Eating three or more bowls of the Japanese delicacy Miso soup every day could cut women's risk of developing breast cancer, researchers suggest.
The soup contains fermented soy paste along with other ingredients including seaweed, bean curd and vegetables.
Most people in Japan eat the soup at least once a day.
Previous studies have suggested that soya-rich foods can help cut women's risk of developing breast cancer.
Soybeans contain isoflavones, chemicals found in plants, which mimic the action of the female sex hormone oestrogen.
The study is too small to be definitive and more work is needed in this promising area
Dr Tim Key, Cancer Research UK
Although isoflavones are found in other plants, they are most concentrated in soybeans.
They are believed to prevent breast tumours developing by blocking the cancer-causing effects of oestrogen.
Researchers at Japan's National Cancer Centre monitored the eating habits of 21,852 women aged between 40 and 59 for 10 years from 1990.
They discovered women who had three or more bowls of miso soup each day reduced
their risk of getting breast cancer by about 40% to those who had only one bowl.
Those who had two bowls daily cut their risk by 26%.
Laboratory studies have shown isoflavones inhibit the development of breast cancer, but previous studies looking at the effects of eating soy-based foods have produced inconsistent results.
Seiichiro Yamamoto, head of the research group, said: "Consumption of miso soup and isoflavones ... was inversely associated with the risk of breast cancer.
"The tendency for lowered breast cancer risk (associated with other soy products) was observed, but we need to do further studies to confirm it."
He added that eating too much miso soup was not advisable because of its high salt content, so people should try to balance their soy intake by eating a number of
Dr Yamamoto said: "Very generally speaking there is a perception that the traditional Japanese diet is healthy.
"We will study what part of it had what kind of effect on people. Some were good, some were bad."
Dr Tim Key, of the Cancer Research UK Epidemiology Unit in Oxford, said: "Over the last 10 years, about a dozen studies have examined the possibility that women who eat relatively large amounts of soya may
have a lower risk for breast cancer than women who eat little or no
"The results are, overall, inconsistent. This new study from Japan does suggest that women who eat a lot of soya may enjoy some
protection against breast cancer, but the study is too small to be definitive and more work is needed in this promising area."
The research is published in the online edition of the Journal of the National Cancer Institute.