NEW YORK (CBSNewYork) – Using computer artificial intelligence is offering some exciting news about better early detection of breast cancer.
Artificial intelligence, or AI as it’s called, is everywhere these days, reports CBS2’s Dr. Max Gomez.
It’s what’s making driverless cars, smart vacuum cleaners and iPhone assistants possible. It’s also being put to more important uses saving lives, such as improving mammogram detection of breast cancer.
Women are telling their stories to a non-profit organization called Dense Breasts Canada.
Their mammograms had missed their cancers because they had dense breasts, which is where mammograms are not as good at seeing cancers.
“They didn’t tell me the implications of dense breasts, I wish I had known,” said one woman.
“I did not know there there are cancers that cannot be seen by mammography,” said another.
Dr. Linda Moy, the head of artificial intelligence breast imaging research at NYU Langone Health, is leading ways to improve on mammograms.
“Mammograms are good but they are not perfect,” said Moy.
A.I. is particularly well suited to doing that work. Human breast cancer experts can show computers what breast cancer looks like on mammograms, and then feed it millions of mammograms so that the A.I. machine begins to get better and better at finding breast cancers.
“There are many subtle features that only the computer can see,” said Moy. “The human eye and the human brain can only process so much.”
A new study in the journal Nature reported that used an A.I. system developed by Google performed somewhat better than radiologists at finding breast cancers and reducing the number of false positives that humans found in mammograms.
It’s not a perfect system. In some cases, A.I. missed a cancer that all the test radiologists found, but A.I. still has some advantages over humans.
“The computer doesn’t get tired, it never needs a cup of coffee,” said Moy.
A.I. is better at finding cancers in dense breasts, but for now, Moy says that the best results will come from combining the expertise of human radiologists with the tireless, sharp eyes of computers to find early and curable breast cancers.