http://www.sciencedaily.com/releases/2007/02/070218194439.htm
"When cells become cancerous, they also become 100 times more likely to genetically mutate than regular cells, researchers have found. The findings may explain why cells in a tumor have so many genetic mutations, but could also be bad news for cancer treatments that target a particular gene controlling cancer malignancy."
I think Blaman was reading your post as if you meant that they mutate in a similar way to bacteria.
Cancer cells can mutate, but they don't spread from one person to another.
And since the mutations are entirely random, the mutation in one cancer is not likely to be there in another cancer.
So, although, as stated in that article, some cancers may be resistant to certain chaemotherapies, not all cancers would be from then on.
It would just simply make it even more harder to treat cancer, because there's an even bigger chance of variation in the situation in one individual case.
It wouldn't be in the same way that say, a form of Tuberculosis is killed with one drug, but has a mutation making it immune, and then all TB is immune. This only happens because the mutated and immune TB is capable of spreading.
Cancer can't do that.