Related Experiment Video
Updated: Jul 13, 2025
06:09
Semi-Quantitative Determination of Dopaminergic Neuron Density in the Substantia Nigra of Rodent Models using Automated Image Analysis
Published on: February 2, 2021
4.5K
Quantitative analysis of prion disease using an AI-powered digital pathology framework
Massimo Salvi1, Filippo Molinari2, Mario Ciccarelli2
1Biolab, PoliTo(BIO)Med Lab, Department of Electronics and Telecommunications, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129, Turin, Italy. massimo.salvi@polito.it.
Scientific Reports
|October 18, 2023
Summary
This study introduces an AI pipeline for identifying prion protein aggregates in brain tissue, improving diagnosis of prion disease. The automated system accurately quantifies protein deposition, aiding neurodegenerative disease research.
Area of Science:
- Neuroscience
- Pathology
- Artificial Intelligence
Background:
- Prion diseases are fatal neurodegenerative disorders characterized by abnormal prion protein (PrPSc) accumulation.
- Current diagnostics rely on immunohistochemical staining of tissue samples.
- Digital pathology and AI offer new avenues for analyzing stained slides.
Purpose of the Study:
- To develop an automated AI pipeline for identifying PrPSc aggregates in digital pathology images.
- To evaluate the efficacy of deep learning and machine learning approaches for PrPSc detection.
- To establish a framework for quantifying PrPSc deposition in brain tissue.
Main Methods:
- Developed an automated pipeline using a vision transformer (deep learning) and traditional classifiers (machine learning).
- Trained and tested the pipeline on 64 whole slide images from 41 patients with confirmed prion disease.
- Evaluated PrPSc identification and quantification in cerebellar and occipital cortex tissue samples.
Main Results:
- The AI framework accurately classified whole slide images from a blind test set.
- The pipeline successfully quantified PrPSc distribution and localization within brain tissue.
- Demonstrated the first framework for evaluating PrPSc deposition in digital images.
Conclusions:
- AI-assisted pathology can significantly improve diagnostic accuracy and efficiency for prion disease.
- The developed pipeline offers a quantitative method for assessing PrPSc pathology.
- This approach holds potential for application in other neurodegenerative diseases, such as Alzheimer's and Parkinson's disease.
