Raman Spectroscopy and Exosome-Based Machine Learning Predicts the Efficacy of Neoadjuvant Therapy for HER2-Positive...

Area of Science:

• Oncology
• Biomarkers
• Nanotechnology

Background:

• Early prediction of neoadjuvant therapy efficacy is vital for HER2-positive breast cancer patients.
• Exosomes show promise as biomarkers for treatment response but face detection challenges.
• Current exosome isolation methods are inefficient, time-consuming, and yield low purity/quantity.

Purpose of the Study:

• To develop a non-invasive method for early prediction of neoadjuvant therapy efficacy in HER2-positive breast cancer.
• To analyze molecular changes in exosomes using Raman spectroscopy and machine learning.
• To introduce an efficient HER2-positive exosome capture and detection system.

Main Methods:

• Raman spectroscopy was used to analyze exosomes from HER2-positive breast cancer patients' sera before and after neoadjuvant therapy.
• Machine learning algorithms (PCA, LDA, SVM) were employed to build a predictive model.
• A novel Magnetic beads@HER2-Exos@HER2-SERS detection nanoprobes (HER2-MEDN) system was developed for exosome capture and analysis.

Main Results:

• A predictive model using Raman spectroscopy and machine learning achieved an AUC > 0.89.
• The HER2-MEDN system enabled efficient extraction and analysis of HER2-positive exosomes.
• The refined predictive model incorporating HER2-MEDN achieved an accuracy > 0.94 for early treatment response prediction.

Conclusions:

• The HER2-MEDN system demonstrates significant potential for accurate early prediction of neoadjuvant therapy response in HER2-positive breast cancer.
• This study offers novel insights and methodologies for assessing treatment efficacy, paving the way for personalized medicine.
• The developed system addresses limitations of current exosome detection methods, improving diagnostic capabilities.