It’s not just a misfolded protein, it’s a misfolded protein that catalyzes other proteins (with the same chemical structure) to misfold. So proximity of a “good” protein to a “bad” protein is sufficient to turn the “good” protein “bad”.

It’s an interesting hypothesis to consider that there could be other prion variations out there that misfold other types of proteins, but they don’t misfold the “good” versions of themselves, so they don’t self-replicate and instead just create one-off instances of a broken protein.