Abstract: TH-PO0741
Nano-High Performance Liquid Chromatography-Mass Spectrometry (NanoHPLC-MS)-Based Urinary Proteomic Signatures Predict Remission and Relapse in Primary Membranous Nephropathy
Session Information
- Glomerular Innovations: Artificial Intelligence, Multiomics, and Biomarkers
November 06, 2025 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Glomerular Diseases
- 1402 Glomerular Diseases: Clinical, Outcomes, and Therapeutics
Authors
- Zeng, Lingyun, The Third Affiliated Hospital of Sun Yet-sun University Department of Nephrology, Guangzhou, Guangdong, China
- Peng, Hui, The Third Affiliated Hospital of Sun Yet-sun University Department of Nephrology, Guangzhou, Guangdong, China
Background
Primary membranous nephropathy (PMN) is a leading cause of nephrotic syndrome in adults, yet reliable prognostic tools remain limited. Although anti-PLA2R antibody testing is widely used, it does not fully capture disease heterogeneity or predict treatment outcomes. This study aimed to identify novel urinary proteomic biomarkers for predicting clinical remission and to uncover relapse-associated molecular subtypes in PMN.
Methods
We collected baseline midstream urine samples from 86 biopsy-confirmed PMN patients and performed quantitative proteomic profiling using NanoHPLC-MS. The primary endpoints were remission status and time to remission. We constructed 101 machine learning model combinations to identify optimal algorithms. A prognostic risk score model was developed using random survival forest and stepwise Cox regression, with internal validation via bootstrap resampling. Molecular subtypes were identified through unsupervised clustering and non-negative matrix factorization based on differentially expressed proteins.
Results
During a median follow-up of 8 months (IQR 3.4-18.5), 76.1% achieved remission. The four-protein risk model (PON1, ACTBL2, RDX, TPP1) was established (C-index=0.729), effectively stratifying patients into high and low risk groups. The combined model integrating clinical features (serum anti-PLA2R antibody, age, eGFR) with the four-protein signature further improved predictive performance (C-index=0.744), outperforming the clinical-only model (C-index=0.636). Among remission cases, molecular profiling revealed three distinct subtypes, PMN2 was independently associated with a significantly higher risk of relapse (OR=10.26, P<0.05).
Conclusion
This study established a non-invasive, NanoHPLC-MS-based urinary proteomic approach for predicting clinical remission and relapse risk in PMN. The integrated model and molecular subtypes offer a novel framework for precision prognostication and individualized management.