This approach assumes that it would be easy to “make” an animal model. Unfortunately, this central assumption is wrong. If we had that model, we would be half there in terms of understanding this problem.
Specifically, what we need is a disease model of PFS. We know from humans that only a small percentage of those that consume finasteride get PFS. There are some predisposing factors that make us different from those that can happily use finasteride and not get hurt. Those are either genetic and/or epigenetic. Some get PFS after only a a few pills, others can take it for a decade before PFS strikes. There again, we need to understand the underlying difference between those extremes.
Only with that knowledge can we hope to engineer a disease model that will get “real” PFS when fed finasteride for a short while.
A good disease (animal) model is the holy grail of medical research. To reach that, we need to understand what the genetic driving factors are behind PFS. A good start to this would be if more patients would participate in our 23andMe project.