A machine learning framework for predicting entrapment efficiency in niosomal particles.

Area of Science:

• Pharmaceutical Sciences
• Computational Chemistry
• Materials Science

Background:

• Niosomes are nonionic surfactant-based vesicles crucial for drug delivery.
• Optimizing drug entrapment efficiency in niosomes is vital but experimentally challenging.
• Current optimization methods are time-consuming and expensive.

Purpose of the Study:

• To develop a machine learning framework for predicting optimal niosome formulation parameters.
• To identify the most critical parameter influencing drug entrapment efficiency.
• To reduce the cost and time associated with niosome formulation optimization.

Main Methods:

• Data extraction from recent literature on niosomes and thin-film hydration.
• Training deep neural network (DNN), linear, and polynomial regression models.
• Sensitivity analysis to determine the most influential parameter (hydrophilic-lipophilic balance - HLB).
• Experimental validation using doxycycline hyclate-loaded niosomes.

Main Results:

• The DNN model achieved the best performance with R-squared of 0.763 ± 0.1.
• Hydrophilic-lipophilic balance (HLB) was identified as the most influential parameter.
• Optimal HLB values were determined for maximizing entrapment efficiency.
• Experimental results confirmed the model's predictions.

Conclusions:

• Machine learning offers an efficient and cost-effective approach to niosome formulation optimization.
• DNN models can accurately predict and optimize drug entrapment efficiency.
• Identifying key parameters like HLB is crucial for targeted niosome design.