Improving Cancer Therapy with Machine Learning 3


Daniel Reker, Jee Won Yang, Natsuda Navamajiti, Ruonan Cao, Dong Soo Yun, Giovanni Traverso, Robert Langer


This image is a juxtaposition of a molecular dynamics simulation and an electron microscopy image of the anti-cancer drug sorafenib. Sorafenib and many other drugs can spontaneously form self-assembled, intricate nanostructures, which dramatically changes how the drug will behave in experiments and, even more importantly, affect patients. By understanding and simulating the underlying mechanisms of such nanostructure assembly, we enable the use of artificial intelligence to rapidly predict the effect of millions of possible adjuvants, solvents, and chemical modifications on such aggregations. With the help of smart algorithms, we can then focus our experimental testing only on the most promising candidates and thereby quickly identify new formulations and treatments with improved delivery and efficacy.