Breast cancer kills because it spreads. We need to find better ways of dealing with the cells which get away, better ways of detecting breast cancer early and better ways of predicting which women are most at risk.

Will the researchers come up with goods? Karol Sikora, Professor of Clinical Oncology at Hammersmith Hospital, is optimistic. Doctors are investigating one way of giving high-dose chemotherapy to patients who have a very high risk of the disease recurring. With chemotherapy the dose has to be adjusted so it kills cancer cells without killing the patient. Stem cell rescue may be a way round the problem.

"Stem cells are precursor cells of bone marrow," says Professor Sikora. "If you give the patient a shot of growth factor, the stem cells come out into the bloodstream after five days. You can then take blood, put it through a cell separator to harvest the cells, and then put the blood back. After the chemotherapy the stem cells can be put back, too. There is evidence to suggest this does reduce the risk of recurrence."

In the years to come Professor Sikora believes it unlikely any 'magic' drugs will be discovered. "But we will develop new hormonal treatments."

New hormonal treatments, he believes, will follow once researchers develop a greater understanding of what makes breast cancer sensitive or resistant to hormonal changes.

He also believes it will be possible to make chemotherapy more selective. "One strategy would be to use gene therapy so you could get cancer cells to selectively activate a drug."

He says gene therapy is exciting. "Once we underst the molecular basis of breast cancer, if we know certain genes switch on breast cancer, you could devise drugs to switch them off.

"And even longer term, once you identify the genes which carry a risk - and more genes will be uncovered - you could begin to tackle the population and concentrate screening on women who were at high risk."

Professor Sikora is confident blood tests to pick up early changes in breast cells will be developed. "Cancer cells behave differently, they switch on a different set of genes and this shows up in the blood. Testing for this needn't be more expensive than testing haemoglobin levels. In ten years' time we won't use mammography to screen for breast cancer at all."