Depending on the mass they may have been able to remove it manually. A colleague used to use paperclips to study the field lines, and those had very little force.
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.