The technique harnesses bone marrow immune response
Some leukaemia patients who do not respond to conventional treatments may benefit from bone marrow transplants selected to target the cancer directly.
The technique, pioneered in Italy, uses transplants from family members who are not a perfect match.
"Natural killer" cells in the new bone marrow then attack the leukaemia.
Survival rates rose after the treatment, but UK experts say these need to be reproduced on a bigger scale and in different types of the disease.
Bone marrow transplantation has been around for half a century, and, in leukaemia, is traditionally given to patients to replace bone marrow destroyed by powerful anti-cancer treatments.
One of the biggest problems in bone marrow transplantation is when the immune cells in the donated bone marrow "reject" their new host.
They launch attacks which can prove fatal in the worst cases, a condition called "graft versus host disease".
To avoid this, patients and donors are carefully screened to produce as perfect a match as possible.
However, some of the most recent research in bone marrow transplantation focuses on using the disease-fighting qualities of bone marrow to destroy cancer cells, either to prevent the disease coming back or to tackle it head-on, even where it is resistant to drugs and other treatments.
Testing for killers
Professor Andrea Velardi, from the University of Perugia in Italy, has been funded by the Anthony Nolan Trust to look for ways to harness these qualities without raising the risk of graft versus host disease.
He has been using donors from the patient's own family who are only a partial match - sharing only 50% of their genetic material.
He found that in some cases, immune cells called "natural killer" cells were active in the donor bone marrow after transplantation, and could launch an effective attack on the leukaemia cells, and that he could predict in advance, using tests, how effective that would be.
In a small group of patients with acute myeloid leukaemia - which makes up approximately a third of all cases - survival rates improved when this kind of transplant was given with the patient already "in remission" - cleared of the disease by chemotherapy.
However, it significantly increased survival - from 2% to 30%, among those patients whose disease had not responded fully to treatment prior to the transplant.
He said: "For patients considered ineligible because of chemo-resistant leukaemia, this is a potentially life-saving advancement.
"It's likely to make enormous changes in the practice of transplantation worldwide."
He said that there were some patients - approximately a third of the population - who could not benefit from the technique, because of a physical makeup which meant that their leukaemia would not respond to the natural killer cells.
In addition, he said, the chances of finding a donor with suitable cells was approximately 50% in the average family size.
Other experts welcomed the research, although were cautious about the implications of the findings.
A spokesman for the Leukaemia Research Fund said that, while exciting, more evidence would be needed that the treatment was effective in bigger trials against all types of leukaemia.
In particular, it should be tested on older patients, who make up a large proportion of those whose disease had failed to respond to conventional treatment.
However, he said: "I don't want to underestimate the importance of this work.
"If you are in the position of a patient who has not responded to treatment, the choice between a 2% chance of survival and a 30% chance is not a difficult one."
Professor Velardi's work is due to be presented at a bone marrow transplantation research conference in London next week.