Google AI Program Better Than Practicing, Licensed Radiologists At Detecting Cancer

While we use technology for countless applications in today’s world, we haven’t yet worked out all the bugs related to artificial technology and technology in general. However, as we build upon the strong base of tech that we already have access to, we’re beginning to have access to technology that is better than people at some of their jobs.

Take, for example, a recent unveiling of an artificial technology network built by Google that was found to detect lung cancer better than trained, licensed, practicing radiologists.

A peer-reviewed study published in Nature Medicine, one of the medical industry’s leading journal, published Monday, May 20, 2019, found that Google’s piece of technology did better at detecting lung cancer than radiologists. The study involved 6,716 cases from the database of National Lung Cancer Screening Trial cases, as well as 1,139 cases outside of the National Lung Cancer Screening Trial’s database.

Here’s what the study consisted of. The researchers, who hailed from the likes of New York University, Northwestern Medicine, and Google AI – where most of them came from – carried out two trials. On both of the trials, both radiologists and Google’s innovative technology participated. One of the trials consisted of scans of patients’ lungs in which a prior scan was available for analysis, whereas the second trial did not feature scans in which more than one scan was available.

Six radiologists tried to identify cases of lung cancer in the computed tomography (CT) scans. Their performance was stacked up against Google’s program. For the first of the two trials, the one in which prior scans were available to participants, Google’s screening technology outperformed the radiologists. In the second trial, the performance of the six radiologists and Google’s technology were roughly the same.

In quantified terms, the artificial intelligence developed by Google outperformed the radiologists by roughly five percent. The five-percent improvement in performance included both not recognizing false positives and detecting cancer that the radiologists simply couldn’t detect.

Although the performance of the technology is nothing short of amazing, the algorithm from Google isn’t going into practice just yet. One of the reasons why Google’s technology isn’t going to be implemented yet is because the study consisted of scans from patients who had already been treated or who had succumbed to lung cancer.

Currently, here in the United States, lung cancer is the number-one killer in terms of all cancers. Smoking tobacco is one of the most common causes of lung cancer across the nation.

Recommended For You


About the Author: admin

Leave a Reply

Your email address will not be published. Required fields are marked *