Related Experiment Video
Updated: Jan 14, 2026
11:50
A Standardized Pipeline for Examining Human Cerebellar Grey Matter Morphometry using Structural Magnetic Resonance Imaging
Published on: February 4, 2022
4.5K
Automated morphological classification and quantification of cerebrospinal fluid extracellular vesicles via AFM and
Mario Kurtjak1, Vera Tomas2, Leon Ivović3
1Advanced Materials Department, Jožef Stefan Institute SI-1000 Ljubljana Slovenia.
Nanoscale Advances
|October 23, 2025
Summary
Optimizing sample preparation is key for accurate extracellular vesicle (EV) morphology analysis using atomic force microscopy (AFM). We developed a machine learning tool to identify EV shapes, revealing that ethanol dehydration and critical point drying best preserve EV structure for diagnostic potential.
Area of Science:
- Biophysics
- Nanotechnology
- Biochemistry
Background:
- Extracellular vesicles (EVs) in cerebrospinal fluid are crucial biomarkers for brain conditions.
- Atomic force microscopy (AFM) in air offers accessible EV visualization but can distort native morphology.
- Accurate EV morphology assessment is vital for reliable diagnostic applications.
Purpose of the Study:
- To systematically compare 24 preparation methods for cerebrospinal fluid EVs using AFM.
- To develop computational tools for automated EV morphology analysis and shape classification.
- To identify optimal preparation techniques for preserving EV morphology during AFM imaging.
Main Methods:
- Compared 24 EV preparation methods, analyzing size, height, aspect ratio, and shape distributions via AFM.
- Defined 5 EV shape categories (round, flat, concave, single-lobed, multilobed) and excluded artefacts.
- Developed a computer program for manual EV observation and shape identification.
- Employed a machine learning (convolutional neural network) model for automated EV shape recognition (85 ± 5% F1 score).
Main Results:
- Fixation methods significantly impact EV capture and protection on substrates.
- Critical point drying preserves EV morphology better than hexamethyldisilazane.
- (3-aminopropyl)triethoxysilane can cause EV flattening; NiCl2 promotes artefact formation.
- Ethanol gradient dehydration followed by critical point drying best preserved EV morphology.
- Chemical dehydration with dimethoxypropane yielded balanced shapes and lower aspect ratios.
- Highest aspect ratios, matching liquid AFM, were achieved with ethanol dehydration and critical point drying on NiCl2-coated mica.
Conclusions:
- Optimized sample preparation is crucial for accurate EV morphology analysis using AFM.
- Machine learning significantly enhances the speed and objectivity of EV shape classification.
- Specific preparation techniques, particularly ethanol dehydration and critical point drying, are superior for preserving near-native EV morphology.
- These findings advance the potential of AFM-based EV analysis for diagnostic purposes.
