Related Experiment Video
Updated: Jan 10, 2026
Supervised Machine Learning for Semi-Quantification of Extracellular DNA in Glomerulonephritis
Published on: June 18, 2020
AI-Driven Quantitative Analysis of Pathological Images for Membranous Nephropathy Across Macro and Micro Modalities
Abstract:
The diagnosis of membranous nephropathy (MN) has been reliant on the identification of glomerular basement membrane (GBM) variations and lesions at both macro and micro levels. At the macro level, light microscopy (LM) has been used to reveal spike- like projections that indicate pathological changes, whereas at the micro level, transmission electron microscopy (TEM) has been employed to identify GBM thickening. However, qualitative diagnosis has been limited by inter-pathologist variability, creating the need for deep learning approaches capable of quantifying pathological changes and predicting MN progression. In this study, an AI-driven framework based on the Mamba model has been proposed, in which the area and proportion of spike- like projections are quantified at the macro level, and GBM thickness is segmented and measured at the micro level. Classical machine learning models are then applied to predict MN progression based on pathological indicators extracted through factor analysis. Unlike prior approaches, the framework has been designed to emulate the diagnostic workflow of pathologists by integrating LM and TEM images for joint analysis. Experiments on an external dataset of 109 cases demonstrated strong performance in glomeruli classification, GBM segmentation, and MN progression prediction. These findings highlight the potential of multi-scale integrated quantification to provide objective, reproducible, and clinically interpretable assessment of MN progression.
Insights
This study introduces an AI framework using Mamba models to analyze kidney biopsy images for diagnosing membranous nephropathy (MN). It quantifies pathological changes for objective and reproducible MN progression prediction.
Area of Science:
- Nephrology
- Medical Imaging
- Artificial Intelligence
Background:
- Membranous nephropathy (MN) diagnosis relies on identifying glomerular basement membrane (GBM) lesions via light microscopy (LM) and transmission electron microscopy (TEM).
- Qualitative diagnostic methods suffer from inter-pathologist variability, necessitating quantitative approaches for accurate MN progression prediction.
Purpose of the Study:
- To develop an AI-driven framework for quantifying pathological changes in MN from LM and TEM images.
- To predict MN progression using integrated multi-scale image analysis and machine learning.
Main Methods:
- An AI framework utilizing the Mamba model was developed to quantify spike-like projections (LM) and GBM thickness (TEM).
- Factor analysis was used to extract pathological indicators for predicting MN progression with classical machine learning models.
- The framework integrates LM and TEM images for joint analysis, mimicking pathologist workflow.
Main Results:
- The AI framework demonstrated strong performance in glomeruli classification and GBM segmentation.
- Accurate prediction of MN progression was achieved using the developed pathological indicators.
- Experiments on an external dataset of 109 cases validated the framework's effectiveness.
Conclusions:
- Multi-scale integrated quantification offers an objective and reproducible method for assessing MN progression.
- The AI framework has the potential to enhance clinical interpretability in MN diagnosis.
- This approach addresses the limitations of qualitative diagnostic methods in nephrology.
