AI Model Can Better Diagnose Human Breast Cancer Than Human Experts: Study
Wed, April 21, 2021

AI Model Can Better Diagnose Human Breast Cancer Than Human Experts: Study

New AI model detects breast cancer by analyzing mammograms to determine whether the disease is present / Credits: Chompoo Suriyo via 123RF

 

One of the most common cancers for women is breast cancer, with more than two million new diagnoses last 2019 alone. More and more women are encouraged to have a check-up to detect early signs of breast cancer. However, interpreting the scans from a mammogram leaves room for error. Reports show that a small percentage of all mammograms return false readings. This can misdiagnose a healthy patient or miss the disease as it spreads. 

But, artificial intelligence has a solution for this. Recently, researchers from Google announced that they have trained an AI model that can detect cancer in breast scans better than medical professionals. It can reduce false positives by 5.7% and false negatives by 9.4%. The researchers collected anonymized mammograms from some 76,000 women in the UK and 15,000 women in the US to build the AI system. After that, they tested it on the X-rays of a different group to determine how often the AI was right about the diagnosis. 

According to Vox, a liberal-leaning American news and opinion website owned by Vox Media, the AI model works by analyzing mammograms to determine whether the disease is present. "The earlier you identify a breast cancer the better it is for the patient. We think about this technology in a way that supports and enables an expert, or a patient ultimately, to get the best outcome from whatever diagnostics they've had,” Dominic King, UK lead at Google Health, said. 

People living in Britain who want a check-up for breast cancer go through a labor-extensive process as all mammograms are reviewed by two radiologists. The Google Health team conducted experiments comparing the computer’s decision with that of the first human scan reader. According to Science Alert, a website that promotes general science news, using AI to verify the first human expert reviewer's diagnosis could save up to 88% of the workload for the second clinician.

"There's the opportunity for this technology to support the existing excellent service of the (human) reviewers,” Ken Young, a doctor who manages mammogram collection for Cancer Research UK, said.