Synthesizing 9,310 Publications Through Natural Language Processing and the Decade-Long Transformation of Artificial Intelligence in Dermatology
Current Research in Health Sciences · 30.06.2026 · 3 (2) · 59-68
DOI: 10.62425/crihs.1924605 ↗Kurz gefasst
Fachartikel in der Zeitschrift Current Research in Health Sciences, online erschienen am 30.06.2026. Verfasst von Orhan Şen. Die Arbeit nennt 23 Literaturangaben. Gefördert von: None.
Abstract (Originalsprache)
Objective
This study aimed to examine the evolution of artificial intelligence (AI) research in dermatology between 2015 and 2025, extending beyond conventional bibliometric indicators through natural language processing (NLP)-assisted content analysis, and to delineate the temporal shifts in disease focus, data types, and technological infrastructure.
Methods
A total of 9,310 publications retrieved from the Web of Science Core Collection were analyzed. Classical bibliometric indicators were assessed using R-Bibliometrix, VOSviewer, and CiteSpace software. In the NLP-based classification phase, which constitutes the original contribution of this study, the title and abstract of each publication were screened using a rule-based algorithm and systematically classified across three dimensions: the targeted dermatological disease, the data type employed, and the clinical task addressed. Thematic shifts between the early period (2015–2019) and the late period (2020–2025) were comparatively examined.
Results
In the early period, oncology (melanoma) dominated the literature with a 32.2% research proportion; however, this rate declined to 20.9% in the late period, during which teledermatology gained momentum. The research focus exhibited a notable clinical shift toward cosmetic/aging (32.4%), surpassing oncology. Concurrently, the clinical tasks targeted by AI expanded beyond diagnosis alone to encompass disease severity scoring (29.7%) and treatment response monitoring (18.0%). Following the release of ChatGPT in 2022, large language model (LLM) research in dermatology demonstrated exponential growth, increasing 25-fold within three years.
Conclusion
The AI literature in dermatology is undergoing a simultaneous three-dimensional transformation. AI is evolving from a specific melanoma screening tool into a “clinical assistant” integrated into routine outpatient practice, encompassing cosmetic and inflammatory diseases. The visionary rise of large language models (LLMs) is accelerating this transformation; however, overcoming dataset asymmetries and algorithmic hallucination risks remains essential for safe clinical integration.
Förderung
- None
Zitieren
APA
Şen, O. (2026). Synthesizing 9,310 Publications Through Natural Language Processing and the Decade-Long Transformation of Artificial Intelligence in Dermatology. Current Research in Health Sciences, 3(2), 59-68. https://doi.org/10.62425/crihs.1924605
Vancouver
Şen O. Synthesizing 9,310 Publications Through Natural Language Processing and the Decade-Long Transformation of Artificial Intelligence in Dermatology. Current Research in Health Sciences. 2026;3(2):59-68. doi:10.62425/crihs.1924605
BibTeX
@article{en2026,
author = {Şen, Orhan},
title = {Synthesizing 9,310 Publications Through Natural Language Processing and the Decade-Long Transformation of Artificial Intelligence in Dermatology},
journal = {Current Research in Health Sciences},
year = {2026},
volume = {3},
number = {2},
pages = {59--68},
doi = {10.62425/crihs.1924605}
}
EUVIMED