Incorrect detection of breast cancer via a mammogram is not a new thing anymore. Google has taken years to come up with an effective way to fix these irregularities in the diagnosis of breast cancer. Google’s health division has developed an artificial intelligence feature to sort out the issue. This model can identify cancer, which is missed by the radiologists during the mammogram. The tech company has made this AI model along with its subsidiary firm Deepmind Institute, Cancer Research UK Imperial Centre and Northwestern University. This method was able to detect cancer missed by mammograms when it was tested against the scans of 25000 women in the UK and around 3000 women in the US. AI model was also able to rule out the problem of false diagnosis in healthy women.
Dominic King the head of Google’s health division in the UK has said that his team is working hard to develop the technology, which can diagnose cancer with greater precision. He has also made it clear that further approvals and clinical testing are required before this technology makes a way to the market. AI will be able to eliminate the need for double screening by the radiologists. It will be able to deliver a real-time indication of cancer in the scans if further studies on AI show a green signal to it. During the trial in the US, AI has performed better than human experts. It has delivered a 5.7 percent less false-positive diagnosis. It has also recorded 9.4 percent less false negative cancer detections.
Researchers said that AI initially detected some cancers overlooked by the radiologists. At the same time, it missed some cancers diagnosed by the radiologists. These findings show that it can also be inaccurate in some cases. Experts are trying to make it more accurate and reliable for the future. The research on this technology has been showing good results. If it is successful, AI will be able to work in sync with health professionals in the time to come. A mammogram is the most commonly used method for the diagnosis of breast cancer. However, it tends to miss one in five breast cancers. As per the data, around 50 percent of women who go through this testing in a span of 10 years in the United States get a false diagnosis of breast cancer.