Automated microscopy for malaria diagnosis in a reference laboratory in nonendemic settings

Area of Science:

• Medical Diagnostics
• Artificial Intelligence in Healthcare
• Parasitology

Background:

• Malaria diagnosis is crucial for effective case management, control, and elimination strategies.
• miLab™ is a digital microscopy system utilizing AI-driven deep learning for automated Plasmodium parasite detection and parasitemia level determination.
• The system differentiates infected from non-infected red blood cells in blood smears.

Purpose of the Study:

• To evaluate the diagnostic performance of miLab™ automated microscopy for malaria.
• To compare miLab™ against conventional microscopy and nested-multiplex malaria polymerase chain reaction (NM-PCR).
• To assess miLab™ utility in a nonendemic country's reference laboratory setting.

Main Methods:

• Prospective analysis of 400 samples from 2021 to 2024.
• Utilized miLab™ automated microscopy, conventional microscopy, and NM-PCR as reference methods.
• Evaluated concordance, sensitivity, specificity, and parasite density correlation.

Main Results:

• miLab™ demonstrated substantial concordance (90.8%) with conventional microscopy (kappa=0.8, sensitivity=92.1%, specificity=89.4%).
• Parasite density correlation was significant (r=0.77), though miLab™ underestimated counts by 11.6%.
• Compared to NM-PCR, miLab™ showed 62.8% sensitivity and 95.4% specificity (kappa=0.4), accurately identifying 63.4% of P. falciparum infections.

Conclusions:

• miLab™ automated microscopy offers comparable sensitivity to conventional methods without requiring expert microscopists and reduces turnaround time.
• It serves as a valuable tool for malaria diagnosis, particularly in nonendemic regions.
• Further enhancements are necessary for precise malaria species identification and accurate parasite quantification.