AI-powered Microscope Detects Malaria Parasite

AI-powered Microscope Detects Malaria Parasite

AI-powewred microscope

Researchers at University College London Hospitals (UCLH) Trust have created an artificial intelligence (AI) system for identifying malaria in travelers. Annually, over 200 million people contract malaria, resulting in over half a million deaths.

The World Health Organization advocates for parasite-based diagnosis prior to commencing malaria treatment. Medical professionals utilize various diagnostic methods including traditional light microscopy, rapid diagnostic tests, and polymerase chain reaction.

Manual light microscopy is the standard for malaria diagnosis, where a specialist examines blood films under a microscope to confirm the presence of malaria parasites. However, accuracy depends on the microscopist’s expertise.

Dr. Roxanne Rees-Channer, a UCLH researcher, explained that the AI system achieved an 88% diagnostic accuracy rate compared to microscopists. This near-expert level performance in clinical settings is a significant achievement for AI algorithms targeting malaria. It suggests the system could be a valuable tool for malaria diagnosis in appropriate contexts.

malaria parasite

The researchers tested the AI and automated microscope system’s accuracy using more than 1,200 blood samples from travelers returning to the UK from malaria-prone regions. Both manual light microscopy and the AI-microscope system were used for evaluation. While 113 samples were manually diagnosed as malaria-positive, the AI system correctly identified 99 samples, yielding an 88% accuracy rate.

The fully automated malaria diagnostic system employs an automated microscopy platform to scan blood films, coupled with a malaria detection algorithm that analyzes images to identify parasites and their quantities.

Dr. Rees-Channer emphasized that even expert microscopists can make errors due to fatigue, particularly under heavy workloads. AI-based malaria diagnosis could alleviate this burden and potentially increase patient capacity.

Despite its high accuracy, the automated system did generate false positives, identifying 122 samples as positive when they were not. This could lead to unnecessary administration of anti-malarial drugs.

Dr. Rees-Channer cautioned that while the AI software shows promise, it is not yet as accurate as a skilled microscopist. The study’s results offer a positive indicator, but further evidence is needed.

In a separate development last year, researchers at the University of Queensland introduced a chemical-free, needle-free method for malaria detection via infrared light through the skin.

Click here.

Leave a Comment