Plasmo detect: malaria diagnosis in a click, or almost...
After winning the Montpellier Innovation Booster in 2024, researcher Grégoire Pasquier is about to finalize a new application capable of accurately detecting, from a blood smear, the type and degree of severity of malaria... Thanks to a smartphone.
A university hospital practitioner and researcher at the Migevec laboratory, Grégoire Pasquier came up with the concept using the Pl@ntNet app, which identifies plants from a photo catalog. Since 2021, in parallel with a thesis on the involvement of Schizophyllum commune in human pathology, he has therefore been working on Plasmo detect, an app capable of detecting malaria as early as possible and in all its forms via a smartphone adapted to the microscope. "Currently, the reference method for diagnosing malaria is microscopy, using what we call a blood smear. But it takes a great deal of expertise to determine which of the five existing parasite species is the cause of the disease, and at what stage. With Plasmo detect, we hope to automate this process, using artificial intelligence to classify images", explains Grégoire Pasquier.
30,000 blood smear images
To begin with, he and his computer scientist photographed and annotated almost 30,000 images of blood smears from people infected with different species of Plasmodium. This catalog is essential not only to identify the species, but also to establish the parasitemia, which corresponds to the percentage of infected red blood cells. " Above 4%, it's a criterion of severity", adds the researcher. "For the moment, we are able to define the correct species in 80% of cases. According to the criteria of the World Health Organization (WHO) quality manual, we are at level 2. To reach level 1, we need to reach 90%. We are therefore in the process of improving the performance of the algorithm and the application", sums up Grégoire Pasquier.
To achieve this, the team is currently expanding the photographic database, and annotating the images already available to include the parasitic stages detected on each blood smear. " We can go into a little more detail than has been done so far", says the scientist, who is hoping for better results in the short term.
Large-scale testing
Winner of the Booster innovation Montpellier (BIM), a support scheme run by the Pôle universitaire d'innovation (PUI), Plasmo detect has been awarded a €30,000 grant to continue the effort and achieve a reliable, usable application by 2027. Next year, a Master 2 student specializing in artificial intelligence will join the team. Grégoire Pasquier also intends to invest part of this sum in a more powerful computer and a precision camera.
Thanks to the BIM workshops dedicated to the creation of start-ups, the project owner was also able to learn about entrepreneurship. " This opens up interesting prospects", he stresses. In the coming weeks, the application could already be in the hands of a telemedicine company, capable of testing it on a large scale. "We've just signed a confidentiality agreement, so it's well on the way."
260 million people affected
On November 25, the founder of the Plasmo detect project will also represent the PUI at the Starthèse ceremony in Lyon. It's an opportunity for Grégoire Pasquier to familiarize himself once again with entrepreneurship, and for the University of Montpellier to showcase innovative projects that address potentially serious and pressing health issues.
Malaria affects some 260 million people worldwide every year, and kills almost 600,000. Most of these cases occur in sub-Saharan Africa, "where few people are trained to read blood smears", explains the researcher. By 2027, Plasmo detect could facilitate diagnosis, guide treatment much more quickly and perhaps even save lives...