Advanced Therapy Medicinal Products (ATMPs) is a rapidly emerging field in medicine with great potential, where there is a growing need for robust and accurate methods to characterize and classify processes to assess safety aspects. This project gathers stakeholders from academia, research institutes and industry from various disciplines to develop AI-based methods for safety assessment of ATMPs.
New safety biomarkers to identify abnormalities at the early stages of the manufacturing process of ATMPs are explored. The project promotes advanced digitalization in the ATMP sector to improve weaknesses in today’s standard methods for predicting quality problems that can potentially lead to large losses of resources and reduced quality of life for patients.
In the project AI models are trained with omics data generated from various stages of the manufacturing of individualized blood vessels, and from data from native blood vessels from the recipient and from the transplants one year after surgery. Siamese Neural Network-models are developed for the calculation of similarity scores to assess the safety of ATMP and improve safety assessment test so that successful transplants can be predicted.
Furthermore, Variational Autoencoders will be developed and trained on integrated omics data to identify novel safety biomarkers by extracting biological information represented in the latent variables. The validated AI models will be combined in a software package for the safety assessment of ATMPs, and with future integration into commercial software, where the scalability of the manufacturing methods for a large number of ATMPs will be improved.
The project uses innovative AI solutions to promote good health and welfare and will greatly benefit Swedish ATMP entrepreneurs.
The project is coordinated by the University of Skövde and is executed in close collaboration with RISE Research Institute of Sweden and our industrial partners VERIGRAFT AB and SciCross AB.