A Stanford-led study shows AI can accurately predict which preterm infants will suffer complications based on metabolite patterns in newborn blood samples.
Key Details
- 1AI model used dried blood spot samples from 13,536 premature infants in California, born more than 10 weeks early.
- 2Validated the algorithm on 3,299 preterm infants from Ontario, Canada.
- 3AI identified six key blood measurements, forming a metabolic health index.
- 4Combined with clinical data (gestational age, birth weight, sex, Apgar scores), the index predicted four major complications (intestinal, eye, lung, brain) with over 85% accuracy.
- 5Research published in Science Translational Medicine (DOI: 10.1126/scitranslmed.adv4942).
Why It Matters
This approach enables much more precise risk stratification for preemies, paving the way for personalized neonatal care, early intervention, and potentially reduced morbidity through targeted resource use. It represents a significant advance in applying AI to high-impact clinical prediction using routine biomedical data.
Source
EurekAlert
Related News
•EurekAlert
AI-Powered OCT Enables Rapid 'Optical Biopsy' for Early Endometrial Cancer Detection
A team at Washington University has developed a catheter-based 3D OCT system with AI to quickly and noninvasively detect early endometrial cancers.
•EurekAlert
AI Clinical Reasoning in Diagnostics and Digital Fatigue in Healthcare
Recent JMIR features explore large language models in clinical diagnostics and digital fatigue among healthcare professionals.
•EurekAlert
KAIST, MIT, Microsoft Develop Efficient AI Image Upsampling for Robotics
KAIST, MIT, and Microsoft have created 'Upsample Anything,' a training-free AI method to restore high-resolution visual data from compressed images with up to 16x improved GPU memory efficiency.