University of Helsinki

Researchers studying puumalavirus-infected patients with acute hemorrhagic fever with renal syndrome (HFRS) needed to quantify specific immune cell markers (HLA-DR, CD14, CD16, CD68) in kidney biopsy specimens. Manual microscopy-based cell counting across stained slides was labor-intensive and prone to variability.

The team deployed Aiforia's deep learning AI platform to automate quantification of monocyte and macrophage markers relative to total cell counts in scanned kidney biopsy specimens. The AI model was trained over approximately 2,500 iterations based on around 500 annotations.

The AI-assisted analysis revealed increased numbers of cells expressing monocyte and macrophage markers in the kidneys of HFRS patients compared to controls, contributing to two peer-reviewed publications in PLOS Pathogens (2021). The approach enabled scalable, consistent quantification across large slide sets.

• Deep learning AI can automate tedious cell quantification tasks in infectious disease pathology research with relatively modest annotation effort (~500 annotations).
• AI-assisted image analysis enables reproducible quantification of immune infiltrates, reducing inter-observer variability in research studies.
• Academic pathology labs can leverage commercial AI platforms to accelerate publication-quality research without building custom tools.