The initial linear part of the curve is weak inversion, then a transition to moderate inversion adn finally strong inversion. So, if we believe the simplified theory that weak inversion is also sub-threshold region, then your threshold is somewhere at the meeting point of weak inversion and moderate inversion. But it is fuzzy.
Well, if you have a different model in mind then its
parameters ought to be documented and the equation
(or equations) that contain "VT" can be back-solved in
the same manner.