... because, by definition, such studies CANNOT prove causation.
First off, epidemiological studies CANNOT prove causation. They are simply a set of questions, such as, "What have you eaten over the past 20 years?" People cannot be accurate in remembering what they ate.
Second, this is China. Their diet over the past 20-50 years has been much different than people in other countries. And it is quite a bit different from their own diet over the past 10-20 years, as their economy has allowed them to live a more prosperous life.
The thing I keep coming back to about cancer is that the way researchers study cancer is to take healthy cells and deprive them of oxygen. Once deprived of oxygen, health cells turn cancerous (and researchers can then study them).
It seems logical, then, that cancer has something to do with cells being deprived of oxygen, which is likely not related to diet much, if at all.