A Scoping Review of Machine Learning Applications for Diagnosis, Classification, and Prognosis in Lupus Nephritis
Kidney360 · 18.03.2026
DOI: 10.34067/kid.0000001196 ↗Kurz gefasst
Scoping Review in der Zeitschrift Kidney360, online erschienen am 18.03.2026. Verfasst von Harishwar Reddy Kasireddy und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Electrical and Computer Engineering, University of Florida, Gainesville, FL, USA. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: NIAMS NIH HHS.
Abstract (Originalsprache)
Background
Over the past decade, machine learning (ML) has transformed histopathology, yet its application in lupus nephritis (LN) remains underexplored. Earlier reviews have centered on pathologist concordance in LN classification, biomarkers, and AI-enabled diagnosis and treatment-response prediction. Here, we conducted the first comprehensive scoping review of ML in LN, spanning across diagnosis, risk of renal flare, remission, treatment response prediction, and LN disease classification tasks.
Methods
PubMed and Scopus databases were searched for peer-reviewed studies published between January, 2000 and April, 2025 that applied ML to histopathological or clinical data in LN. Inclusion criteria encompassed any supervised or unsupervised ML approach addressing diagnosis, renal flare, treatment response, remission prediction, or LN disease classification. Two reviewers independently screened titles, abstracts, and full texts.
Results
Of 20 qualifying studies, 12 met the inclusion criteria. Additionally, through manual citation and reference screening we added eight more. A large set of studies ( n = 11) are focused on tackling mainly histologic subtyping diagnosis and LN classification. Following this, two studies focused on developing renal flare prediction models leveraging traditional ML models and deep learning (DL) models. Moreover, three studies focused on developing remission prediction models ( n = 3) using both clinical and histological data. Finally, four studies addressed treatment response prediction employing ML and DL using laboratory results, urinary biomarkers, and histology data.
Conclusions
ML has shown promise across multiple LN related tasks, from diagnosis to prognosis. To accelerate clinical translation, future efforts should prioritize larger multicenter, multimodal datasets, standardized annotation protocols, rigorous external validation, integration of explainable AI techniques, usage of whole slide image data, multi-omics data, development of webtools, use of advanced agentic AI, and digital twins to develop personalized actionable items in patients.
Autor:innen und Institutionen
- Department of Electrical and Computer Engineering, University of Florida, Gainesville, FL, USA.
- Seed Production Innovation, Crop Science Division, Bayer Company, St. Louis, MO.
- Department of Pathology, University of Michigan, Ann Arbor, MI, USA.
- Division of Medicine - Quantitative Health, University of Florida, Gainesville, FL, USA.
ORCID-Profile: Harishwar Reddy Kasireddy, Jeffrey B Hodgin, Pinaki Sarder
Förderung
- NIAMS NIH HHS
Zitieren
APA
Kasireddy, H. R., La Rosa, P. S., Demeke, D., Hodgin, J. B., & Sarder, P. (2026). A Scoping Review of Machine Learning Applications for Diagnosis, Classification, and Prognosis in Lupus Nephritis. Kidney360. https://doi.org/10.34067/kid.0000001196
Vancouver
Kasireddy HR, La Rosa PS, Demeke D, Hodgin JB, Sarder P. A Scoping Review of Machine Learning Applications for Diagnosis, Classification, and Prognosis in Lupus Nephritis. Kidney360. 2026. doi:10.34067/kid.0000001196
BibTeX
@article{Kasireddy2026,
author = {Kasireddy, Harishwar Reddy and La Rosa, Patricio S and Demeke, Dawit and Hodgin, Jeffrey B and Sarder, Pinaki},
title = {A Scoping Review of Machine Learning Applications for Diagnosis, Classification, and Prognosis in Lupus Nephritis},
journal = {Kidney360},
year = {2026},
doi = {10.34067/kid.0000001196}
}
EUVIMED