sure =). I'm an engineer and not an oncologist so don't quote me on this. The main msg is that treating cancer is really, really hard; as another commenter noted biology is incredibly complicated.
multi-pronged approaches do seem to work best; some variation of that already happens. Large tumors are surgically removed, anti-angiogenesis drugs work by blocking new blood vessels from growing, I mentioned some others. The problem is that even the most aggressive combinations of treatments haven't proved to actually work in practice, and death rates for most cancers have barely budged in the last 30 yrs. A reflection of this is that people will pay $50k+ for a drug that will increase life expectancy by a few weeks (or their insurance will, that's another topic).
People like me do train their entire lives to improve this, and we are working on many new possibilities, for example tracking circulating tumor cells and figuring out what DNA mutations they have to see if you can make a drug that is targeted towards deactivating whatever fraction of them actually lead to metastatic disease. Look up the National Cancer Institute's nanotechnology in cancer program to see more of this kind of stuff.
I don't know what the "unlocalized" side effects are and why things are done that way - I will say any kind of advance in treaments is incredibly difficult to achieve because you have to convince MDs and patients to go with something new when there is a tried and tested method, and it is their lives in the balance.