I think for there to have been a double failure so close in time, it's necessary to look for areas where parameters would be fairly close in range. The Spar valves shut at the same time could initiate cavitation but would make both environments (fuel lines) dependent on the "same" "set" of conditions. The volume of the lines, though similar, would have been susceptible to varying results from suction due to connections, tubing strength, indeed even the "position" of the valves themselves.
When I think of cavitation, I visualize a smallish tightly sealed space subject to great energy at one end and collapse close by (shear). I'm looking for a block close by the HP's, one that reacted to parameters present in spite of all engineered precautions and anticipated challenges.
On the other hand, the delay of 7 seconds could represent the "differences" in systems' reactions to the fault: partial "collapse" of the piping to fixed values, time for two spar valves to close completely having lagged, run up of the lobes of the HP's to max, all other junctions to reach maximum deflection due to demand, etc.
Question: Could cavitation itself have been responsible for loss of commanded power and failure to respool? Only in the sense that it represents the "onset" of fault, the plug would have been responsible for the transition from "cavitation as evidence" to starvation and ongoing low power. 1.02 and 1.03 EPR mean not much is happening in the thrust department.
FS: I think the empty center tank would have performed more as an insulator than a chiller. Think Thermos.