To be honest, if you're working with .qidf files .... I'm not sure. I guess it's possible but I'm not sure about the maths. Perhaps Roy or someone else could comment. Perhaps there's a trick that I've forgotten.
Because I am mostly a Piezography user I don't have .qidf files for my .quad files. Instead I use a droplet that Roy released a few years ago - QTR-Linearize-Quad.exe. You'll find it in C:\Program Files (x86)\QuadToneRIP\Eye-One. The idea is that you select both your measurement data file and the quad file in question and drag and drop them both onto this file and you get a relinearised .quad file with -lin appended to the name. If no-one offers you any suggestions about how to combine two linearisations then you could use this technique for a second round if and when you need it.
It's rare that a second round is required in my experience. If I have a difficult paper or one that I'm uncertain about, then I will try all the candidate curves - either MK or PK - and choose the one that's closest to linear to use for a relinearisation for the paper in question. In my experience what you want to try to avoid are curves with deep hooks in the shadows in their linearisation plots, since they don't relinearise all that well. In some case they won't relinearise at all - the underlying maths can't handle the shape.