King's College London

The risk score - which is based on a routine test called flow cytometry that has been validated for clinical use - allows doctors to quickly and cheaply assess the chances of death or progression to Acute Myeloid Leukaemia (AML) in patients with myelodysplastic neoplasms (MDS).

Myelodysplastic neoplasms (also known as myelodysplastic syndromes) are bone marrow cancers that occur most often in older people. They arise when blood stem cells gain genetic and epigenetic - when behaviours or environment cause changes to how the genes work - that impair blood production, causing anaemia, infections, bleeding and sometimes progression to acute leukaemia.

Doctors already use prognostic systems to assess risk of disease progression. However, these advanced tests, such as molecular testing, can take days to complete, are expensive, and are not available in many healthcare centres. They are also limited in measuring bone marrow cell development and immune cell patterns that can help predict disease progression.

Shahram Kordasti, Professor of Systems Cancer Immunology, and his team of researchers, in collaboration with Professor Mieke Van Hemelrijck, wanted to investigate whether data gathered using a technique called flow cytometry – commonly used in clinical labs to diagnose people with MDS – could also be utilised by doctors to work out a patient’s chances of survival. Flow cytometry is a quick, relatively cheap and widely available technique that involves using a laser to measure proteins on the surface of cells to identify cancerous cells.

Published in the journal HemaSphere, the study describes how the researchers looked at flow cytometry data from more than 500 MDS patients in healthcare centres at Dresden University in Germany, Amsterdam University in the Netherlands, and sites across the UK. They used a computational model with machine learning to analyse the data.

They identified 55 different flow cytometry measurements that gave relevant insight into disease progression. Unlike traditional flow cytometry scores, which focus only on abnormal blood stem cells, they also incorporated variables from the immune system. This reflects an increasingly accepted understanding that MDS is not just a disease of defective blood stem cells, but it is a disease where abnormal stem cells interact constantly with immune cells in the bone marrow.

The researchers built four models which they tested against the 55 parameters. They identified the best performing model, and then whittled down the number of parameters required for the model to perform effectively to just six.

The final score strongly predicted survival in patients and performed better than previous flow cytometry scores. Strikingly, the score identified different types of MDS – a varied disease where patients can have very different outcomes. By incorporating elements of the immune system, the score was able to more accurately pin down the subtle variances in MDS between patients, which could support clinicians with risk stratification.

By successfully validating the risk score, the study opens the door to quick, cheap and widely available prognostic testing for MDS patients. While the current study looked at the effectiveness of using flow cytometry in bone marrow, Professor Kordasti explained that future studies will be carried out to see whether it can be performed on peripheral blood samples instead of bone marrow, which would be less invasive for patients.

Read the full study here.