The findings, from a study of nearly 335,000 European women, do not prove that French fries, sweets and white bread contribute to breast cancer.
However, they do hint at a potential factor in a little understood form of breast cancer, according to a researcher not involved in the work.
Specifically, the study found a connection between high "glycemic load" and breast cancers that lack receptors for the female sex hormone estrogen.
A high glycemic load essentially means a diet heavy in foods that cause a rapid spike in blood sugar. The usual culprits include processed foods made from white flour, potatoes and sweets. A sweet, juicy piece of fruit can also raise blood sugar quickly. However, since fruits are low in calories, they do not contribute as much to a diet's glycemic load.
So-called estrogen receptor (ER)-negative tumors account for about one-quarter of breast cancers. They typically have a poorer prognosis than ER-positive cancers because they tend to grow faster and are not sensitive to hormone-based therapies.
In this study, postmenopausal women whose diets were very high in glycemic load had a 36% higher risk of ER-negative breast cancer, compared with women whose diets had the lightest load.
In general, a diet with a high glycemic load is not a particularly healthy one, noted Christina Clarke, a research scientist at the Cancer Prevention Institute of California in Fremont, and a consulting assistant professor at Stanford University.
"These types of diets have been associated with many negative health outcomes," said Clarke, who was not involved in the study.
So although the current findings do not prove cause-and-effect, they can give women another reason to make healthier diet choices, according to Clarke.
Lead researcher Isabelle Romieu, of the International Agency for Research on Cancer in Lyon, France, did not respond to an email request for an interview.
From a scientific standpoint, Clarke said the results are interesting because so little is known about what causes ER-negative breast cancers. Most breast tumors - the ER-positive ones - have their growth fueled by estrogen.
"We really don't know anything about what causes (ER-negative) tumors," Clarke said. "This study gives us a really important clue for future research."
Diets with a high glycemic load are associated with a bigger secretion of insulin, a hormone that regulates blood sugar. High insulin levels, in turn, have been linked to certain cancers, possibly because insulin helps tumors grow.
The current findings hint at a role for "insulin pathways" in ER-negative breast cancer, according to Clarke. "But there's definitely more work that needs to be done," she said.
The findings, which appear in the American Journal of Clinical Nutrition, are based on a long-running European study on nutrition factors and cancer risk.
Of nearly 335,000 women in the study, 11,576 developed breast cancer over a dozen years. Overall, there was no link between breast cancer risk and glycemic load - estimated from diet questionnaires the women completed at the study's start.
But the picture changed when the researchers focused on postmenopausal women with ER-negative cancer. Among women in the top 20 percent for glycemic load, there were 158 cases of breast cancer, versus 111 cases in the bottom 20 percent.
When breast tumors also lacked receptors for the hormone progesterone, the gap was a bit more pronounced.
Still, the numbers "weren't huge," Clarke noted. And there are many other factors that could be different between those groups of women, although the study did account for some of them, including weight, exercise habits, calorie intake and smoking.
Clarke pointed out that there is no single factor in any woman's risk of breast cancer. But, she said, the findings offer more incentive to eat a balanced diet that limits refined carbs in favor of healthier fare - like lean protein, vegetables, "good" fats and high-fiber grains.
"Really, you want to avoid these (high glycemic load) diets anyway," Clarke said.