Recent disappointing results have challenged the earlier results of ecologi
cal studies that launched the hypothesis of a strong relationship between d
iet and cancer. The present state of knowledge regarding the relationship b
etween cancer and diet is considered and discussed here. Steps for improvin
g the understanding of the relationship and the content of recommendations
for cancer prevention and survival are proposed, such as determining the po
ssible food effect at each step of the carcinogenesis process, considering
the dietary pattern instead of a single nutrient or food, introducing the d
iet quality index for evaluating cancer risk and developing more comprehens
ive statistical methods in nutritional epidemiology. In support of these pr
opositions, previous, recent and on-going studies are reviewed and discusse
d. A holistic model of diet is described as a conclusion.