Self-supervised artificial intelligence system for differentiating mycosis fungoides and benign inflammatory dermatoses
British Journal of Dermatology · 23.06.2026 · 195 (1) · 136-145
DOI: 10.1093/bjd/ljag129 ↗Kurz gefasst
Fachartikel in der Zeitschrift British Journal of Dermatology, online erschienen am 09.04.2026. Verfasst von Jie Zhao und 8 weiteren Autor:innen. Beteiligte Einrichtung: Department of Dermatology, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 30 Literaturangaben. Gefördert von: National Key Research and Development Program of China.
Abstract (Originalsprache)
Background
Mycosis fungoides (MF), the most common cutaneous T-cell lymphoma, is challenging to diagnose in its early stages because the clinical and histopathological features often overlap with benign inflammatory dermatoses (BIDs), leading to misclassification and delayed treatment.
Objectives
To evaluate whether a self-supervised artificial intelligence (AI) system can support diagnostic decision-making in MF and common inflammatory mimics.
Methods
This retrospective two-centre study included patients with confirmed diagnoses of MF or BIDs. A self-supervised multimodal system integrating whole-slide histopathology and routine clinical variables was developed to perform multiclass differential classification across MF and common BIDs. Cases were partitioned into training, internal validation and independent external validation cohorts at the patient level. Clinical utility was assessed in a reader study in which dermatopathologists reviewed cases with and without system assistance.
Results
Across 786 whole-slide images from 532 patients, the multimodal model achieved macro-area under the curves > 0.85 in both validation sets, with a macro-balanced accuracy of 0.837 [95% confidence interval (CI) 0.778–0.897; n = 106] internally and 0.762 (95% CI 0.724–0.802; n = 260) externally, representing improved performance over the unimodal histopathology model. In the reader study, AI assistance was associated with improved macro-balanced accuracy among junior (from 0.778 to 0.818) and senior (from 0.805 to 0.859) dermatopathologists, accompanied by consistent improvements in macro-averaged sensitivity and specificity across all diagnostic categories. Interpretability analyses generated heatmaps that were generally consistent with recognized histopathological features of MF and BIDs, aligning with dermatopathologist interpretation.
Conclusions
This multimodal, self-supervised system may support dermatopathologists by providing interpretable, probability-based guidance for the classification of MF and its common inflammatory mimics.
Autor:innen und Institutionen
- Department of Dermatology, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China.
- Dermatology Department, Dermatology Hospital, Southern Medical University, Guangzhou, China.
ORCID-Profile: Juan Bai, Jianjun Qiao
Förderung
- National Key Research and Development Program of China
Zitieren
APA
Zhao, J., Bai, J., Li, G., Sun, Q., Li, S., Wang, C., Li, H., Xue, R., & Qiao, J. (2026). Self-supervised artificial intelligence system for differentiating mycosis fungoides and benign inflammatory dermatoses. British Journal of Dermatology, 195(1), 136-145. https://doi.org/10.1093/bjd/ljag129
Vancouver
Zhao J, Bai J, Li G, Sun Q, Li S, Wang C, et al. Self-supervised artificial intelligence system for differentiating mycosis fungoides and benign inflammatory dermatoses. British Journal of Dermatology. 2026;195(1):136-145. doi:10.1093/bjd/ljag129
BibTeX
@article{Zhao2026,
author = {Zhao, Jie and Bai, Juan and Li, Guomin and Sun, Qingmiao and Li, Sheng and Wang, Chen and Li, Hui and Xue, Ruzeng and Qiao, Jianjun},
title = {Self-supervised artificial intelligence system for differentiating mycosis fungoides and benign inflammatory dermatoses},
journal = {British Journal of Dermatology},
year = {2026},
volume = {195},
number = {1},
pages = {136--145},
doi = {10.1093/bjd/ljag129}
}
EUVIMED