Scientists have been able to "turn off" genetic switches and stop breast cancer developing.
Human breast cancers were studied in mice
The research in mice, carried out by a team at the University of California, could lead to the design of new drugs to target the disease.
They suggested it may be possible to reverse tumour growth or prevent breast cancer altogether.
The research was presented to the International Association for Breast Cancer Research.
It focussed on a gene called beta-1 integrin which helps cells stick to a framework to form tissue.
In healthy breast tissue, it also regulates growth and survival.
But previous research has linked the gene to cell migration and the development of breast cancer metastases - the appearance of secondary tumours in other parts of the body.
Scientists genetically modified mice so they had a particularly aggressive form of human breast cancer.
But when beta-1 integrin was turned off, a large number of tumours - including those which were most aggressive and advanced - regressed to the point where they could no longer be detected on scans.
The team say their findings show the gene plays a crucial role in the development of breast cancers.
Although many of the tumours went into complete remission, a substantial number of cancers recurred over the next year.
However, the researchers say this is also promising, because it replicates the natural pattern of many women's breast cancers where a small number of cells remain and can cause cancer some years later.
'Making cancers disappear'
William Muller, professor of biochemistry at McGill University in Montreal, who worked on the study, said: "This shows that it is absolutely essential to have the beta-1 integrin gene present in order for mammary gland tumours to develop.
"We now have a good target for biological drug development, and the challenge now is to develop an agent that can block its activity."
Fellow researcher Lewis Chodash of the University of Pennsylvania, added: "We're extremely encouraged that we have been able to demonstrate in laboratory animals that we can make mammary cancers essentially disappear by reversing just one mutation."
Charles Streuli, professor of cell biology at the University of Manchester, told BBC News Online the research was interesting, but said he was cautious because it had not yet been published in a peer-reviewed journal.
"We know that integrins are essential for many cell types to respond to growth and survival signals.
"It will be important to find out if the lack of beta-1 integrin causes the oncogene-expressing tumour cells [which cause healthy cells to change into cancer cells] to all die or whether they remain dormant and are unable to respond to proliferative signals."
Dr Michelle Barclay of Breakthrough Breast Cancer said: "The concept of altering a gene to prevent or reverse breast cancer is very interesting.
"However this research has only been demonstrated in animal models and, although exciting, we do not yet know how this would translate to humans and further research is needed."