A novel image analysis approach for evaluation of mixing uniformity in drug-filled silicone rubber matrix

• Drugs-polymer matrix blend uniformity was analyzed by SEM images processing.
• In dispersion level, frequency of drug(s) aggregates’ vs. particle size was traced.
• Studying distribution patterns by two approaches resulted in the same conclusion.
• Assuming uniform drug localization, an ideal matrix was developed.
• The ratio of white (drug) to black + white pixels (matrix + drug) was calculated.

Nowadays, there is a lot of interest in developing long-acting drug delivery devices for human or veterinary applications including monolithic systems. Drug content uniformity of a monolithic device is highly dependent on the uniform distribution of drug particles within the polymeric matrix both in dispersion and distribution levels. Here, a range of formulations were prepared which consisted of progesterone (1%w/w) and estradiol benzoate (0.1%w/w) dispersed in a silicone rubber matrix. Blend uniformity of the compounds was analyzed by image analysis of SEM micrographs obtained from the cross-sections of the devices by a new image processing approach. Efficiency of mixing was investigated at the dispersion level by plotting the relative frequency of drug particles versus “projected area diameter” of their aggregates. Based on the particle size distribution results, a significant improvement was observed in the dispersion pattern of drug particles by adding silicone oil (9%w/w). Distribution pattern of the particles was investigated by transforming the micrographs into algebraic matrices. An “ideal matrix” was developed by assumption of uniform localization of the drug particles. Real matrices obtained for all of the formulations were compared with this ideal matrix as a reference. Closer similarity between the two matrices was observed for silicone oil-containing (9%w/w) samples showing the best dispersive and distributive mixing quality.

Figure optionsDownload high-quality image (175 K)Download as PowerPoint slide