The assumption that "it does make a difference between death an survival" in not necessarily born out by the data.

There has been a big push back recently against using an improvement in 5 year survival rate as an indication of success in the war on cancer. For example, if a new diagnostic test lets you find out you have cancer earlier, but you still end up dying at the same age you would anyway, the 5 year survival rate would go up, with no real benefit to you. In fact, there could be lots of real negatives, like anxiety and an increase in other tests and procedures that don't actually improve your lifespan or quality of life.