Liposome Particle Size Prediction by In-Line Process Analytical Technology (PAT)-Integrated Machine Learning.

Area of Science:

• Pharmaceutical Sciences
• Biotechnology
• Chemical Engineering

Background:

• Accurate control of liposome size is essential for effective drug delivery systems.
• Current methods for liposome size determination can be time-consuming and require extensive experimental data.

Purpose of the Study:

• To develop an in-line process analytical technology (PAT) integrated machine learning model for predicting liposome particle size.
• To achieve high accuracy and generalization in liposome size prediction using limited experimental data.

Main Methods:

• Development of a machine learning model integrated with in-line PAT.
• Utilizing physicochemical membrane characteristics as input features.
• Validation of the model using experimental data to assess accuracy and generalization.

Main Results:

• The developed model achieved high accuracy in predicting liposome particle size with a root mean square error of 7.18 nm.
• The model demonstrated strong generalization capabilities, with a root mean square error of 7.53 nm when incorporating physicochemical membrane characteristics.
• The model provides interpretability, offering insights into the factors influencing liposome size.

Conclusions:

• The study establishes a practical and accurate framework for advanced liposome particle size control.
• The PAT-integrated machine learning approach offers a significant advancement for optimizing liposome manufacturing processes.
• This methodology has the potential to enhance the development and efficacy of liposome-based drug delivery systems.