Health & Wellness
PathoScan, an offline AI microscope for malaria diagnosis
Kamugisha Edson, a student at Kyambogo University, is building PathoScan, an offline AI-driven digital microscopy system that automates malaria parasite detection in rural and low-resource health facilities.
Most people in rural areas are treated for malaria based on their symptoms, Kamugisha Edson says, and that has increased antimalarial resistance. Misdiagnosis across sub-Saharan Africa is the problem he is working on.
The reason clinicians fall back on symptoms is that reliable diagnostics are out of reach. Across many rural health facilities in sub-Saharan Africa there are too few trained microscopists, so the work leans heavily on Rapid Diagnostic Tests. Those have limits in both directions. They return false negatives where parasite density is low, and where emerging parasite strains have deleted the pfhrp2/3 target genes. They return false positives because target antigens such as HRP2 can stay in a patient's bloodstream for weeks after treatment has cleared the parasites, and because of cross-reactivity with rheumatoid factors. The result is presumptive treatment, which accelerates antimalarial drug resistance and leaves other febrile illnesses untreated.
“Innovation in global health shouldn't be constrained by resources or infrastructure. Every line of code and hardware iteration brings us one step closer to making automated malaria detection accessible to every community.”
Kamugisha Edson
His answer is PathoScan: an offline, AI-driven digital microscopy system built for resource-constrained health settings. Custom edge computing hardware paired with optimised deep learning models automates parasite detection and quantification, and it runs fully offline. The aim is to give frontline health workers rapid, laboratory-grade diagnostic confidence, and therefore accurate treatment decisions at the point of care, whatever the local infrastructure. He studies at Kyambogo University.
The device is still in development.