Desenvolupament d’una base de dades d’imatges microscòpiques per millorar el diagnòstic de malària mitjançant aplicacions digitals mòbils

Abstract

Malaria is a worldwide health problem, where almost half of the world’s population is at risk of infection, and in 2015 it caused 500.000 deaths. This disease is has a global distribution, but mainly found in developing areas. One of the Sustainable Development Goals established by the WHOc for 2030 is to eradicate this disease. Thus, prevention and transmission control are the keys to getting it. This may be possible if there is a thorough and rapid diagnosis to establish the most appropriate treatment. Despite the new diagnosis tools developed in recent years, optical microscopy remains the tool for excellence in the diagnosis of malaria. It has a high sensitivity and specificity if it is performed by a skilled and trained microscopist. Nevertheless, in endemic areas of the disease, most technicians of health centers are not familiarized with malaria parasites or do not have a high experience. Therefore, an automatic diagnostic system, which is based on microscopy but does not require the presence of an expert technician, is a good solution for these areas and it can help to control the disease and its transmission. This automatic diagnosis is based on the development of a mobile application. For this, it needs an extensive database of images to train the application software and thus reduce false positives and negatives that may occur due to other forms similar to parasites. The images have been obtained from blood samples stained with Giemsa. In total 62 samples of the four species causing human malaria were examined to get the corresponding image. 1714 images were obtained between positive (1328) and negative (386). Most of the images obtained correspond to the specie Plasmodium falciparum because of its relevance in the disease. The first results of the application are encouraging because it is able to detect 80% of the parasites examined. In conclusion, an automatic diagnosis allows rapid and careful diagnosis, and also lower cost than current and rapid detection techniques: a drop of 5 to 9 € versus 30 € for the rapid screening test