UV imaging for the rapid at-line content determination of different colourless APIs in their tablets with artificial neural networks.

This paper presents a novel high-resolution and rapid (50 ms) UV imaging system, which was used for at-line, non-destructive API content determination of tablets. For the experiments, amlodipine and valsartan were selected as two colourless APIs with different UV induced fluorescent properties according to the measured solid fluorescent spectra. Images were captured with a LED-based UV illumination (385-395 nm) of tablets containing amlodipine or valsartan and common tableting excipients. Blue or green colour components from the RGB colour space were extracted from the images and used as an input dataset to execute API content prediction with artificial neural networks. The traditional destructive, solution-based transmission UV measurement was applied as reference method. After the optimization of the number of hidden layer neurons it was found that the relative error of the content prediction was 4.41 % and 3.98 % in the case of amlodipine and valsartan containing tablets respectively. The results open the possibility to use the proposed UV imaging-based system as a rapid, in-line tool for 100 % API content screening in order to greatly improve pharmaceutical quality control and process understanding.

Raman-based real-time dissolution prediction using a deterministic permeation model.

The purpose of this study was to develop a deterministic permeation model (DPM) that predicts the in vitro release profile of an active ingredient (API) embedded in hydroxypropyl-methylcellulose (HPMC) matrix tablets based on Raman spectra. So far in the literature, such mechanistic models were utilized only for formulation optimization (off-line dissolution prediction), while the real-time prediction of dissolution profiles based on Process Analytical Technology (PAT) data was performed by empirical methods such as Partial Least Squares (PLS) regression. Our work represents a novel conceptual approach that utilizes a mechanistic model to predict dissolution profiles based on data yielded by PAT tools. Tablets containing various API- and HPMC-amounts were produced using different compression pressures according to a 33 full factorial design, their Raman spectra were recorded before dissolution testing. The DPM was constructed using one-third of the measured dissolution profiles and is presented as a system of differential equations together with its analytical solution. The parameters of DPM were estimated by the training data set containing the spectroscopically determined API- and HPMC- amounts and the tableting pressures used, then the release profiles of the remaining two-thirds of the tablets were predicted. The Raman spectra-based predictions of DPM were compared with predictions of an Artificial Neural Network (ANN). It was found that the two methods yield similar results, however, the mechanistic approach has the benefit of requiring a lower amount of training samples. Although the model is based on a remarkable simplification of reality, it facilitates a deeper understanding of the behavior of the formulation. The DPM could improve our understanding of the effect of HPMC and tableting pressures on the release kinetics of the HPMC matrix tablets and participate in the development of PAT-based new surrogate dissolution methods for Real-Time Release testing (RTRt).