You cannot prove that people who get lung cancer and smoke will not get lung cancer if the didn't smoke. A controlled experiment on that scale would require controlling for every other factor which causes lung cancer, which is basically impossible. And highly illegal.

I will admit there is a distinct correlation between the two, but it proves nothing. Perhaps a susceptibility to lung cancer causes becoming addicted to smoking. Perhaps people who aren't smokers aren't checked for lung cancer as thoroughly, whereas people who smoke are checked closer, leading to a disparity in the statistics Linolaques mentions. You cannot prove that smoking causes lung cancer.

The rest of your argument is also ####, but isn't as easily mocked, so I'll stick with this.