Data Science Investment Counter —Funding raised by UK data science companies in 2018.
£ 5.640 Million

Google-backed study shows AI can be used to detect lung cancer

A study by Google and various US medical centres has shown how artificial intelligence (AI) can be used to detect lung cancer and provide doctors with more accurate computerised tomography (CT) scans readings.

The paper with the new findings was published in Nature Medicine on May 20. In it, the researchers compared the results of the AI system with those of six board-certified US radiologists with up to 20 years of clinical experience each.

When prior LDCT scans were not available, the study claimed, the AI “model outperformed all six radiologists with absolute reductions of 11% in false positives and 5% in false negatives”. When previous imaging was available, the AI did just as well as the radiologists.

It is not the first time AI is used to improve scans reading for cancer treatment. In November last year, the UK government funded an AI start-up to help with breast cancer screenings.

And more recently, scientists from Heidelberg University Hospital and the German Cancer Research Centre developed a machine learning method for the automated image analysis of brain tumours.

However, the new study makes use of Google’s supercomputer to greatly enhance the AI’s capabilities.

“In order to build the AI to view the CTs in this way, you require an enormous computer system of Google-scale,” Study co-author Dr Mozziyar Etemadi said. “The concept is novel, but the actual engineering of it is also novel because of the scale.”

Dr Etemadi also added that systems like this could help not only in cancer treatment but also to avoid costly procedures for people that are mistakenly diagnosed with cancer.

“Not only can we better diagnose someone with cancer,” he said, “we can also say if someone doesn’t have cancer, potentially saving them from an invasive, costly, and risky lung biopsy.”

While it prays the potential of AI in the medical field, however, the study concludes by warning that in order for this technology to become widespread, it is first necessary to validate these results in larger cohorts.

Image via Wikipedia and Pixabay.


Co-working space and blog dedicated to all things data science.

Subscribe to our newsletter