There is another aspect to this, and that is the rate of heat transfer from the oil fire. Lightweight components heat up very quickly, but massive structures much more slowly. The IP rotor disc is a composite of massive areas (the disc bore) and much thinner areas (the drive arm). Therefore I suspect the drive arm overheated much more quickly than the disc proper.
This reinforces the scenario in which the drive arm failed first (after the oil fire escaped the sump). This released the disc assembly, blades included, and since it was subject to a big gas pressure load on the front face, it was immediately blown aft until it was seated on some static parts, establishing a "false bearing". It's still turning 7000 rpm, and the false bearing is quickly friction-heated producing liquid metal, which has some lubricating properties.
And it's still being driven by core turbine exhaust airflow, and it has lost the normal torque load of the IP compressor.
Guaranteed overspeed, as previously stated.