Design of controlled-release drugs using image-based modelling and machine learning

Controlled-release tablets are manufactured to dissolve slowly and release active ingredients at a predictable rate so they can sustain a small amount of medication into a patient's system over an extended period. Such systems give physicians better control over the patient's health. However, designing a micro-structure that guarantees such controlled release is subject to uncertainty and demands complex numerical models. In this project, we use image-based and machine learning techniques to simulate the dissolution process. Actual 3-D images of a model pharmaceutical tablet obtained by micro computed tomography (micro-CT) will be used and by digital optimization of the tablets micro-structures the structure that leads to the desired rate of drug release will be explored. The outcome of this project will be an intelligent computational model that can be used to design new controlled-release tablets tailored to the need of patients. ​