The problem is that normal MRI math tries its damnedest to avoid actually solving the right equations. Instead, with a flat enough field, you can assume linearity and just FFT the thing. They'll physically place bits of metal and magnets at various places on the big magnet to calibrate and better adjust the field to being approximately linear. A hunk of metal bigger than a shim sounds like it would mess with that.

Shim it with some more pens?

It would come off ok, this happens from time to time, but that facility needs to lift its game.

Peripheral staff (nurses, anaesthetic techs etc) visiting are the usual source of these accidents.