Drug product characterization by macropixel analysis of chemical images.
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
Traditional monitoring of pharmaceutical manufacturing combines physical sampling and analytical methodologies (e.g. HPLC). Process analytical technology (PAT) can be implemented to collect real-time measurements, although successful monitoring requires that sampling be representative. The maximum spot size for a spectroscopic tool (e.g. near-infrared; Raman) should be equivalent to a single dosage size. A smaller spot size may provide a PAT tool that is sensitive to monitoring process changes, but if too small, produces non-reproducible data. The current study uses chemical imaging to determine appropriate spot size. A chemical image is an array of pixels which maps the chemical composition of the sample. "Macropixel Analysis" is introduced as a measure of image heterogeneity based on clusters of pixels (macropixels) within near-infrared chemical images. Analyses were conducted using non-overlapping tiles of macropixels (Discrete-Level Tiling) and all possible macropixels of the image (Continuous-Level Moving Block). Both methods minimize the variance between macropixel intensities by varying the size of the macropixels. Spot size is then chosen as the minimum macropixel size for which the range of macropixel intensities falls within an acceptable criterion. Both imaging-based algorithms provide useful quantitative information about the heterogeneity of pharmaceutical products.
