PruEV-AI: a Simple Approach Combines Urinary Extracellular Vesicle Isolation with AI-Assisted Analysis for Prostate...

Area of Science:

• Biomarkers
• Extracellular Vesicles
• Artificial Intelligence

Background:

• Urinary extracellular vesicles (uEVs) show promise for non-invasive prostate cancer (PCa) diagnosis.
• Current uEV isolation and single-marker assays have limitations in efficiency and accuracy.

Purpose of the Study:

• To introduce PruEV-AI, an integrated system for rapid uEV isolation and AI-based biomarker analysis.
• To develop a highly accurate, non-invasive diagnostic approach for PCa.

Main Methods:

• PruEV platform uses amine-modified zeolites (AZ) and carbohydrazide (CDH) for rapid uEV and miRNA isolation (<30 min).
• RT-qPCR validated 6 candidate miRNAs from urine samples of PCa patients and controls.
• A deep learning (DL) model analyzed combinations of 6 miRNAs and PSA to identify optimal biomarkers.

Main Results:

• Individually, candidate miRNAs and PSA showed modest diagnostic performance (AUC 0.6–0.8).
• The DL model identified an optimal biomarker combination with an AUC of 0.9556.
• The PruEV-AI system achieved 93.33% sensitivity, specificity, and accuracy for PCa detection.

Conclusions:

• The PruEV-AI system offers a robust and non-invasive diagnostic method for PCa.
• This approach supports improved screening protocols and therapeutic strategies for PCa.