[ad_1]
Google unveiled an artificial intelligence (AI) system that has demonstrated an amazing ability to detect lung cancer in early tests and even surpassed radiologists.
According to a study published in Nature Medicine, the technology giant has driven the algorithm on 42,000 CT scans taken in a clinical trial of the National Institutes of Health and outperformed six radiologists to determine whether patients were suffering from or not a cancer.
The AI has been able to detect 5% more cancers and suppress false positives – that is, when a cancer is suspected but that a nodule is actually harmless – 11% by reviewing a single sweep, says the study.
The crooks use FACEBOOK to make people think that they gave money to ISIS
"These people have a technology that will dramatically improve the accuracy of screening," Dr. Otis Brawley, a professor of oncology and epidemiology at Johns Hopkins University told STAT. did not participate in the study.
Brawley told the media that he was generally skeptical about lung cancer screening, but pointed out that Google's performance in reducing false positives was a significant advance.
"It will prevent other bad things for people who are screened," Brawley added.
A radiology technologist examines the scan of a lung with a tumor at the UCSF Cancer Center in San Francisco.
(Getty Images)
Artificial intelligence has long been considered a way to potentially improve screenings and allow doctors to spot a range of malignancies with much greater precision.
GOOGLE'S ERIC SCHMIDT: BREAKING A BIG TECH WOULD BE ILLEGAL & # 39;
However, Google's system will have to undergo more rigorous testing before it can be put into medical practice. Since the study was limited to previously treated patients, it is not known if the system would lead to better results if it was used in new patients.
Lung cancer is the leading cause of cancer deaths in men and women. According to the American Cancer Society, about 228,000 new cases of lung cancer and 142,000 deaths result each year.
CLICK HERE FOR THE FOX NEWS APP
[ad_2]
Source link