A new approach to tackling the spread of malaria in sub-Saharan Africa, which combines affordable, easy-to-administer blood tests with machine learning and unbreakable encryption, has generated encouraging early results in Uganda.
Malaria is one of the world's leading causes of illness and death, affecting around 228 million people each year, more than 400,000 of them fatally. Like SARS-CoV-2, malaria can spread asymptomatically, making widespread field testing vital to containing outbreaks and saving lives.
One significant challenge for making field testing widely available is that the most common and accurate malaria blood test is based on the polymerase chain reaction (PCR) process, which requires trained staff to draw blood, and laboratory conditions to test the samples. In remote areas of sub-Saharan Africa, malaria infections often break out hundreds of miles away from trained staff and laboratory facilities, making effective infection control very difficult.
While more portable lateral-flow tests for malaria have been developed and delivered in recent years, their reliability has been questioned, with some studies suggesting they may be only 75% accurate.
Over the past few years, biomedical engineers from the University of Glasgow and the Ministry of Health in Uganda have worked together to develop a more reliable, low-cost ‘origami’ alternative to PCR and lateral-flow tests.
It uses sheets of folded wax paper to prepare patient samples for a different type of detection process – loop-mediated isothermal amplification (LAMP), which can be delivered in the field. Previous field tests in Uganda have shown the ‘origami’ test technique is 98% accurate.
A blood sample taken from a patient via fingerprick is placed on in a wax channel in the paper. The paper then is folded, directing the sample into a narrow channel and then three small chambers which the LAMP machine uses to test the samples' DNA for evidence of Plasmodium falciparum, the mosquito-borne parasitic species which causes malaria.
In a new paper published recently in Nature Electronics, the researchers describe how they have developed a secure smartphone app to pair with their ‘origami’ tests, which uses deep learning to allow more accurate diagnosis, and could facilitate better surveillance of community transmission.
The app, paired with a 3D-printed stand containing a simple heating element, controls the temperature of the ‘origami’ test, heating it in around 10 minutes to the temperature the LAMP process requires to work.
Then, the LAMP results are analysed using a cloud-based machine-learning process to ensure they are being administered correctly, enabling users of varying skill levels to conduct the test properly. A positive or negative diagnosis of the patient’s malaria infection is provided via lines on a lateral-flow strip similar to those used for home SARS-CoV-2 testing.
The patient’s results are stored securely on a blockchain-based ledger to ensure their privacy, and shared with the local authorities to allow anonymised monitoring of local infections.
The researchers’ paper, titled ‘Smartphone-based DNA malaria diagnostics using deep learning for local decision support and blockchain technology for security’, is published in Nature Electronics.
