*(Source: The Rundown AI Newsletter)* Johns Hopkins researchers have developed AbdomenAtlas, an AI-powered dataset containing 45,000 3D CT scans with 142 annotated anatomical structures, significantly improving medical image analysis for early cancer detection. The dataset, sourced from 145 hospitals worldwide, is 36 times larger than its closest competitor. AI and expert radiologists completed in two years what would have taken humans 2,500 years, achieving a 500-fold speedup in organ annotation and a 10-fold improvement in tumor identification. While AbdomenAtlas marks a major advancement, it still represents only 0.05% of annual U.S. CT scans, emphasizing the early stage of AI-driven medical imaging. https://medicalxpress.com/news/2025-02-ai-powered-abdomen-cancer-early.html