Advancing cell therapy manufacturing: an image-based solution for accurate confluency estimation

Abstract

Cell therapies represent a transformative approach for treating diseases resistant to conventional therapies, yet their development and manufacturing face significant hurdles within the biopharmaceutical sector. A critical parameter in the production of these therapies is cell confluency, which serves as both an indicator of biomass in adherent cultures and a determinant of product quality. However, existing methods for measuring confluency are often inadequate for the large-scale cultivation systems used in industry, and current software solutions lack comprehensive automation capabilities necessary for a manufacturing environment. This article introduces a novel image-based software application designed for accurate cell confluency estimation, integrated with a high-throughput microscopy system. Utilizing a machine-learning model for pixel classification, the application facilitates efficient image and metadata processing in a cloud environment, delivering results through an interactive web interface. By incorporating methods from process analytical technologies, manufacturing data digitalization, and data science, this platform enables automated image acquisition, storage, analysis, and reporting in near-real time. The proposed solution aims to streamline the manufacturing process of cell therapeutics, ultimately enhancing the reliability and speed of delivering these innovative treatments to patients.