High-Speed Imaging-Based Particle Attribute Analysis of Spray-Dried Amorphous Solid Dispersions Using a Convolution...

Area of Science:

• Pharmaceutical Technology
• Materials Science
• Chemical Engineering

Background:

• Spray drying is crucial for amorphous solid dispersions (ASDs) to enhance drug bioavailability.
• Characterizing particle attributes like size and morphology in spray-dried intermediates (SDIs) is vital for drug product performance.
• Existing microscopy methods lack the high-throughput needed for SDI analysis.

Purpose of the Study:

• To develop a rapid, high-throughput method for characterizing particle size and morphology distributions in spray-dried ASDs.
• To create a "morphological fingerprint" for SDIs using advanced imaging and machine learning.
• To demonstrate the method's utility in guiding spray drying process development.

Main Methods:

• Utilized high-speed dynamic imaging combined with laser diffraction (LD).
• Developed unsupervised and supervised convolutional neural network (CNN) models for image analysis.
• Applied the method to SDIs containing hypromellose acetate succinate, varying spray drying parameters.

Main Results:

• Demonstrated that SDI batches are mixtures of diverse particle subpopulations.
• The CNN model rapidly computed volumetric composition, generating a morphological fingerprint.
• Results correlated well with LD and electron microscopy data, showing robustness.

Conclusions:

• The developed high-speed imaging-based fingerprinting approach is robust and high-throughput.
• This method enables accurate quantification of particle size and morphological distributions in SDIs.
• The technique can guide spray drying process development for optimized drug product performance.