Evaluating the Utility of Generative AI Models for Synthetic Medical Image Generation: A Study on Panoramic Dental X-Rays

Year 2025, Volume: 1 Issue: 2, 190 - 219, 28.07.2025

Elif Yılmaz , Abdullah Huzeyfe Köse , Rahime Yılmaz

https://doi.org/10.26650/d3ai.1726054

Abstract

Integrating artificial intelligence (AI) into dental practice can streamline processes that are otherwise manual and time-consuming. However, implementing AI in dentistry poses challenges, primarily due to the limited availability of clinical data, especially for rare cases, compounded by ethical and regulatory constraints. To address these challenges, synthetic data generation has increasingly been recognised as a promising solution. In this study, we aimed to evaluate the clinical usability of AI-generated panoramic dental X-rays produced by large-scale multimodal generative AIs, specifically GPT o4 mini high and Copilot, using textual prompts. These synthetic images were generated across multiple categories, enabling a comprehensive evaluation of the models’ capabilities. The primary objective was to determine whether AI generated images, produced from structured textual prompts and real sample descriptions, could be used for learning-based model training. For this purpose, an expert dentist evaluated the 61 generated synthetic images from a clinical perspective, focusing on their visual realism, anatomical accuracy, clinical/educa tional utility, image clarity, and absence of artefacts. Both models generated images rated satisfactorily across all quality dimensions, with GPT o4 mini high achieving significantly superior anatomical fidelity and realism than Copilot. It was observed that neither system could consistently reproduce fine anatom ical details with pixel-level accuracy, a limitation that may prevent its use in applications requiring exact replication. Despite these limitations, the high quality and realism of the synthetic images highlight the potential of multimodal generative AIs. Thus, the study’s findings suggest that synthetic panoramic images can reduce the reliance on scarce patient data, thereby accelerating the development of AI applications in dentistry.

Keywords

Synthetic Data Generation , Dental Radiography , Multimodal Generative AI , Clinical Usability , Data Scarcity in Dentistry

