Quantum machine learning-based electrokinetic mining for the identification of nanoparticles and exosomes with...

Area of Science:

• Nanotechnology and Biomaterials Science
• Quantum Computing and Machine Learning
• Biomedical Engineering and Diagnostics

Background:

• Nanoparticles (NPs) and extracellular vesicles (EVs) are crucial in biomaterials, oncology, and drug delivery.
• Traditional characterization methods (e.g., electron microscopy) are costly and time-consuming.
• Electrokinetic attributes offer a rapid, cost-effective alternative for particle characterization.

Purpose of the Study:

• To introduce the first application of quantum machine learning (QML)-based electrokinetic mining for NP and EV identification.
• To assess QML's ability to differentiate various NPs and EVs, including those from cancer patients.
• To evaluate QML's performance with minimal training data for clinical applications.

Main Methods:

• Utilized QML and classical machine learning (ML) for electrokinetic data analysis.
• Applied cross-validation, train-test splits, confusion matrices, and ROC curves for analysis.
• Tested QML on green-synthesized NPs and exosomes from cell lines and patient plasma.

Main Results:

• Classical ML accurately identified NPs and EVs based on electrokinetic data.
• QML demonstrated superior differentiation capabilities between various NPs and EVs.
• QML successfully distinguished EVs in colorectal cancer (CRC) patient plasma from healthy individuals.
• QML achieved high prediction performance in identifying NPs and EVs in CRC patient plasma and mice, even with limited training data.

Conclusions:

• QML-based electrokinetic mining is a powerful tool for identifying NPs and EVs.
• This method offers high accuracy and efficiency, especially with minimal training data.
• QML facilitates novel clinical development, particularly in liquid biopsies for cancer detection.