max FR4 PCB temperature and lifetime

FR-4 isn't a material as such, more of a qualifier for a bunch of materials. Get the exact manufacturer and part number and look at their data. Isola, Rogers, Getek (or whatever) Nelco, etc all make "FR-4" materials and all are slightly different chemistries. Your best bet is to get details and ask the manufacturer rep.

On _any_ PC board? Including the phenolic stuff you find in clock radios that appears to be made out of old trash?

The functional failure modes of the material application are too varied and interdependant. To produce a simple temperature chart for the basic material is pretty meaningless without a specific physical or chemical property being identified. Specific failure modes are the subject of study, combined with other environmental factors, like moisture or impurities. (conductive filament formation, relative tracking index etc etc)

You might be better off concentrating on SJ reliability, or assembly reliability, which is more frequently and more readily evaluated.

Hot boards mean hot components with readily applied mtbf.

RL

Not graphs. Probably best to contact the mfg of the material. For example, page 10 on this datasheet specifies -65C to 150C max for flex but I cannot imagine that to be good for the FR-4 parts in there:

For more detailed data things become expensive, problem is you can't a priori find out how good a paper or book is unless you have quick access to a university library:

With new packages such as D2PAK and such, these guys recommend 100C but state that the FR-4 can be rated as high as 125C:

I am going to have a similar challenge soon, heat-sinking DPAKs into the board for a unit that is supposed to be used in very hot climates. That part of the design won't be fun.