Get started

Please fill in the form below and an Ezra representative will contact you within 1 business day.

Notify me

Please fill in the form below and we will be in touch when we launch in your city.

Blog / Artificial Intelligence, Cancer, Early Detection, Healthcare Trends, Lung Health

AI Can Help Radiologists Identify Lung Tumors While Decreasing False Positive Rates

Dec. 03 2019 by Sheherzad Raza Preisler Blog Editor
AI Can Help Radiologists Identify Lung Tumors While Decreasing False Positive Rates

A group of scientists in Seoul, Korea published an exciting retrospective study in Radiology in mid November. It showed that an artificial intelligence (AI)-powered deep learning software can improve clinicians’ capabilities to identify lung cancer on chest x-rays. 

The study looked at 800 random radiographs, and found that, with the help of the AI, the average doctor’s capacity to find an existing cancer increased from 65.1% to 70.3%. What’s more is that the rate of false positives decreased with the AI’s help, too. This is important because certain qualities found in lung lesions–such as location, size, and density–prove difficult for doctors to find nodules on x-rays. However, the AI software may be able to help bridge this gap, according to Byoung Wook Choi, the radiologist and professor at Yonsei University College of Medicine, who was a senior author on the study.

Choi and his colleagues evaluated 800 radiographs they collected from four separate imaging centers; around 200 of these were considered “normal,” while 600 contained at least one malignant nodule, and 704 contained confirmed tumors. Afterwards, a group of radiologists interpreted the same collection of x-rays and re-read them with the assistance of deep convolutional neural networks (DCNN). It was then that Choi discovered the approximately 5% increase in a radiologist’s ability to detect existence cancer with the help of the AI, as well as the decreased rate of false positives (it dropped from 0.2 to 0.18). 

The finding regarding the decreased number of false positives could certainly address a major issue with AI’s, because many tend to exhibit a high rate of them. In a statement, Choi remarked: “Computer-aided detection software to detect lung nodules has not been widely accepted and utilized because of high false positive rates, even though it provides relatively high sensitivity. DCNN may be a solution to reduce the number of false positives.”

If your Ezra medical provider determines that you’re at an increased risk of developing lung cancer, we may include a low dose CT of the lungs with your scan at no additional cost. If you’d like to learn about our screening options, you may do so at this link.