{
  "id": 9151899,
  "title": "AI assistant learns from radiology reports to spot problems in abdominal scans",
  "url": "https://urgent.news/2026/09/22/ai-assistant-learns-from-radiology-reports-to-spot-problems-in",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-22T14:40:07.000Z",
  "source": {
    "name": "Medical Xpress",
    "slug": "medical-xpress",
    "url": "https://medicalxpress.com/news/2026-09-ai-radiology-problems-abdominal-scans.html"
  },
  "original_language": "en",
  "account": "An AI system named RADAR has been developed to assist doctors in interpreting abdominal CT scans. It works by leveraging a vision-language framework that learns directly from clinical reports containing 424,911 examinations, 1.5 million image-text pairs, and over 15 million anatomy-specific pairs. This approach eliminates the need for manual data labeling, making the system scalable and cost-effective. When tested on real-world hospital examinations and emergency cases from eight independent hospitals, RADAR achieved an average diagnostic accuracy score of 0.913, outperforming existing medical vision-language AI models. The tool accurately identified 18 anatomical structures and 146 imaging signs and diseases, demonstrating its capability to handle the complexity of abdominal CT scans. RADAR's ability to boost radiologist sensitivity by around 10% and reduce reading time by approximately 31% suggests its potential to enhance diagnostic accuracy for radiologists of all experience levels. However, further testing is needed to ensure its effectiveness across diverse patient populations and clinical environments.",
  "summary": "An abdominal CT scan packs an extraordinary amount of information into a single image, layering dozens of organs and hundreds of possible conditions. It's widely considered one of the most complex scans in medicine to read, and even expert radiologists need time to carefully untangle what they're looking at. To help ease some of their workload, researchers presented a new AI assistant called…",
  "key_points": [
    "RADAR AI system assists in interpreting abdominal CT scans",
    "Trained on 424,911 examinations and 15 million anatomy-specific pairs",
    "Achieves 0.913 diagnostic accuracy, outperforming existing models"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}