Liver fibrosis and inflammation,quantified across the whole section

In preclinical MASH studies, Aiosyn's AI quantifies fibrosis and lobular inflammation across the full liver section of every animal in your cohort. Read-outs are returned per animal and per experimental group, produced by one fixed method on every slide.

Discuss your study Biopharma overview

Full section quantified, no sampled fields
Same section, same read-out, every run

Drag the handle. Left: the stained liver. Right: the same tissue, with fibrosis segmented edge to edge.

Fibrotic tissue

Fibrosis read-outs tracked expert pathologist scoring

Together with TNO, we validated our fibrosis read-out on mouse liver sections spanning treated and untreated groups. An expert pathologist scored the percentage of each section covered by fibrosis, and the AI measured it independently on the same sections.

A whole mouse liver section with fibrosis segmented across the tissue by the Aiosyn AI, shown in green. — One whole liver section from the validation, fibrosis segmented across the entire tissue with 25% coverage.

AI against expert pathologist

Fibrotic tissue, whole liver section

Correlation: r = 0.95
Variance explained: R² = 0.91
Sections: 76

Each point is one section: the pathologist's score on the horizontal axis, the AI's on the vertical, against the line of perfect agreement.

We ran this validation with TNO.

Inflammation, detected on the same section

Inflammatory-cell aggregates are detected across the full liver section and returned as a density per mm², on the same tissue that gave the fibrosis read-out. The individual cells inside each aggregate are counted on the same pass, so cells per aggregate and total counts come back with it.

A steatotic liver field with one inflammatory-cell aggregate outlined by the AI. — The aggregates that carry the read-out, the cells the AI counted inside them, and a view of where inflammation concentrates across the whole section.

A steatotic liver field with individual inflammatory cells detected and marked one by one. — The aggregates that carry the read-out, the cells the AI counted inside them, and a view of where inflammation concentrates across the whole section.

From liver section to MASH read-outs

1
Share your sections
Whole-slide images of the liver sections in the formats you already scan. No change to how they are cut, stained or digitized.
2
Every section is read
Fibrosis is segmented, inflammatory cells are detected, and macrovesicular and microvesicular steatosis are quantified as affected tissue, on every slide in the cohort under one fixed method.
3
Read-outs come back quantified
Per-animal and per-group numbers, with the segmentations and detections behind them.

A liver read-out this page does not cover

Another stain, another species or an endpoint specific to your model is built on your own images as custom AI development.

Co-developed with TNO Health & Work

Through our partnership, we co-develop these models together with TNO. Their preclinical and pharmaceutical research expertise sits behind the study design and the tissue, our AI behind the read-outs, and the models are applied within TNO's preclinical studies with more study areas in development.