Real-time release testing of dissolution based on surrogate models developed by machine learning algorithms using NIR...

Area of Science:

• Pharmaceutical Sciences
• Materials Science
• Computational Chemistry

Background:

• Predicting drug dissolution profiles is crucial for sustained-release tablet quality.
• Critical Material Attributes (CMAs) and process data are key factors influencing drug release.
• Particle Size Distribution (PSD) of matrix polymers significantly impacts drug release kinetics.

Purpose of the Study:

• To develop and compare machine learning models for predicting in vitro dissolution profiles of sustained-release tablets.
• To evaluate the impact of including matrix polymer PSD data on dissolution prediction accuracy.
• To integrate Process Analytical Technology (PAT) data with CMAs for Real-Time Release Testing (RTRT).

Main Methods:

• Spectroscopic measurements (NIR), process data, and CMAs were utilized.
• Three machine learning methods were employed: Artificial Neural Networks (ANN), Support Vector Machines (SVM), and Ensemble of Regression Trees (ERT).
• Models were developed with and without PSD data to assess its significance.

Main Results:

• Matrix polymer PSD was identified as a significant factor influencing drug release.
• Artificial Neural Networks (ANN) provided the most accurate dissolution profile predictions compared to SVM and ERT.
• The inclusion of PSD data improved the predictive performance of the models.

Conclusions:

• Machine learning models, particularly ANN, can accurately predict sustained-release tablet dissolution profiles.
• Integrating PAT data with CMAs, including PSD, enables effective Real-Time Release Testing (RTRT).
• This approach enhances quality control and reduces the need for extensive dissolution testing.