There's a related approach that creates in effect an "and" gate on the cells. They attack another cell only if it expresses two specific proteins on its surface. This should enable a wider range of cancer types to be targeted.
Something I've been thinking/wondering about. I'm clearly not an oncologist, so what i've been thinking about may already be standard - One of the features of cancer cells is that as cell division becomes dysregulated, karyotypes become deranged. There are missing, duplicated, truncated, hybrid chromosomes. If we could find ( or more drastic introduce ) a protein that is always expressed from each chromosome or can be induced to be expressed through a drug in every cell, how could you come up with a treatment that is deactivated by the presence of these proteins. While tumor cells in an individual are heterogeneous, if you had a treatment consisting of a compound that stimulates the production of a certain enzyme, and a drug that is deactivated by a certain enzyme, could you use this to kill all cells that lack the chromosome that code for that enzyme?
Perhaps this is already done in chemotherapy, where a compound induces it's own breakdown in healthy cells, but is this done in a chromosome by chromosome strategy?