Key Takeaways from the Article:

1. The World Health Organization (WHO) has set ambitious targets for neglected tropical diseases (NTDs), emphasizing the need for improved diagnostics to achieve these goals.

2. Current diagnostic methods for NTDs, such as microscopic examination, have limitations in terms of sensitivity, reproducibility, and manual read-out errors.

3. The WHO's Diagnostics and Technical Advisory Group (DTAG) is working on target product profiles (TPPs) to guide the development of new diagnostics, but the translation of biomarkers into affordable tests is challenging.

4. Artificial intelligence-based digital pathology (AI-DP) is a promising alternative for NTD diagnostics, offering improved reproducibility, automation, and cost reduction.

5. The study presents a proof-of-concept for an affordable AI-DP prototype for detecting soil-transmitted helminths (STH) and intestinal Schistosoma (SCH) eggs in stool samples.

6. An early prototype of a whole slide imaging (WSI) scanner was engineered for capturing digital images of STH and SCH eggs in Kato-Katz (KK) stool thick smears.

7. A database of over 1.3 million images was created, with annotations for eggs of different helminth species.

8. Using a deep learning approach, the AI model achieved high precision and recall for detecting and classifying STH and SCH eggs.

9. The prototype WSI scanner demonstrated the technical feasibility of scanning KK stool thick smears in field settings.

10. The development of affordable and fieldable AI-DP scanners shows promise in supporting the WHO's roadmap for NTDs and improving diagnostics for STH and SCH infections.

Further research and development are necessary to adapt the prototype for field settings, conduct end-to-end testing in the field, and evaluate performance across different infection intensities.

Read more: https://doi.org/10.1371/journal.pntd.0010500