AVO Get in touch

AI-driven cancer detectionAI that reads every mammogram.

AVO detects masses and calcifications, scores each finding for malignancy, and compares it against the patient's earlier studies — a second read on every screening mammogram, in under three minutes.

The problem

Every fourteen seconds, a woman is diagnosed with breast cancer.

1

Diagnosed since you opened this page

94 countries out of 194 have a screening programme.

0 of 194 countries screen

AVO makes a second read cheap enough to scale — including into mobile units, at the point of screening.

A second read can take three weeks.

Days to a second read

Day 1 — AVO, same session Day 21+ — conventional second read
~50

Scans a radiologist reads in a day. Demand is rising; capacity is not.

10–30%

Of breast cancers may be missed — higher in dense tissue.

A consistent, fatigue-free second read, in under three minutes.

Reported miss rates, reading volumes and review waits from published screening literature.

What the AI doesSix things the model does to every study.

Mass detection

Finds suspicious masses and marks them on the image itself, where the radiologist is already looking.

Calcification detection

Picks out microcalcification clusters, including in the dense tissue where they are hardest to see.

Malignancy scoring

Gives every finding a probability of malignancy, and the study a cumulative risk score out of ten.

Temporal comparison

Puts this year's findings beside the previous mammograms, so what changed is visible rather than remembered.

Worklist prioritisation

Thousands of screening studies ordered by risk, so the queue reflects urgency instead of arrival.

Learning loop

Radiologist feedback returns to the model, so it keeps improving on the population it actually reads.

A batch of processed studies in AVO, each with a risk score, the
                    abnormalities found and a classification.
A batch as the model leaves it: every study scored, with what it found and how it reads.

A second read in under three minutes, inside the radiology workflow you already have.

What comes backOne study, read and scored.

Sample outputRead time 2m 40s
Cumulative risk spectrum score

9.4/ 10

Finding
Suspicious mass, suspicious calcifications
Malignancies
2 found
Flag
Priority review
A mammogram in the AVO viewer. One region is circled, magnified beside the
                    image, and labelled 08 Mass, Suspicious.
The same finding in the viewer: circled on the image, magnified, and labelled where the radiologist is already looking.

Who's behind it

TenX

An enterprise AI provider. Computer-vision and software engineers, delivering across five countries.

Shaukat Khanum Memorial Cancer Hospital and Research Centre

A cancer registry running since 1994. 147,120 neoplasms recorded, over 8,700 patients treated a year. JCI Enterprise re-accredited.

Clinical validation

1,300+

Real-world patients reviewed

3,000+

Mammograms annotated by trained radiologists

JCI

Accredited partner hospital

The clinical study has concluded. The models continue to improve through a feedback loop with the radiologists reviewing them. AVO is being established as its own company, with its own management.

Where it runsInto the reading room, or out to the van.

AVO arrives through your PACS, on-premise or in the cloud, to match your regulator — and runs the same way inside a mobile screening unit, where the scan is taken hundreds of miles from the nearest radiologist.

Mobile screening unit Hospital reading room R CC L CC R MLO L MLO PACS L CC AVO on-premise or cloud 3 min 9 6 4 Back in the worklist
Studies reach AVO the same way wherever they were taken, and come back into the queue the radiologist already works from. The outlines are traced from a real mammogram.

Get in touchIf this is the problem you have, we would like to hear from you.

info@avolabs.ai

Or find AVO on LinkedIn.

Please don't send patient data or images by email.