MicrOscopic VisuAlization of BLood cElls for the Detection of Malaria and CD4+
- THOMA Yann
February 2014 -
Context and objectives
Early diagnostic based on Point of Care (POC) devices allow to more rapidly act on a patient, based on its illness, and they can be qualitative or quantitative. Within this context the MOVABLE project aimed at developing a quantitative POC system able to count infected and normal blood cells with optical methods. This vast goal was decomposed into two applications : quantification of CD4+ cells for the evaluation of an infection degree, and quantification of psalmodium infected cells for the detection of malaria. This initial goal was also to use smartphones for taking images, but the research allowed to conclude that such devices are not yet ready to supply images with a quality good enough for such detection.
The objective of the malaria subproject was to automatize the detection of infected blood cells by malaria and to estimate the ratio of infected cells to the total number of red blood cells. This ratio is an important indicator of how to treat the patient and current practice let a technician perform this analysis, with a very high error rate. As the evaluation of smartphones showed that the resulting images are not good enough for the size of plasmodium, the project exploited images previously acquired by the team of Dr. Guy Prod'Hom (CHUV).
The software developed at HEIG-VD is based on a supervised learning algorithm (KernelBoost, a generalization of AdaBoost). It exploits the knowledge of a pratician in order to learn what are the psalmodium infected cells so as to become more and more accurate. A graphical user interface has also been developed and offers the visualization of images and more specifically with the identification of the infected cells.
The next image shows the graphical user interface:
The final software set is a unique tool now available on github https://github.com/reds-heig/movable
with no equivalent.
Finally, the software is currently under clinical evaluation by the team of Dr. Guy Prod'Hom. It will allow to validate its potential in terms of clinical use. However it is clear that a certification process should be started as this software is a medical device.