Detecting breast cancer from changes in the structure of hair could cut down false alarms from mammograms.

The test bombards strands of hair with X-rays from a synchrotron particle accelerator. In hair from healthy people, the pattern produced by the X-rays is a series of arcs, while in people with breast cancer a distinctive ring is superimposed on top of the arcs.

Though the test first showed promise in 1999 (Nature, vol 398, p 33), other researchers failed to repeat the results. Now Peter French and Gary Corino of the company Fermiscan in Sydney, Australia, say this is because hair is damaged by products such as dyes, by stretching as it is held in the X-ray beam, or because it is wrongly aligned in the beam. Their device holds hair in the correct position, does not stretch it and tests only the untreated hair a few millimetres from the scalp.

So far, they have correctly identified 12 out of 15 women already diagnosed with breast cancer (International Journal of Cancer, vol 122, p 847). Although the test generates a high rate of false positives, incorrectly identifying 2 out of every 10 women as having breast cancer, Fermiscan says it is useful. In an ongoing trial, the people it has flagged as false positives tend not to be the same ones incorrectly flagged by mammograms, suggesting that together the two tests could reduce the overall false positive rate.

No one is sure what causes the signature ring. One suggestion is that proteins released by cancer cells alter the hair follicle to produce hair with a different structure.

Cancer - Learn more about one of the world’s biggest killers in our comprehensive special report.