Artificial Intelligence-Driven Nephrology: The Role of Large Language Models in Kidney Care
American Journal of Nephrology · 28.08.2025 · 57 (4) · 476-490
DOI: 10.1159/000548208 ↗Kurz gefasst
Übersichtsarbeit in der Zeitschrift American Journal of Nephrology, online erschienen am 28.08.2025. Verfasst von Carmine Zoccali und 6 weiteren Autor:innen. Beteiligte Einrichtung: Renal Research Institute, New York, New York, USA, carmine.zoccali@icloud.com. Hauptthemen laut MEDLINE: Artificial Intelligence, Renal Insufficiency, Chronic. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 39 Literaturangaben.
Abstract (Originalsprache)
Background: Artificial intelligence (AI) increasingly impacts medicine and medical specialties, including nephrology. Technologies such as large language models (LLMs), decision-support AI, and machine learning-powered predictive analytics enhance clinical care. These AI-driven tools show great potential in areas such as predicting the risk of chronic kidney disease (CKD), managing dialysis, supporting kidney transplantation, and treating CKD and diabetes-related kidney issues. Summary: General AI platforms like ChatGPT, Bard, and Google Gemini are useful for education and synthesizing knowledge. In contrast, specialized medical AI systems such as KidneyIntelX and DreaMed Advisor provide clinically validated decision support systems that aid physicians in patient care. Retrieval-augmented generation (RAG) enhances LLMs by accessing real-time medical data and research insights, reducing misinformation risks, and ensuring accurate, verified medical responses. However, LLMs still face challenges in adapting to complex patient cases. The effectiveness of RAG depends on the quality of the data retrieved and adherence to ethical and confidentiality standards, with human oversight often necessary. Key Messages: (i) Improving AI accuracy, increasing model transparency, and ensuring seamless integration into clinical settings maximize AI benefits in nephrology. (ii) Regulatory approvals and validation are essential to build trust among patients, physicians, and healthcare institutions. (iii) When integrated correctly into clinical workflows, AI can transform nephrology practice by providing efficient, data-driven insights, improving patient outcomes, and reducing administrative burdens. (iv) Ethical, responsible adoption with stringent oversight is crucial for successfully implementing AI in nephrology.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Renal Research Institute, New York, New York, USA, carmine.zoccali@icloud.com.
- Institute of Molecular Biology and Genetics (Biogem), Ariano Irpino, Italy, carmine.zoccali@icloud.com.
- Associazione Ipertensione Nefrologia Trapianto Renale (IPNET), Reggio Calabria, Italy, carmine.zoccali@icloud.com.
- Department of Nephrology, Royal Preston Hospital, Fulwood, UK.
- Department of Surgery Transplantation and Gastroenterology, Semmelweis University, Budapest, Hungary.
- Division of Nephrology, Department of Internal Medicine, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands.
Zitieren
APA
Zoccali, C., Floyd, L., Cseprekal, O., Eisenga, M. F., Mirioglu, S., Caravaca-Fontàn, F., & Mallamaci, F. (2025). Artificial Intelligence-Driven Nephrology: The Role of Large Language Models in Kidney Care. American Journal of Nephrology, 57(4), 476-490. https://doi.org/10.1159/000548208
Vancouver
Zoccali C, Floyd L, Cseprekal O, Eisenga MF, Mirioglu S, Caravaca-Fontàn F, et al. Artificial Intelligence-Driven Nephrology: The Role of Large Language Models in Kidney Care. American Journal of Nephrology. 2025;57(4):476-490. doi:10.1159/000548208
BibTeX
@article{Zoccali2025,
author = {Zoccali, Carmine and Floyd, Lauren and Cseprekal, Orsolya and Eisenga, Michele F and Mirioglu, Safak and Caravaca-Fontàn, Fernando and Mallamaci, Francesca},
title = {Artificial Intelligence-Driven Nephrology: The Role of Large Language Models in Kidney Care},
journal = {American Journal of Nephrology},
year = {2025},
volume = {57},
number = {4},
pages = {476--490},
doi = {10.1159/000548208}
}
EUVIMED