Johns Hopkins University BIOS Division: Advancing Intracerebral Hemorrhage Treatments Through AI
Johns Hopkins University BIOS Division is dedicated to improving clinical trials and medical image reviews through data science-driven approaches. To accelerate brain hemorrhage image analysis and improve treatment outcomes, the division used Google Cloud, including Cloud Healthcare API, Compute Engine, Cloud Dataflow, and Google Kubernetes Engine. They developed AI-driven models for brain scan analysis that reduce clinical trial brain scan review times from five hours to 30 seconds and improve diagnostic accuracy. The solution integrates DICOM medical imaging and automates machine learning training and inference pipelines for near real-time clinical trial capabilities. The implementation reduced research and infrastructure costs, accelerated insights generation, and democratized NIH study-quality core lab capabilities for more widespread healthcare impact.
- Organization
- Johns Hopkins University BIOS Division
- Industry
- Healthcare
- Location
- United States
- Published
- June 2024
Reported outcomes
30 seconds
timeTime & speed
Strategic outcomes
Primary read
Use case focus
Showing 3 of 3
- 1Medical Image Analysis
- 2Clinical Trial Acceleration
- 3Machine Learning
- They built a cloud-based machine learning system integrating Cloud Healthcare API for handling medical images, Compute Engine and Cloud Dataflow for pipeline automation, and Google Kubernetes Engine for automated inference.
- A robust AI algorithm was developed and trained on a curated library of brain hemorrhage patient images.
- The AI-driven system could analyze brain scans in about 30 seconds instead of 5 hours, allowing near real-time clinical trial operations.
- Clinical brain scan analysis time reduced from 5 hours to 30 seconds, improving speed and efficiency.
- AI-driven image analysis enhanced accuracy of brain hemorrhage data points with a dice coefficient of 0.93, comparable to expert human analysis.
- Research and infrastructure costs for clinical trials decreased significantly due to cloud adoption.
- Enables real-time capability in trial core laboratory functions, improving patient care and outcomes.
Architecture
The architecture integrates Cloud Healthcare API to manage DICOM medical images, Compute Engine and Cloud Dataflow for machine learning training and data pipelines, and Google Kubernetes Engine to automate inference pipelines, enabling fast, accurate brain hemorrhage image analysis.
Implementation partners1
Sources & evidence1
AI-generated summary. Verify important details with the linked sources before relying on this case.
Explore related AI use cases
Was this useful?