Doctors have dramatically increased the success rate for chemotherapy in breast cancer by targeting patients with the most sensitive tumours – using a new genetic test.

By using the test to select patients, they say chemotherapy could eradicate the tumour in 70 per cent of those treated, compared with 44 per cent now.

Reporting their findings in The Lancet Oncology, the researchers from University of Bordeaux, France, say: "This is such a big improvement that we need to do another study to make sure it is really true but if it is confirmed, it will make a big difference to the way doctors treat breast cancer."

The second study to confirm the findings is under way and the results will be known in six to nine months, said Professor Herve Bonnefoi, who heads the research team, yesterday.

The aim of the research is to develop "tailored therapy" for breast cancer, so that women aresubjected to chemotherapy, which has toxic side effects and is expensive, only if they are likely to respond.

It will not increase the overall number of patients who can be successfully treated but it will mean that chemotherapy can be targeted on those most likely to respond. This will spare women with the "wrong" kind of cancer from enduring the unpleasant effects of chemotherapy, as well as avoiding the cost.

Chemotherapy is given before or after surgery to shrink the tumour or to mop up any remaining cancer cells and has been shown to improve survival. New treatments, including drugs called taxanes, have been introduced that are more toxic and more expensive but benefit only a limited number of patients. At least three cycles are given at a cost of £2,100 a cycle.

Professor Bonnefoi said: "Treatment for breast cancer has become more efficient in the past decade but the downside is that treatments have become more aggressive and have to be taken for longer. It is to avoid this that we are doing this research. We are trying to identify groups of patients who need lighter chemotherapy or shorter treatment. We are trying to develop tailor-made treatment and this study is a good signal that we are going in the right direction."

The study involved 125 patients with oestrogen-negative breast cancer, who comprise a third of all breast cancer cases. They were randomly allocated to receive two chemotherapy regimens – one including taxanes and one without.

The women then had surgery to remove the tumour. Tests on the removed tissue revealed that the presence of two separate genetic signatures could be used to predict which chemotherapy regimen would be successful in eradicating the tumour. This was defined as the disappearance of all but a few scattered cancer cells.

"Selection of the treatment regimen with our genomic signatures has the potential to increase the pathological response rate from 44 per cent to around 70 per cent," the authors write.

Professor Bonnefoi said those patients who were not suitable for chemotherapy could be included in trials of new drugs aimed at different targets. "I have been working on this for 10 years ... So far researchers have identified two targets – hormone receptors, tackled by the drug tamoxifen, and Her2 [receptor] status, tackled by Herceptin. Our data suggest we have identified a gene signature that may be linked with a new target."