References

• R. Baza. 2024. Evaluating artificial intelligence in dental radiography. google scholar
• F. Umer and N. Adnan. 2024. Generative artificial intelligence: synthetic datasets in dentistry. BDJ Open 10, 1, 13. https://doi.org/ 10.1038/s41405-024-00144-9 google scholar
• F. Schwendicke and J. Krois. 2022. Data dentistry: how data are changing clinical care and research. J. Dent. Res. 101, 1 (Jan. 2022), 21–29. https://doi.org/10.1177/00220345211041658 google scholar
• L. T. Reyes, J. K. Knorst, F. R. Ortiz, and T. M. Ardenghi. 2021. Scope and challenges of machine learning-based diagnosis and prognosis in clinical dentistry: A literature review. J. Clin. Transl. Res. 7, 4, 523. google scholar
• T. Wang, Y. Lei, Y. Fu, J. F. Wynne, W. J. Curran, T. Liu, and X. Yang. 2021. A review on medical imaging synthesis using deep learning and its clinical applications. J. Appl. Clin. Med. Phys. 22, 1, 11–36. google scholar
• L. R. Koetzier, J. Wu, D. Mastrodicasa, A. Lutz, M. Chung, W. A. Koszek, et al. 2024. Generating synthetic data for medical imaging. Radiology 312, 3, e232471. google scholar
• A. Hosseini and A. Serag. 2025. Is synthetic data generation effective in maintaining clinical biomarkers? Investigating diffusion models across diverse imaging modalities. Front. Artif. Intell. 7, 1454441. google scholar
• A. B. Abdusalomov, R. Nasimov, N. Nasimova, B. Muminov, and T. K. Whangbo. 2023. Evaluating synthetic medical images using artificial intelligence with the GAN algorithm. Sensors 23, 7, 3440. google scholar
• M. Farhadi Nia, M. Ahmadi, and E. Irankhah. 2025. Transforming dental diagnostics with artificial intelligence: Advanced integration of ChatGPT and large language models for patient care. Front. Dent. Med. 5, 1456208. google scholar
• Z. Yang, Z. Yao, M. Tasmin, P. Vashisht, W. S. Jang, F. Ouyang, et al. 2025. Unveiling GPT-4V’s hidden challenges behind high accuracy on USMLE questions: Observational study. J. Med. Internet Res. 27, e65146. google scholar
• S. Tian, Q. Jin, L. Yeganova, P. T. Lai, Q. Zhu, X. Chen, et al. 2024. Opportunities and challenges for ChatGPT and large language models in biomedicine and health. Brief. Bioinform. 25, 1, bbad493. google scholar
• H. Fukuda, M. Morishita, K. Muraoka, S. Yamaguchi, T. Nakamura, I. Yoshioka, et al. 2025. Evaluating the image recognition capabilities of GPT-4V and Gemini Pro in the Japanese national dental examination. J. Dent. Sci. 20, 1, 368–372. google scholar
• N. Sharmin and A. K. Chow. From text to tissue: Artificial intelligence-generation of biological images. J. Dent. Educ. (in press). google scholar
• M. Silhadi, W. B. Nassrallah, D. Mikhail, D. Milad, and M. Harissi-Dagher. 2025. Assessing the performance of Microsoft Copilot, GPT-4 and Google Gemini in ophthalmology. Can. J. Ophthalmol. (in press). google scholar
• D. Stephan, A. Bertsch, M. Burwinkel, S. Vinayahalingam, B. Al-Nawas, P. W. Kämmerer, and D. G. Thiem. 2024. AI in dental radiology —improving the efficiency of reporting with ChatGPT: Comparative study. J. Med. Internet Res. 26, e60684. google scholar
• A. Mittal, A. K. Moorthy, and A. C. Bovik. 2012. No-reference image quality assessment in the spatial domain. IEEE Trans. Image Process. 21, 12, 4695–4708. google scholar
• Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli. 2004. Image quality assessment: From error visibility to structural similarity. IEEE Trans. Image Process. 13, 4, 600–612. google scholar
• R. Zhang, P. Isola, A. A. Efros, E. Shechtman, and O. Wang. 2018. The unreasonable effectiveness of deep features as a perceptual metric. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit., 586–595. google scholar
• M. Büttner, R. Rokhshad, J. Brinz, J. Issa, A. Chaurasia, S. E. Uribe, et al. 2024. Core outcomes measures in dental computer vision studies (DentalCOMS). J. Dent. 150, 105318. google scholar
• H. M. S. S. Herath, H. M. K. K. M. B. Herath, N. Madusanka, and B. I. Lee. 2025. A systematic review of medical image quality assessment. J. Imaging 11, 4, 100. google scholar
• S. Pedersen, S. Jain, M. Chavez, V. Ladehoff, B. N. de Freitas, and R. Pauwels. 2025. Pano-GAN: A deep generative model for panoramic dental radiographs. J. Imaging 11, 2, 41. google scholar
• F. Khader, G. Müller-Franzes, S. Tayebi Arasteh, T. Han, C. Haarburger, M. Schulze-Hagen, et al. 2023. Denoising diffusion probabilistic models for 3D medical image generation. Sci. Rep. 13, 1, 7303. google scholar
• G. Müller-Franzes, J. M. Niehues, F. Khader, S. T. Arasteh, C. Haarburger, C. Kuhl, et al. 2023. A multimodal comparison of latent denoising diffusion probabilistic models and generative adversarial networks for medical image synthesis. Sci. Rep. 13, 1, 12098. google scholar
• B. De Wilde, A. Saha, M. de Rooij, H. Huisman, and G. Litjens. 2023. Medical diffusion on a budget: Textual inversion for medical image generation. arXiv preprint arXiv:2303.13430. https://arxiv.org/abs/2303.13430 google scholar
• K. Kokomoto, R. Okawa, K. Nakano, and K. Nozaki. 2021. Intraoral image generation by progressive growing of generative adversarial network and evaluation of generated image quality by dentists. Sci. Rep. 11, 1, 18517. google scholar
• S. Y. Jeong, E. J. Bae, H. S. Jang, S. Na, and S. Y. Ihm. 2024. Study on virtual tooth image generation utilizing CF-fill and Pix2pix for data augmentation. Sci. Rep. 14, 1 (2024), 26772. google scholar
