How GE Healthcare used AWS to build an AI model for MRI scans

GE Healthcare has developed the first full-body three-dimensional MRI foundation model, marking a significant advance in medical imaging AI technology.

Key innovation; The new AI model processes complete 3D body scans, moving beyond the traditional two-dimensional approach that has limited medical imaging analysis.

• Built from scratch on AWS infrastructure, the model incorporates over 173,000 images from more than 19,000 medical studies
• The system achieves training efficiency by requiring 80% less computing power than comparable models
• Mass General Brigham has been selected as an early research partner to test the technology’s capabilities

Technical capabilities; The foundation model demonstrates advanced multimodal functionality that bridges the gap between visual and textual medical data.

• Enables sophisticated image-to-text searching capabilities within medical databases
• Shows superior performance in disease classification, including prostate cancer and Alzheimer’s detection
• Achieves 30% accuracy in matching MRI scans with text descriptions, a tenfold improvement over existing solutions

Infrastructure and architecture; The model leverages advanced cloud computing and GPU technology to deliver optimal performance.

• Implements a “resize and adapt” methodology to handle diverse imaging datasets
• Utilizes Amazon SageMaker for distributed training across multiple Nvidia A100 GPUs
• Integrates Amazon FSx and S3 for efficient data storage and management
• Optimizes costs through strategic use of Amazon EC2 instances

Clinical potential; The technology shows promise in transforming several aspects of medical imaging and patient care.

• Could reduce scan times for various imaging procedures, including X-rays
• May improve radiation therapy planning and execution
• Has potential applications across multiple medical specialties through fine-tuning

Looking ahead; While currently in the research phase, the model’s impressive early results suggest it could become a foundational technology in medical imaging, though regulatory approval and clinical validation will be critical next steps before widespread adoption in healthcare settings.