A team of researchers from USC has developed a groundbreaking AI algorithm called RED (Rare Event Detection) that can spot tiny traces of cancer cells in blood samples within just 10 minutes — cutting manual review time by over 1,000×. Unlike traditional diagnostic systems, RED doesn’t rely on predefined cancer images or labeled data. Instead, it studies millions of normal cells and flags anything that looks unusually different. In testing, this self-learning approach achieved an incredible 99%