Related Experiment Video
Updated: May 3, 2026
09:16
Supervised Machine Learning for Semi-Quantification of Extracellular DNA in Glomerulonephritis
Published on: June 18, 2020
6.8K
PICASO Set Operator for Computational Nephropathology
Samira Zare1, Huy Q Vo1, Nicola Altini2
1Department of Electrical and Computer Engineering, University of Houston, Houston, Texas, USA.
Kidney360
|March 3, 2025
Summary
PICASO, a novel deep learning tool, enhances digital nephropathology by dynamically aggregating histopathologic features. It significantly improves the detection of active crescent lesions in IgA nephropathy and classification of antibody-mediated rejection in kidney transplants.
Area of Science:
- Nephropathology
- Digital Pathology
- Artificial Intelligence in Medicine
Background:
- Digital nephropathology integrates deep learning for improved diagnostics.
- PICASO is a novel permutation-invariant set operator for dynamic histopathologic feature aggregation.
- Applications include detecting active crescent lesions in IgA nephropathy and classifying antibody-mediated rejection (AMR) in kidney transplants.
Purpose of the Study:
- To introduce and evaluate PICASO, a Transformer-based set operator for nephropathology.
- To assess PICASO's performance in identifying active crescent lesions and classifying AMR.
- To compare PICASO against other set operators in diagnostic accuracy.
Main Methods:
- PICASO utilizes a Transformer architecture with static memory (Histopathologic Vectors) and dynamic updates.
- Applied to 6206 glomerular crops for IgA nephropathy active crescent detection.
- Used on 1655 glomerular crops from 89 biopsies for kidney transplant AMR classification.
Main Results:
- PICASO achieved superior AUROC of 0.99 (internal) and 0.96 (external) for active crescent detection, outperforming other methods (P<0.001).
- Attained the highest AUROC of 0.97 for case-level AMR classification (P=0.02).
- Demonstrated significantly higher AUPR, recall, and accuracy compared to baselines (P<0.001).
Conclusions:
- PICASO shows potential to advance nephropathology diagnostics.
- Dynamic feature aggregation by PICASO improves diagnostic performance.
- The tool offers a promising approach for integrating AI into kidney disease diagnosis.
More Related Videos
Related Concept Videos
01:24Renal Tubule and Collecting Duct
4.4K
The renal tubule is divided into three parts: the proximal convoluted tubule (PCT), the Loop of Henle (LOH), and the distal convoluted tubule (DCT).
Proximal Convoluted Tubule (PCT):
The PCT is the initial segment of the renal tubule, extending from the Bowman's capsule that encloses the glomerulus. Its convoluted structure and microvilli-lined cells increase the surface area for reabsorption. The PCT reabsorbs glucose, amino acids, sodium, and water from the filtrate, ensuring essential...
Proximal Convoluted Tubule (PCT):
The PCT is the initial segment of the renal tubule, extending from the Bowman's capsule that encloses the glomerulus. Its convoluted structure and microvilli-lined cells increase the surface area for reabsorption. The PCT reabsorbs glucose, amino acids, sodium, and water from the filtrate, ensuring essential...
4.4K
01:28Imaging Studies I: Kidney, Ureter, and Bladder Studies
871
Kidney, Ureter, and Bladder (KUB) StudiesKidney, Ureter, and Bladder (KUB) studies are standard diagnostic imaging procedures used to assess the anatomy of the urinary system. They are commonly utilized for patients experiencing abdominal pain or urinary symptoms. By using a simple X-ray of the abdomen, KUB studies can reveal structural and pathological abnormalities within the kidneys, ureters, and bladder. These studies are particularly valuable in diagnosing kidney stones, urinary...
871