With chemotherapy, we have literally have so much empirical evidence that the benefits far outweigh the risks. We understand the mechanism of action of chemotherapy enough to assess its risk profile.
With lobotomies, we understand enough about the brain now to know that the risk profile is extremely bad. Additionally, empirical evidence has shown that outcomes are pretty bad anyway! We have better tools and ways to understand the brain now to not give people a brain-destroying "treatment" that is basically a random, low chance of success. Would you reasonably prescribe getting shot in the head as a treatment if it had worked on one person?