Predicting Calcein Release from Ultrasound-Targeted Liposomes: A Comparative Analysis of Random Forest and Support...

Area of Science:

• Biomedical Engineering
• Drug Delivery Systems
• Machine Learning Applications

Background:

• Targeted liposomes offer precise drug delivery.
• Ultrasound-triggered release allows for spatiotemporal control.
• Predictive models are crucial for optimizing drug delivery parameters.

Purpose of the Study:

• To compare the predictive performance of Random Forest (RF) and Support Vector Machine (SVM) algorithms.
• To evaluate these models for predicting calcein release from ultrasound-triggered, targeted liposomes.
• To assess model performance under varying low-frequency ultrasound (LFUS) power densities.

Main Methods:

• Synthesis and characterization of liposomes loaded with calcein and targeted with seven different moieties.
• Utilized Dynamic Light Scattering and Bicinchoninic Acid assays for characterization.
• Trained and evaluated RF and SVM models using metrics like MAE, MSE, R², and a20 index.

Main Results:

• Random Forest (RF) models consistently showed superior performance compared to SVM.
• RF achieved R² scores exceeding 0.96 across all tested LFUS power densities.
• RF demonstrated a strong correlation with experimental data, especially at higher power densities.

Conclusions:

• Random Forest (RF) is a more effective algorithm than Support Vector Machine (SVM) for predicting drug release from targeted liposomes.
• Both models have potential applications depending on specific predictive requirements.
• The findings support the use of machine learning for optimizing ultrasound-mediated drug delivery.