Researchers develop smartphone app to support malaria diagnosis
August 19, 2015
By Diego Freire
Agência FAPESP – To make malaria diagnosis faster and easier in remote parts of the world, researchers at the US National Library of Medicine (NLM) are developing a smartphone app that uses an automated system to detect and count malaria parasites.
The technology was presented at the 28th IEEE International Symposium on Computer-Based Medical Systems (CBMS 2015), held in June with FAPESP’s support at the University of São Paulo’s Mathematics & Computer Science Institute (ICMC-USP) in São Carlos and the same university’s Ribeirão Preto Medical School (FMRP-USP) in Brazil.
“Malaria is a curable disease, but inadequate diagnosis and resistance to emerging drugs are still barriers to a reduction in mortality. It’s important to develop vaccines and control the mosquito that transmits malaria, among other measures, but the development of a reliable rapid diagnostic test is one of the most promising ways to combat the disease,” said Sameer Antani, a scientist with NLM, which is part of the National Institutes of Health (NIH).
The project requires interaction among health professionals, computer scientists and engineers, he explained.
“A dialogue between the clinical requirements of healthcare for people in rural areas and computer engineering led to this new approach to combating malaria, developed to meet the urgent demand for control of the disease but adapted to the reality of work in areas with highly specific needs,” Antani said.
Agma Traina, a professor at ICMC-USP and a member of the CBMS Steering Committee, stressed that this interaction has given rise to contributions to medical practice, but more dialogue is needed.
“The CBMS 2015 program was designed to cover different ways of integrating computers and medicine for the benefit of people generally. These two areas already talk to each other a great deal, as evidenced by the many research findings and technologies presented during the symposium. However, they need to interact even more. On one hand, computer scientists and mathematicians need to know more about and understand the needs of health professionals in diagnosing and treating diseases. On the other, physicians and researchers in the area need to understand the potential assistance offered by computer engineering,” Traina said.
Antani said the project to automate the detection and counting of malaria parasites was supported by the US Department of Health & Human Services and a research partnership between Thailan’s Mahidol University and Oxford University (UK), which together set up the Mahidol Oxford Tropical Medicine Research Unit.
“Watch it, parasite!”
Malaria is caused by Plasmodium, a single-cell protozoan parasite transmitted through the bites of female Anopheles mosquitoes. It is diagnosed with the help of light microscopy to identify the parasites in blood samples.
According to Antani, some 170 million blood films are examined every year for malaria, and in most cases the number of parasites is counted manually to determine whether a case is severe or uncomplicated in the classification recommended by the World Health organization (WHO).
“Accurate parasite counts are essential to diagnosing malaria correctly and successful treatment, influencing drug effectiveness, for example. However, microscopic diagnostics isn’t standardized and depends heavily on the experience and skill of the microscopist,” he said.
This leads to incorrect diagnostic decisions in the field and hinders the control of malaria in low-resource settings where the incidence of the disease is high, Antani explained.
“For example, false negatives aren’t just a health hazard for infected individuals but also entail a second consultation, lost days of work and unnecessary expense.
In the case of false positives, misdiagnosis entails unnecessary use of anti-malaria drugs and suffering from their potential side-effects, such as nausea, abdominal pain, diarrhea, and sometimes severe complications,” he said.
The NLM researchers’ project, “Watch it, Parasite!”, is developing a fully automated system for parasite detection and counting in blood film, a thin smear of blood spread on a microscope slide and stained to enable various blood cells to be examined.
Antani explained that automatic parasite counting offers several advantages compared to manual counting.
“Automation provides more reliable and standardized interpretation, reduces diagnostic costs, and enables more patients to be processed in less time, making health workers more productive in areas of high malaria incidence,” he said.
The system uses machine learning methods derived from artificial intelligence in which algorithms and other techniques enable computer-based devices to learn to recognize certain patterns and make data-driven predictions to improve their performance in specific tasks.
The system first “learns” the typical patterns produced by parasites in manual training images. It then uses digital images acquired from blood films on standard light microscopy equipment to detect whether parasites are present, counts them, and discriminates between infected and uninfected cells.
Adapted for use as a smartphone app, the technology is inexpensive, highly portable, and can be used to analyze blood film images in the field. For Antani, the app will benefit the populations of several countries that still suffer from malaria, including Brazil.
“Malaria kills more than 600,000 people per year, mostly in Africa, where a child dies every minute from the disease,” he said. “Many survive but have to live with neurological disabilities. Malaria is actually a global problem, though. There are some 200 million cases worldwide, including in Brazil. The development of a low-cost, high-precision, portable technology to diagnose malaria offers great potential to assist in fighting the disease and contributes to efforts to wipe it out completely.”
According to the WHO, Brazil has reduced the number of infections by 75% since 2000, but the incidence of malaria is still high, especially in the Amazon. The number of cases diagnosed throughout Brazil in 2013 reached 177,767, and 41 deaths were recorded.
Antani also spoke at the CBMS 2015 symposium on the development of advanced algorithms for automated screening of digital chest X-ray images for pulmonary abnormalities with a special focus on tuberculosis. The goal is to simplify the procedure and take it to remote areas of the world where there are shortages of radiologists.
The algorithms can be used in a system installed in mobile X-ray trucks capable of traveling through rural areas. The researchers have a pilot project running in Kenya and other parts of Africa where there are many cases of opportunistic HIV and TB co-infections.
“When the system receives a chest X-ray, it evaluates segments of lung regions using a graph cuts optimization approach that combines X-ray information with personalized lung atlases derived from models used to train the system in a set of texture and shape characteristics. All this enables X-rays to be classified as normal or abnormal using a binary classifier,” Antani explained.
According to Antani, the performance of the proposed computer-assisted diagnostic system comes close to that of human experts and should help combat tuberculosis in remote areas.
“A comparison of the system’s performance with that of radiologists shows 84% accuracy, which represents substantial hope for populations that lack easy access to these professionals and suffer from tuberculosis. When the disease is undiagnosed and patients are therefore untreated, death rates are high,” he said.
Agência FAPESP licenses news reports under Creative Commons license CC-BY-NC-ND so that they can be republished free of charge and in a straightforward manner by other digital media or by print media. The name of the author or reporter (when applied) must be cited, as must the source (Agência FAPESP). Using the button HTML below ensures compliance with the rules described in Agência FAPESP’s Digital Content Republication Policy.