Search Menu

Machine learning help predict leukaemia recurrence after stem cell transplant

9/11/2018 10:25

FRC Blood Service has studied the relevance of hereditary variation among stem cell transplant patients in terms of the treatment outcome. The results indicate that genome data could be used more extensively to assess risks in stem cell transplant treatment.

​Blood stem cell transplants are used as a last-resource treatment for quickly advancing haematological cancers against which any cancer medication is inefficient. In allogeneic stem cell transplantation treatment, the donor’s bone marrow replaces the patient’s own bone marrow and immune system, destroying any malignant tumour cells in the process.

Although response to treatment is steadily improving, stem cell transplantation is still a demanding and dangerous procedure in which life-threatening complications such as graft-versus-host effect or relapse are common. About a third of the patients relapse despite a seemingly favourable transplantation, and currently there are no good tools based on genomics for predicting relapse.

Basis for more individualised treatment

The Blood Service, where all cross-matching between patients and donors is carried out, has now been studying the correlation between the graft recipient’s genes and the relapse risk.

The study involved more than 500 patients who had received stem cell transplants in Finland. The variation of their genome was determined by means of sequencing and a DNA chip. The correlation between variation and relapse was studied after the treatment by means of calculations and statistical methods, attempting to take into account, for example, the effect of variation caused by age or graft type on the outcome. The patients received their grafts from their siblings. The machine learning model was created for a sample of 150 patients and its reliability was re-assessed in two independent patient series.

The results suggest that there is a statistical correlation between certain genes and cancer relapse. Future research into genetic factors may help to understand the disease mechanisms and a more accurate assessment of the patient’s risks.

- Genomics information can help us develop more accurate tests for matching the stem cell donor and the recipient, says Director of Research Jukka Partanen at Finnish Red Cross Blood Service.

The research project is funded by the Academy of Finland, and these first results have been published in Leukemia, a leading journal on the subject: J Ritari et al.: Genomic prediction of relapse in recipients of allogeneic haematopoietic stem cell transplantation

More information:

Jukka Partanen, Director of Research, Finnish Red Cross Blood Service, tel. 029 300 1759

Jarmo Ritari, Project Researcher, Finnish Red Cross Blood Service, tel. 029 300 1806