EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Effect of Fitzpatrick Skin Type Prompting on Diagnostic Accuracy in Multimodal Large Language Models: A Within-Image Experimental Study

Manoj Bhagwat iD, Tyler Wittles iD, Jeffery Tan iD, Joshua Mijares iD, Neil K Jairath, Syril Keena T Que

Bioengineering · 18.08.2026 · 13 (8) · 932

DOI: 10.3390/bioengineering13080932 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Bioengineering, online erschienen am 18.08.2026. Verfasst von Manoj Bhagwat und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Dermatology, Indiana University School of Medicine, Indianapolis, IN 46202, USA. Die Arbeit nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multimodal large language models are increasingly used for dermatologic queries, but whether Fitzpatrick skin type (FST) labels affect diagnostic accuracy is unknown. We evaluated 656 biopsy-confirmed photographs from the Diverse Dermatology Images (DDI) dataset under no-FST, DDI-concordant FST, and two DDI-discordant FST conditions using ChatGPT 5.2 Edu and Gemini 3.1 Pro browser configurations (5248 evaluations). Confirmed malignant diagnosis omission from the top three differential diagnoses (“lethal miss”) was a prespecified exploratory outcome. Gemini had higher ordinal accuracy than ChatGPT (odds ratio 2.32; 95% confidence interval 2.01–2.67; false discovery rate-adjusted p < 0.001). No FST prompt condition significantly improved accuracy over no-FST prompting; an equal-image-weighted sensitivity analysis yielded similar results. Among malignant images eligible for extreme discordance, extreme DDI-discordant prompting was associated with more ChatGPT lethal misses than no-FST prompting (paired odds ratio 5.0; p = 0.043); Gemini showed no significant paired shift (p = 1.00). Post hoc binary malignancy detection showed low no-FST specificity for ChatGPT (42.5%) and Gemini (22.9%), indicating frequent overcalling. Gemini showed lower no-FST accuracy for malignant FST V–VI images than for FST I–II and III–IV images. FST prompting provided no measurable diagnostic benefit. Neither configuration supports autonomous dermatologic use.

Schlagwörter

Stichwörter der Autor:innen

ChatGPTDiverse Dermatology ImagesFitzpatrick skin typeGPT-5.2Geminiartificial intelligencediagnostic accuracylarge language modelsprompt engineeringskin of color

Autor:innen und Institutionen

  • Department of Dermatology, Indiana University School of Medicine, Indianapolis, IN 46202, USA.
  • Ronald O. Perelman Department of Dermatology, New York University Grossman School of Medicine, New York, NY 10016, USA.

ORCID-Profile: Manoj Bhagwat, Tyler Wittles, Jeffery Tan, Joshua Mijares

Zitieren

APA

Bhagwat, M., Wittles, T., Tan, J., Mijares, J., Jairath, N. K., & Que, S. K. T. (2026). Effect of Fitzpatrick Skin Type Prompting on Diagnostic Accuracy in Multimodal Large Language Models: A Within-Image Experimental Study. Bioengineering, 13(8), 932. https://doi.org/10.3390/bioengineering13080932

Vancouver

Bhagwat M, Wittles T, Tan J, Mijares J, Jairath NK, Que SKT. Effect of Fitzpatrick Skin Type Prompting on Diagnostic Accuracy in Multimodal Large Language Models: A Within-Image Experimental Study. Bioengineering. 2026;13(8):932. doi:10.3390/bioengineering13080932

BibTeX
@article{Bhagwat2026,
  author = {Bhagwat, Manoj and Wittles, Tyler and Tan, Jeffery and Mijares, Joshua and Jairath, Neil K and Que, Syril Keena T},
  title = {Effect of Fitzpatrick Skin Type Prompting on Diagnostic Accuracy in Multimodal Large Language Models: A Within-Image Experimental Study},
  journal = {Bioengineering},
  year = {2026},
  volume = {13},
  number = {8},
  pages = {932},
  doi = {10.3390/bioengineering13080932}
}
Kostenlos starten →