Skip to content

Expert cell and tissue analysis, in minutes.

VSQRD AI measures and describes every sample in your study, from H&E to multiplex fluorescence to flow cytometry, without the hours and cost of manual analysis.

The problem

Most of what a sample shows is never measured.

  1. A single desk, drawn in isometric projection, with one document on it.

    Setting up a method

    Each stain, panel and measurement needs its own method. An analyst draws regions, labels cells or sets gates by hand until the method finds the right thing.

    • A new measurement takes days to set up
    • A classifier can need hundreds of drawn regions
    • Only one analyst knows how the method was built
  2. A hopper narrowing to a bottleneck: work items funnel in at the top and queue at the neck.

    Running it on a study

    The method runs. Then someone checks every sample by hand: folds, pen marks and bubbles on a slide, debris and drift in a flow file.

    • Artefacts are found sample by sample
    • Settings get changed for a sample that does not fit
    • Every study waits for the same analyst
  3. A vertical stack of papers, one sheet pulled part-way out for review.

    Deciding what to measure

    Each method costs days, so a study measures only the endpoints in the quote. The rest of what the samples show is never counted.

    • Endpoints are cut to fit the budget
    • A change someone sees is written down and never counted
    • A new question about old samples means a new project

How it works

Ask on one sample, and get the answer on every sample.

  1. 01Ask on one sample

    Ask for any measurement in plain words. In minutes, the AI builds the method on one sample and shows you every outline and gate it draws.

  2. 02Lock the method

    A pathologist or scientist reads the method, asks for changes and locks it. Its steps and settings stay fixed for the whole study.

  3. 03Measure, review and sign

    Code measures every sample, with no AI in the count. The AI then reads each sample beside its numbers, describes what it sees, and points out problems such as folds, blur or debris. A person accepts each change and signs the study.

What it measures

Methods for the stains and panels labs run most.

Each method is written from published guidance. Your lab can change any of them or write its own, and an owner approves each version.

  • H&E

    Fatty liver, inflammation and more

    Fat droplet area with macro and micro droplets apart, inflammation and necrosis, fat cell size, and heart muscle cell size.

  • Special stains

    Fibrosis and vessels

    Collagen area on picrosirius red, vessel wall thickness on elastic van Gieson, and muscle cell size on WGA.

  • IHC and ISH

    Any marker

    Positive area, positive cells with an H-score, a measurement inside one region of the tissue, double stains, RNA dots per cell, and vessel density.

  • Fluorescence

    Single and multiplex

    Positive area, and cells found and named by the markers they carry, on single stains and multiplex panels.

  • Flow cytometry

    Common panels

    T cell subsets and memory subsets, TBNK absolute counts, intracellular cytokines, phospho signalling, receptor occupancy, cytotoxicity, basophil activation and high-parameter phenotyping.

Why the numbers hold

Every number traces back to the step that made it.

A server and a database, drawn in isometric projection, standing together.
Only measured numbers
A report can quote a number only if code measured it. The AI never counts by eye.
Every sample, in full
Code reads every tile of a slide at the scanner's own resolution, and every event in a flow file. The AI checks samples that the software picks.
One method for the study
Settings stay fixed for the whole study. A sample that does not fit is flagged for a person.
Every step kept
Each view, setting and result is saved and cannot be changed later. A run can be replayed exactly.
A person signs
Clean-ups and steps wait for a person to accept them. Someone else approves the result.

Where to next

Read more, or talk to us.

Book a call with the two of us who build it. We turn the analysis on for your lab after it.

Start with one study you have already reported.

Bring its slides or flow files and the numbers you sent the client. We go through them with you and show you what else your data can tell you.

Book a Discovery Session