sGAN: A Digital Sandbox for Drug Product Optimization

Developing drug products often relies on a costly make-and-test approach, and the final results can be complicated by manufacturing variability. As a result, formulation and scale-up can be time-consuming and subject to significant uncertainty.

In this webinar, we will introduce sGAN (structural Generative Adversarial Network), a new tool designed to reduce guesswork in drug product development by creating a high-fidelity digital twin of a test product from real microstructure imaging data.

We will discuss how this software streamlines product development, reduces the number of physical samples required, and helps shorten time to market.

We will also show how this tool can be combined with digiM’s dissolution and release modeling tools to extend in silico modeling capabilities, giving users greater freedom to explore potential parameter space while further improving quality and accuracy.

We will also present two partner case studies: one with Merck focused on long-acting injectable development, and one with Genentech focused on oral tablet formulation optimization.

‍