'If you identify these latent DCIS cancers and biopsy them, you have traumatised the area,' he says. 'You immediately trigger the natural healing mechanisms which involve angiogenesis. So the biopsy could be considered as an angiogenic switch. You take a latent cancer that would never hurt a woman, biopsy it, turn on the angiogenic switch, and it ceases to be latent - it becomes an aggressive disease.'

The evidence appears to bear Baum out. Three separate studies show that breast cancer 'almost always occurs at the original biopsy site' (Cancer, 1986; 57: 197-208; Cancer, 1989; 63: 618-24; Br J Cancer, 1990; 61: 869-72).

Poor track record The third arm of breast-cancer treatment sometimes used for DCIS is chemotherapy, where powerful cell-destroying drugs are infused into the bloodstream. In this case, Swedish cancer statisticians did the massive job of marshalling together all the evidence from over 200 separate trials of the world’s major chemotherapy drugs.

The most positive news they could find was that chemotherapy reduces death rates in some breast cancer cases by, at best, 12 per cent. But, for most women with the disease, chemotherapy fares very much worse.

The open secret in medicine is that the conventional treatments for breast cancer, as with DCIS itself, is still a largely a mystery. Despite all the fanfare about winning the war on cancer, the best statistics still come from alternative practitioners.

Tony Edwards