« Back to Farmacia Journal 4/2021

ARTIFICIAL NEURAL NETWORKS FOR INVESTIGATION OF THE MOST IMPORTANT FACTORS OF INDUSTRIAL TABLET MANUFACTURING ON THE DISSOLUTION OF ACTIVE PHARMACEUTICAL INGREDIENTS AS CRITICAL QUALITY ATTRIBUTES

MARCEL PRIKERŽNIK ^1,2*, STANKO SRČIČ ¹

Download Full Article PDF

This research aimed to show the usefulness of artificial neural networks (ANNs) to investigate significant factors on the biopharmaceutical relevant parameter. It is the first time ANNs were used on such a comprehensive set of data collected during two-years of industrial production of tablets containing two active pharmaceutical ingredients (APIs). The feature selection technique was applied to the ANNs models to recognize critical material attributes and process parameters. Altogether, we have identified seven critical material attributes related to APIs, croscarmellose sodium, magnesium stearate, and colloidal silicon dioxide, together with five critical process parameters related to the roller compaction and tablet compression. The study demonstrates the usefulness of ANNs in facilitating a higher level of tablet manufacturing understanding.