Machine-learning methods to improve CAR-T multiple myeloma therapies

Felipe Prosper

Clínica Universidad de Navarra (CUN), Pamplona, Spain

The development of CAR-T therapies using the patient’s own genetically modified immune cells has yielded promising results in various haematological diseases over the past decade. However, in a significant number of cases, despite an initial positive response, patients treated with this type of therapy stop responding after several months, particularly those diagnosed with multiple myeloma. This is because the modified CAR-T cells do not persist and the tumour cells develop resistance mechanisms.

Although work is ongoing to improve these therapies, the mechanisms that regulate the function of genetically modified T cells after they are administered to patients are still not fully understood. This project will try to decipher them. To achieve this, samples will be used from patients currently receiving CAR-T therapies.

Using new machine learning methods and recent advances in single-cell analysis technologies, which provide an unprecedented level of resolution, the project aims to shed light on the regulatory mechanisms. Ultimately, this will enable the efficacy of CAR-T cell therapies to be improved, with a particular focus on those aimed at treating multiple myeloma.

• Juan Roberto Rodríguez Madoz, Centro de Investigación Médica Aplicada (CIMA), Universidad de Navarra, Spain

• Nir Yosef, Weizmann Institute of Science, Rehovot, Israel

Uncovering resistance mechanism to CAR-T cells via deep learning in scRNAseq for improved therapies

€991,550.00