I would be very surprised if you'll be able to find a Kel-F seat outside of NASA.
You are absolutely right, Kel-f is a bit too specialized for the SCUBA industry, that was a poor example.
Furthermore, while the spontaneous ignition temps of PEEK and FKM are double that of EPDM (or Nitrile), they're only a hundred degrees apart.
Meanwhile, the theoretical max temperature due to adiabatic compression is three times ALL of the spontaneous ignition temperatures, yet we rarely see fires.
The auto ignition is only part of the story I believe. I know of one study that specifically looked at VIRPs (Valve integrated pressure regulators) in the medical industry. While not everything from it is transferable to SCUBA, a lot is. If I remember correctly, one of the key takeaways was not only the auto ignition temperature, for the reasons you mentioned. A big impact seemed to have the heat of the resulting combustion. In essence, how hot something was burning was at least equally important as to when it starts burning.
The higher combustion temperatures greatly increased the risk of the surrounding materials, like the metals catching fire.
Pair a high temperature burning EPDM O-Ring with an aluminium body for example. Even in a high O2 environment you did not create a fire just yet. Add heat from adiabatic heating, particle impact or similar and you start to burn that EPDM. As it burns so hot, it will start to burn the aluminium, consuming the whole system.
If that same system would have been pair with brass, nickel, copper, monel, etc. the EPDM would have burned just the same. But the brass or similar has a much better resistance to O2 fires than the aluminium or stainless steel did.
Sidenote: Monel burns rapidly if in a mesh form or similar, otherwise it exhibits great O2 resistance.
Conversely, you may have gotten away pairing that very same aluminium body with a FKM O-Ring, as the heat generated would be much lower.
....In other words, there are LOTS of factors involved...
I could not agree more! An oxygen fire sort of lies on a continues line. A plethora of things have to go wrong to result in an accident. That starts with the basic mechanical layout, continues to material selection and stops with the user. All must have made some poor choices for something to happen.