iMAGING, an app to diagnose malaria using artificial intelligence

BEACON: Design and construction of a proof of concept of LIDAR for its use in autonomous underwater vehicles
January 30, 2024
N2B2: A system to optimise heating and cooling with predictive control
January 22, 2024

25/01/2024

A multidisciplinary team whose participants include the Microbiology Service of the Vall d’Hebrón University Hospital, the Microbiology research group of the Vall d’Hebron Research Institute (VHIR), the Computational Biology and Complex Systems Group (BIOCOM-SC UPC), the Image and Video Processing Group (GPI) and the Database Technologies and Information Management Group (DTMI) at the Universitat Politècnica de Catalunya - BarcelonaTech (UPC), and Fundación Probitas, have presented a new diagnostic method for malaria based on artificial intelligence.


Malaria is an infectious disease that is transmitted by mosquito bites and caused by parasites of the Plasmodium genus. The World Health Organization (WHO) calculated that in 2022 there were 249 million cases worldwide, 93% of which occurred in the African region, which also accounted for 95% of malaria deaths. The same report warns that climate change and globalisation are causing an expansion of the mosquito to new areas that have little preparation and resources to address this problem. Currently, the reference method for diagnosing malaria is visualisation of the parasites by an expert, using an optical microscope and blood samples. This is a manual, long, repetitive procedure that, added to the lack of laboratory technicians and instruments, leads to a high level of underdiagnosis. To date, any steps to automate the process has increased the costs exponentially, which has made it prohibitive in countries with few health resources.

This is a system created from artificial intelligence that combines a low-cost mobile application with an automated microscope, with the idea that this will be a useful and effective method in countries with few resources, where this disease is endemic. The results of the first iMAGING prototype have been published in the journal Frontiers in Microbiology. The prototype has been trained with over 2,500 images and has shown a reliability of over 96% in the laboratory with high-density samples and 94% with low-density samples. False positives and negatives were below 5% in all cases. 

The proposed solution, iMAGING, is a mobile application (available for Android) that uses artificial intelligence to process digital images of blood samples to determine whether or not there is an infection. In positive cases, it also determines the density and the stage of the parasitic infection. To capture the images, an automated microscope has been created using a normal optical microscope and parts created using 3D printing, which has reduced their cost.

The app is connected via Bluetooth to the microscope and controls its movements and focus to automatically analyse the sample and achieve the images that are needed for the diagnosis. Technical staff only need to prepare the samples, which reduces their workload greatly and the possibility of errors.

It is planned to continue to train the artificial intelligence to introduce improvements in other areas. One of these is to be able to differentiate between the five species of parasites that cause the disease. This will enable the treatment to be personalised to a greater extent, which will improve its effectiveness. 

 

Budget and funding

The project is part of science and technology work for human development promoted by the UPC’s Centre for Development Cooperation (CCD). The project is supported by the WHO as part of its initiative for digital imaging diagnosis of blood parasites in low and medium-income countries. It has a budget of €200,000 contributed by the three organisations that are involved: the Fundación Probitas, Vall d’Hebrón Research Institute and the UPC. Currently, funding is being sought to expand the project to diagnose other diseases such as schistosomiasis, filariasis, intestinal parasites or Chagas disease.


 
 

Related Projects