Gary is right on the 75%, that comes from a peer-reviewed article by Mashberg et al. Although I think that given that decade old article and what we know about the emerging new
HPV etiology, that number is due for revision. Incidence of oral cancer in the US will increase 11% this year alone. Given that tobacco use in the US has declined every year for the last ten years, there has to be a replacement cause keeping the number consistent for that period of time, and actually increasing it this year. HPV16 is going to fill that slot. And Gary is right, cancer is not contagious, but a virus that can cause it can be easily transferred from one person to another. Familial histories which might lead a person to believe in a heredity link to some disease processes, when looked at carefully, can reveal that a family that shares similar genes, also shares similar poor lifestyle choices in diet, exercise, smoking, alcohol use, and as discussed below lack of involvement with healthcare professionals.
Tobacco continues to be a factor for more than a decade after cessation, but studies into this that are finite in terms of when that damage might be mitigated do not exist. Anecdotally, treating docs that I work with frequently say to me that a person that was a multi pack a day smoker, even two decades past, is still at risk. This comes from their clinical experiences, not from published articles. The issue of alcohol and tobacco acting synergistically is well established. The alcohol thins the cell wall membranes of the mouth allowing the easier access of the combustion by-product carcinogens to the cell. Alcohol in high quantities can be a risk factor all on its own in non-smokers.
What you have to be careful of, and many studies do not consider well, are things that create a bias in the data. For instance, think about these potential skewing factors to a study. People that smoke often engage in other risk factors. This is a lifestyle issue. Public health studies show that smokers have an education disadvantage when you look at them as a whole. That doesn