AI-powered automated microscopy for PARAsite DETECTion

Background

Parasitic diseases, like malaria, pose major global health challenges, especially in tropical regions where soil-transmitted intestinal worm infections impact around 1.5 billion people. Climate change and increased migration led to a continuous rise in imported tropical diseases in non-tropical countries, including the Netherlands. The primary method for diagnosing parasitic infections is microscopic examination. Although alternatives like nucleic acid amplification methods (e.g., real-time PCR) exist, microscopy remains widely used due to its versatility and global accessibility. However, this method is labour-intensive, requires highly trained technicians, and is prone to errors with false-negative rates exceeding 10%.

Aim of the project

To improve accuracy and efficiency in parasite detection, there is a need for development of automated detection methods. Leven Vision, a Dutch start-up company, is developing an automated parasite detection system using artificial intelligence (AI). Their initial algorithms, based on publicly available digital images, have shown promise in detecting gastrointestinal parasites in stool samples. However, further development is required, including integration with commercially available scanning devices, additional training using validated clinical specimens, and validation in accredited laboratories.
Leven Vision partnered with Erasmus MC Rotterdam, an expert centre in clinical parasitology. This collaboration focuses on optimising the AI algorithm using validated clinical specimens and expertise in proper image acquisition. Our end goal is to enhance the AI’s accuracy and sensitivity, leading to a cost-effective, superior, automated microscopy system, which will facilitate more efficient and precise parasite detection. Once developed, the product will be validated in ISO-15189 certified laboratories to confirm its effectiveness compared to manual examination.
In summary, automation of parasitic detection in clinical samples is urgently needed, and Leven Vision, in collaboration with Erasmus MC, is advancing towards a robust AI-driven solution to meet this demand (visual summary video: https://files.leven.vision/emctki.mp4 5min).

Health Holland / Erasmus MC-TKI-LSH Pilot grant

Leven Vision B.V.

• Ihor Feoktistov (Leven Vision)
• Rob Koelewijn
• Jaap van Hellemond