team, CryptoLab Inc., Seoul, Republic of Korea.
  • Department of Conservative Dentistry and Oral Science Research Center, Yonsei University College of Dentistry, Seoul, Republic of Korea.
  • Department of Dental Education, Dental and Life Science Institute, School of Dentistry, Pusan National University and Dental Research Institute, Busan, Republic of Korea.
  • Department of Dental Education and Dental Research Institute, School of Dentistry, Seoul National University, Seoul, Republic of Korea.

ORCID profiles: Ah-Young Cho, Soo-Heang Eo, Mi-Jeong Jeon, Jae-Hoon Kim, Jungjoon Ihm, Deog-Gyu Seo

Cite

APA

Cho, A. Y., Eo, S. H., Jeon, M. J., Kim, J. H., Ihm, J., & Seo, D. G. (2026). Multimodal Large Language Models for Dental Chart Image Interpretation: Cross-Sectional Benchmarking Study With Students and Clinicians. JMIR Medical Education, 12, e91809-e91809. https://doi.org/10.2196/91809

Vancouver

Cho AY, Eo SH, Jeon MJ, Kim JH, Ihm J, Seo DG. Multimodal Large Language Models for Dental Chart Image Interpretation: Cross-Sectional Benchmarking Study With Students and Clinicians. JMIR Medical Education. 2026;12:e91809-e91809. doi:10.2196/91809

BibTeX
@article{Cho2026,
  author = {Cho, Ah-Young and Eo, Soo-Heang and Jeon, Mi-Jeong and Kim, Jae-Hoon and Ihm, Jungjoon and Seo, Deog-Gyu},
  title = {Multimodal Large Language Models for Dental Chart Image Interpretation: Cross-Sectional Benchmarking Study With Students and Clinicians},
  journal = {JMIR Medical Education},
  year = {2026},
  volume = {12},
  pages = {e91809--e91809},
  doi = {10.2196/91809}
}
Start for free →