Accurate and fast monitoring of leukemia patients

CHALLENGE
Multiparameter flow cytometry is the golden standard in analysing cells in the blood. This technology enables rapid measurement of multiple physical and chemical characteristics of individual cells.
Examples of clinical applications of this analysis are the measurement of Minimal Residual Disease (MRD) in leukemia and inflammatory diseases. The complex data produced is still analysed by hand, which makes interpretation of the information complicated for technicians, clinicians and researchers. Pattern recognition requires experienced people and interpretation is time-consuming and human error sensitive.

OPPORTUNITY
Each year more than 3000 leukemia patients in the Netherlands need MRD monitoring. This monitoring is done before and after the treatment by using flow cytometry. Especially after the treatment, it is extremely important to know how many leukemia cells are still present in the blood and bone marrow, as this helps to diagnose a possible relapse. Since
the relevant cells are in low numbers, pinpointing the malignant cells requires a lot of training and is therefore a time consuming task.

SOLUTION
Researchers and clinicians from UMC Utrecht and Radboud University have addressed this problem by developing an algorithm that uses raw data from a flow cytometry machine and analyses each cell which differs compared to the patient’s prior analysis. A case study performed by the researchers shows that the algorithm is able to detect MRD more precisely (processing 8M cells in one go) and much faster (2 minutes versus >15 minutes expert analysis). Moreover, the results are presented to the technician in a simplified 2D image, showing all the cell populations. Interpretation can thus be done by less experienced people, in less time and more accurate. More importantly, the increased precision means that the amount of false negatives can be reduced.

STATUS
The source code of the algorithm is finished and already tested in patient samples. The Graphical User Interface is in the process of being developed. The clinical validation of this decision support algorithm will follow.