That would be hard to do. If you break down all the static and dynamic air forces on a turbine, they sum to a large AFT component. In other words, the shaft is in tension, and if it ever fails, the turbine rotor shifts aft.
In a multistage LPT, it's generally possible to contrive a system in which the rotor and stator airfoils will clash and thus destroy the driving torque; ergo the rotor grinds safely to a stop.
In the case of the HPT, as soon as the shaft separates, there is no torque driving the HPC, so it soon quits supplying air to the cycle, and the HPT coasts down.
The IPT is another animal. If the core HPC/HPT keeps running, why wouldn't the IPT overspeed? I don't think I'd want to be anywhere in the same county!
