00:00A landmark study out of a major U.S. research institution has produced results
00:04that are reshaping how the medical community thinks about cancer detection.
00:08An artificial intelligence system trained on millions of medical records was able to identify
00:13early markers of certain cancers up to four years,
00:17before a human physician would have noticed anything abnormal.
00:20The study focused on pancreatic cancer, lung cancer, and colorectal cancer,
00:25three of the deadliest cancers in America,
00:28and found that AI analysis of routine blood tests, imaging,
00:32and patient records could flag risk patterns invisible to the human eye.
00:37Pancreatic cancer, which kills more than 50,000 Americans a year,
00:41has a survival rate above 40% when caught early, but drops to just 3% at stage 4.
00:47The potential impact of early AI detection on survival rates is enormous.
00:52The study is now being reviewed for FDA authorization.
00:55Several major U.S. health systems have already begun pilot programs.
01:00This may be the most consequential application of artificial intelligence in medicine in history.
Comments