Not sure about nanite’s changes;
If you get false positives the most possible answer is that the physmat has the same uv issues people get on materials when you apply nanite to a landscape.
Basically it tiles the texture once on every component from what was shared on the forums.
If you really want to sove the value yourself in your own function (or whatever, evene in BP)… You also have to assume that you need a visual debug of the UVs being applied to the physmat mask avaliable.
Make some sort of debug view that swaps out the shader (material instance) to one using the same UVs applied to the physmat during the calculations you already have.
If it looks the same, then good, one less thing to worry about…
Also, prior to nanite, the physmat mask retuned correct results - at least on meshes back when I tested it…