I think this might evolve into an interesting research tool, but I don't think we will any time soon see a direct therapeutic tool that consists of creating digital twins for everybody and actually prescribing therapy based on what the digital model predicts. Since the actual biological system is incredibly complex, the digital twin would be equally complex and would need to be based on a complete understanding of the biological system. I can't imagine the massive amount of research, testing, and validation that would be required for FDA to ever approve such a system.
Also, drug companies will likely oppose, and certainly will not fund research into, such a system if they perceive that the use of digital twins will reduce the use of their drugs. Why would say Janssen want to fund a system that cuts their market share for Remicade by identifying a bunch of patients for whom Remicade won't work very well? They would likely prefer the current system where a doc tries Remicade first for a year on a new patient before moving on to a perhaps more effective option.
