ChatGPT-4 demonstrated near-perfect accuracy in extracting clinical variables from MRI and CT scans of pancreatic cyst patients, equalling manual chart review by radiologists.
Key Details
- 1Study involved 991 patients with 3,198 MRI and CT scans from Memorial Sloan Kettering (2010–2024).
- 2ChatGPT-4 extracted 9 critical clinical variables with accuracy rates of 92%–99%.
- 3AI matched the manual approach, considered the gold standard, for classifying high-risk variables such as cyst size and solid component presence.
- 4Efficiency and cost-effectiveness cited as major advantages of the AI approach.
- 5The study was published in the Journal of the American College of Surgeons and NIH/NCI partially funded the work.
Why It Matters
Demonstrating LLMs’ ability to accurately extract structured data from radiology imaging reports could streamline large-scale research and surveillance tasks, reducing costs and improving patient care efficiency. This proof-of-concept signals the increasing feasibility and trustworthiness of leveraging advanced AI models in radiological workflows.
Source
EurekAlert
Related News
•EurekAlert
Imaging Reveals Skull Changes and Immune Impact in Glioblastoma
Advanced imaging uncovers that glioblastoma affects the skull and immune system, not just the brain.
•EurekAlert
AI-Based CT Analysis Predicts Outcomes in Fibrotic Lung Disease
AI analysis of one-year CT changes predicts disease progression and survival in fibrotic interstitial lung disease.
•EurekAlert
AI and Multiphoton Microscopy Achieve 96% Accuracy in Pancreatic Tumor Detection
University of Arizona researchers combined label-free multiphoton microscopy with neural networks to accurately classify pancreatic neuroendocrine neoplasms in tissue samples.