Back

Project Details

#NT100is5 One To Watch: Zebra

Transforming disease diagnostics with AI

By Zebra

Project URL: https://www.zebra-med.com/
Project Twitter: @ZebraMedVision

The problem

When Phil Teare’s wife died from cervical cancer at an early age, he chose to dedicate the rest of his life to helping medical professionals to detect diseases at earlier stages of onset. In 2012, there were an estimated 527,000 cases of cervical cancer, of which 266,000 sadly proved fatal. The incidence of fatalities from cervical cancer is even greater in developing countries.

The response

Teare taught himself machine learning and, working with Israeli medical startup Zebra, developed an algorithm to detect breast cancer. Zebra uses AI to scan medical images to detect cancerous cells, achieving a reported 91% accuracy; this is a significant improvement on a typical radiologist’s rate of 88%, and with fewer false positives.

Thanks to a partnership with Google Cloud, Zebra analysis is available for radiologists to download from the cloud at just $1 per scan, making the service hugely accessible. The system is already used by more than 50 hospitals worldwide, including the Cedars-Sinai network in the US, and Oxford University in the UK.

Zebra’s AI can detect 11 diseases, including breast cancer, brain trauma and excess coronary calcium. Zebra is not the only company using AI for medical diagnostics: IBM has taught its AI Watson to read medical scans, and startup Face2Gene uses AI to diagnose a range of conditions from a simple photograph.

The potential

Zebra’s co-founder Elan Benjamin believes the tech will make doctors’ jobs easier. “In five or seven years, radiologists won’t be doing the same job they’re doing today,” he told Wired magazine. “They’re going to have analytics engines or bots like ours that will be doing 60, 70, 80 per cent of their work.”

The company is planning to release 35 additional diagnostic products and expand into a further 20 countries in 2018.

Last updated: 07th of February, 2018

Cookies on the Social Tech Guide

Social Tech Guide uses cookies to provide you with an enhanced user experience. To remove this message, please click continue.

Continue