• S. Pedersen, S. Jain, M. Chavez, V. Ladehoff, B. N. de Freitas, and R. Pauwels. 2025. Pano-GAN: A deep generative model for panoramic dental radiographs. J. Imaging 11, 2 (2025), 41. google scholar
• M. Chaichuk, S. Gautam, S. Hicks, and E. Tutubalina. 2025. Prompt to polyp: Clinically-aware medical image synthesis with diffusion models. arXiv preprint arXiv:2505.05573. https://arxiv.org/abs/2505.05573 google scholar
• J. Rousseau, C. Alaka, E. Covili, H. Mayard, L. Misrachi, and W. Au. 2023. Pre-training with diffusion models for dental radiography segmentation. In Proc. Int. Conf. Med. Image Comput. Comput.-Assisted Interv., Cham, Switzerland, 174–182. google scholar
• R. Di Via, F. Odone, and V. P. Pastore. 2025. Self-supervised pre-training with diffusion model for few-shot landmark detection in X-ray images. In 2025 IEEE/CVF Winter Conf. on Applications of Computer Vision (WACV), Waikoloa, HI, USA, 3886–3896. google scholar
• I. E. Hamamci, S. Er, E. Simsar, A. Sekuboyina, M. Gundogar, B. Stadlinger, et al. 2023. Diffusion-based hierarchical multi-label object detection to analyze panoramic dental X-rays. In Proc. Int. Conf. Med. Image Comput. Comput.-Assisted Interv., Cham, Switzerland, 389–399. google scholar
• F. Khader, G. Müller-Franzes, S. Tayebi Arasteh, T. Han, C. Haarburger, M. Schulze-Hagen, et al. 2023. Denoising diffusion probabilistic models for 3D medical image generation. Sci. Rep. 13, 1 (2023), 7303. google scholar
• A. L. Y. Hung, K. Zhao, H. Zheng, R. Yan, S. S. Raman, D. Terzopoulos, and K. Sung. 2023. Med-cDiff: Conditional medical image generation with diffusion models. Bioengineering 10, 11 (2023), 1258. google scholar
• Y. Shi, A. Abulizi, H. Wang, K. Feng, N. Abudukelimu, Y. Su, and H. Abudukelimu. 2024. Diffusion models for medical image computing: A survey. Tsinghua Sci. Technol. 30, 1 (2024), 357–383. google scholar
• F. Umer and N. Adnan. 2024. Generative artificial intelligence: Synthetic datasets in dentistry. BDJ Open 10 (2024), 13. https://doi. org/10.1038/s41405-024-00144-9 google scholar
• S. Pedersen, S. Jain, M. Chavez, V. Ladehoff, B. N. de Freitas, and R. Pauwels. 2025. Pano-GAN: A deep generative model for panoramic dental radiographs. J. Imaging 11, 2 (2025), 41. google scholar
• C. Gao, B. D. Killeen, Y. Hu, R. B. Grupp, R. H. Taylor, M. Armand, and M. Unberath. 2023. Synthetic data accelerates the development of generalizable learning-based algorithms for X-ray image analysis. Nat. Mach. Intell. 5, 3 (2023), 294–308. google scholar
• S. A. Koohpayegani. 2024. Improving efficiency of deep learning models. Ph.D. Dissertation. University of California, Davis, CA, USA. google scholar
• D. B. Dangi, B. B. Dangi, and O. Steinbock. 2025. Evaluation of GPT-4o and GPT-4o-Mini’s vision capabilities for compositional analysis from dried solution drops. ACS Omega 10, 18 (2025), 18955–18959. google scholar
• M. Sourek. 2025. Artificial Intelligence in Architecture and the Built Environment: The Revolution Yet to Come. CRC Press, Boca Raton, FL, USA. google scholar
• P. Rouzrokh, B. Khosravi, S. Faghani, M. Moassefi, M. M. Shariatnia, P. Rouzrokh, and B. Erickson. 2025. A current review of generative AI in medicine: Core concepts, applications, and current limitations. Curr. Rev. Musculoskelet. Med. (2025), 1–21. google scholar
• Y. Zhang, F. Ye, L. Chen, F. Xu, X. Chen, H. Wu, et al. 2023. Children’s dental panoramic radiographs dataset for caries segmentation and dental disease detection. Sci. Data 10, 1 (2023), 380. google scholar
• A. Mittal, A. K. Moorthy, and A. C. Bovik. 2011. Blind/referenceless image spatial quality evaluator. In Proc. 45th Asilomar Conf. on Signals, Systems and Computers (ASILOMAR), Pacific Grove, CA, USA, 723–727. https://doi.org/10.1109/ACSSC.2011.6190099 google scholar
• Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli. 2004. Image quality assessment: From error visibility to structural similarity. IEEE Trans. Image Process. 13, 4 (Apr. 2004), 600–612. google scholar
• R. Zhang, P. Isola, A. A. Efros, E. Shechtman, and O. Wang. 2018. The unreasonable effectiveness of deep features as a perceptual metric. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 586–595. google scholar
• Z. Hanusz, J. Tarasinska, and W. Zielinski. 2016. Shapiro–Wilk test with known mean. REVSTAT Stat. J. 14, 1 (2016), 89–100. https:// doi.org/10.57805/revstat.v14i1.180 google scholar
• T. W. MacFarland and J. M. Yates. 2016. Kruskal–Wallis H-test for oneway analysis of variance (ANOVA) by ranks. In Introduction to Nonparametric Statistics for the Biological Sciences Using R. Springer, Cham, Switzerland, 177–211. google scholar
• D. G. Pereira, A. Afonso, and F. M. Medeiros. 2015. Overview of Friedman’s test and post-hoc analysis. Commun. Stat. Simul. Comput. 44, 10 (2015), 2636–2653. google scholar
• R. F. Woolson. 2005. Wilcoxon signed-rank test. In Encyclopedia of Biostatistics, P. Armitage and T. Colton (Eds.). John Wiley & Sons, Hoboken, NJ, USA. google scholar
• J. Hauke and T. Kossowski. 2011. Comparison of values of Pearson’s and Spearman’s correlation coefficients on the same sets of data. Quaest. Geogr. 30, 2 (2011), 87–93. https://doi.org/10.2478/v10117-011-0021-1 google scholar