Paris Law

In a CT fracture characterization test, Since we neglect data beyond a certain crack length (due to onset of considerable plasticity),

1. Should we also use Paris law to compute the remaining life only until the predominantly elastic regime (The limiting K value)? Or can we use it until the fracture toughness is reached?
2. Though I have a limiting K value, the crack growth curve looks log linear almost until the final stages of growth. So can I include the entire curve? What are the consequences if I use all the data points?
3. Why shouldn't I optimize the parameters so that the calculation also includes the inelastic regime? (I understand that SIF is an LEFM parameter). But since we are only discussing about an empirical approach