In-Process Control Assay of Pharmaceutical Microtablets Using Hyperspectral Imaging Coupled with Multivariate Analysis

Abstract

Monitoring the amount of active pharmaceutical ingredient (API) in finished dosage form is important to ensure the content uniformity of the product. In this report, we summarize the development and validation of a hyperspectral imaging (HSI) technique for rapid in-line prediction of the active pharmaceutical ingredient (API) in microtablets with concentrations varying from 60 to 130% API (w/w). The tablet spectra of different API concentrations were collected in-line using an HSI system within the visible/near-infrared (vis/NIR; 400-1000 nm) and short-wave infrared (SWIR; 1100-2500 nm) regions. The ability of the HSI technique to predict the API concentration in the tablet samples was validated against a reference high-performance liquid chromatography (HPLC) method. The prediction efficiency of two different types of multivariate data modeling methods, that is, partial least-squares regression (PLSR) and principle component regression (PCR), were compared. The prediction ability of the regression models was cross-validated against results generated with the reference HPLC method. The results obtained from the PLSR models showed reliable performance for predicting the API concentration in SWIR region. The highest coefficient of determination (Rp) was 0.96, associated with the lowest predicted error and bias of 4.45 and -0.35%, respectively. Furthermore, the concentration-mapped images of PLSR and PCR models were used to visually characterize the API concentration present on the tablet surface. Based on these results, we suggest that HSI measurement combined with multivariate data analysis and chemical imaging could be implemented in the production environment for rapid in-line determination of pharmaceutical product quality.

Hyperspectral imaging sensor for optimization of small molecule formulations

“…The beta coefficient plot from the PLS model presented in Figure was used to confirm the peaks related to the chemical structure of the API, which was composed of benzene and carboxylic acid groups. Therefore, in Figure , the peaks at 1,143 nm (C‐H second overtone), 1,446 nm (C‐H combination) and 1,685 nm (C‐H first overtone) are characteristic of the aromatic structure, while the peak at 1,890 nm corresponds to the carboxylic group (Kandpal, Tewari, Gopinathan, Boulas, & Cho, ). These above groups were attributed to the overall overtones of the API structure.…”

Spectral imaging and spectral LIDAR systems: moving toward compact nanophotonics-based sensing

“…The spectral imaging technology, which is able to obtain both spatial and spectral information, was originally applied in Earth remote sensing [3]. Currently, it has been widely utilized in remote and indoor sensing, covering from Earth observation, geo-information study [4][5][6][7] to optical sorting and pharmaceutical analysis [8,9].…”

New perspectives of hyperspectral imaging for clinical research

“…13,14,18,19 Today, HSI is a promising non-invasive and non-ionizing technique that supports the rapid acquisition and analysis of diagnostic information in several fields. This technology is nowadays used for remote sensing, 20,21 archeology, 22,23 drug identification, 24–27 forensics, 28–30 defense and security, 31,32 agriculture, 33,34 food safety inspection and control. 30,35,36…”