The biggest problem is that the magnets will "quench", which is what happens when a superconducting electromagnet suddenly stops being superconducting.

There's a lot of energy stored in that magnet, and when it quenches the energy all turns to heat in a very short time. Any remaining helium will flash boil, turning into an explosive expansion of gas, and the thermal shock will seriously damage the machine

Yeah. The magnet quench flash boils a bunch of helium which is itself expensive, and presents a nice asphyxiation hazard as well. And then, assuming the quench damaged nothing, you have to set up the magnet again by getting the coils back down to superconducting temperatures... to get there, you end up boiling off a lot more helium. And then you have have to bring an engineer in to get the electrons spinning through the coil again and wait for the wobbles in the current to stabilize.

Or so I think. I work with NMR spectrometers and not MRIs, but it's essentially the same technology.