Development of Co-Amorphous Systems for Inhalation Therapy-Part 1: From Model Prediction to Clinical Success.

Area of Science:

• Pharmaceutical Sciences
• Computational Chemistry
• Materials Science

Background:

• Machine learning (ML) and artificial intelligence (AI) are transforming pharmaceutical processes.
• An ML model was developed to predict binary co-amorphous systems (COAMSs) for inhalation therapy.
• The model's capacity for creating suitable dry powder formulations was assessed.

Purpose of the Study:

• To experimentally validate an ML model for predicting COAMS formation.
• To develop and characterize a rifampicin-ethambutol COAMS for tuberculosis inhalation therapy.
• To evaluate the aerosolization and dissolution properties of the developed COAMS.

Main Methods:

• Extended experimental validation using co-milling and X-ray diffraction for 18 active pharmaceutical ingredient (API) combinations.
• Development of a rifampicin-ethambutol (RIF-ETH) COAMS via spray-drying and co-milling.
• Assessment of aerosolization properties (e.g., emitted dose, fine particle fraction) and dissolution profiles.

Main Results:

• The ML model achieved 79% prediction accuracy for 35 APIs; experimental validation yielded 72% accuracy.
• Stable RIF-ETH COAMS were successfully developed using both spray-drying and co-milling.
• Co-milled RIF-ETH COAMS exhibited superior aerosolization and more reproducible dissolution characteristics.

Conclusions:

• ML is a valuable tool for predicting COAMS, reducing experimental screening time and costs.
• The developed RIF-ETH COAMS show promise for tuberculosis treatment via inhalation.
• This predictive modeling approach accelerates the development of effective inhalation therapies.