Evaluating the diagnostic performance of miLab™ for detection of malaria parasites using nPCR as reference standard
miLab™ showed strong malaria detection performance versus nPCR and outperformed microscopy in a Ghana hospital study.
Malaria Journal
|February 11, 2026
|Ebenezer Kojo Addae, et al.
Abstract
This hospital-based cross-sectional study in three malaria-endemic communities in Kumasi, Ghana, evaluated the diagnostic performance of the AI-assisted automated miLab™ platform using nested PCR (nPCR) as the reference standard. Among 300 suspected malaria patients, miLab™ achieved a sensitivity of 94.23%, specificity of 98.98%, and accuracy of 97.33% compared with nPCR, outperforming adjudicated microscopy, which showed 85.58% sensitivity, 97.96% specificity, and 93.67% accuracy. Parasite density estimates by miLab™ were within clinically useful ranges and the findings support its use as a reliable, rapid diagnostic alternative in malaria-endemic, resource-limited settings.
| Sensitivity (%) | Specificity (%) | Accuracy (%) | |
|---|---|---|---|
| miLab™ AI | 94.23 | 98.98 | 97.33 |
| Adjudicated Microscopy | 85.58 | 97.96 | 93.67 |
Table 1: Clinical Performance of miLab™ AI vs. Adjudicated Microscopy for Malaria Detection
Key Highlights
- Conducted as a hospital-based cross-sectional study in Kumasi, Ghana, from August 2024 to June 2025.
- Evaluated 300 patients with suspected malaria across three endemic sites.
- miLab™ was tested against nPCR, with microscopy as a comparator.
- miLab™ reached 94.23% sensitivity, 98.98% specificity, and 97.33% accuracy.
- Automated AI-based testing may help reduce dependence on expert microscopy in resource-limited settings.