Octo · H&EBuilt on Kraken

Octo Cellularity

Tumour cellularity and tumour–stroma ratio on H&E

Tumour cellularity and tumour–stroma ratio from routine H&E, nucleus by nucleus, before the block goes to sequencing and in the pathology report.

Request a demo See the product sheet

Octo Cellularity output — Tumour cells detected across the section on H&E · drag to compare

Octo Cellularity: raw slide — Tumour cells detected across the section on H&E · drag to compare

How it works

From slide to cellularity result

1Scan the blockH&E slide from any scanner
2Draw or accept the regionthe area you would macro-dissect
3Every nucleus classifiedtumour, stroma, immune, other
4Cellularity and TSRwith the region saved alongside
5Send to sequencingor into the report

How much of the slide is read

By eyeEstimated by eye on a few fields

Octo CellularityEvery cell on the slide

Illustrative proportions.

Where it applies

NGS block selectionMacro-dissection guidanceTumour–stroma ratio in colorectal cancerTrial eligibility on tumour content

One model, two applications

One cell map. Two answers.

Octo Cellularity classifies every nucleus in the region once. From that single map it reports the tumour fraction for molecular testing and the tumour–stroma ratio for the pathology report.

One readEvery nucleus classifiedtumour · stroma · immune · other

Application 1Tumour cellularityTumour nuclei as a share of all nuclei in the region, so the right block goes to sequencing.For molecular testing

Application 2Tumour–stroma ratioStroma as a share of the tumour area, reported as stroma-low or stroma-high.For the pathology report

The problem

Cellularity is currently estimated by eye.

Before NGS, someone looks at the H&E and guesses the tumour fraction. Under-called cellularity fails the run; over-called cellularity wastes it, and both delay the patient’s result.

Sequencing depends on the tumour fraction. Sequencing assays need a minimum tumour fraction. A reproducible cellularity, with the region it was measured on, means fewer failed runs and a defensible basis for the choice of block.

What changes

For the pathologist

Every nucleus counted

Every nucleus in the region classified as tumour or not, from 14.1 million labelled cells.

Same block, same number

The same percentage for the same block, whoever marks it up and whenever.

The whole region

Whole-region measurement instead of a representative field; region editable by the pathologist.

Outputs

What you get back

Output	What you get	Format
Tumour cellularity	percentage of nuclei classified as tumour in the selected region	data
Tumour–stroma ratio	stroma as a share of the tumour area, reported as stroma-low or stroma-high	data
Tumour cell count	absolute count and density per mm²	data
Cell map	every nucleus classified, tumour vs. non-tumour	overlay
Region export	coordinates for the macro-dissection template	JSON / PDF

Reported to standard

Tumour cellularity reported as required for molecular testing in RCPath tissue pathways · Reference →

Reported to standard

Tumour–stroma ratio reported as required by [guideline name, year] · Reference →

Evidence

83.1 %

agreement with consensus

414

cases in validation

Lancet Oncol.

tumour cell density, ARISTOTLE post-hoc (2025)

< 2 min

per slide

Tumour cell density quantified by AI enabled precise chemo-radiotherapy planning in a post-hoc analysis of the ARISTOTLE trial (The Lancet Oncology, 2025). Regulatory status: research use only.

Also in the Octo family

Octo

Try Octo Cellularity

Run it on a slide in the platform, or ask for the product sheet with the full output specification.

See it in your slide Request a demo