This startup develops AI-powered software that automates the analysis of cardiac MRI images. Their software provides automated interpretation of these images, helping clinicians improve efficiency and make more informed decisions.
Funding
Funding not disclosed
Founders
Product
Problem
Cardiac Magnetic Resonance (CMR) imaging is the gold standard for assessing cardiac anatomy and function, but current analysis methods are time-consuming, require specialized expertise, and can be subjective, leading to variability in interpretation. This complexity limits the scalability and accessibility of CMR, hindering timely and accurate diagnoses of cardiovascular diseases.
Solution
AI4MedImaging offers AI4CMR, a software-as-a-medical-device (SaMD) solution that automates the interpretation of CMR images, providing clinicians with rapid and objective analysis of cardiac function and structure. The AI-powered software performs fully automatic cardiac segmentation, enabling instant quantification of ventricular volumes, myocardial mass, and ejection fraction. AI4CMR automatically identifies cases with normal left ventricle wall motion, serving as a safety net and allowing physicians to focus on complex or abnormal cases. The cloud-based service integrates seamlessly with existing PACS or VNA viewers, improving clinical workflow and reducing reporting time by more than 50%.
Target Audience
The primary target audience includes cardiologists, radiologists, and other medical professionals involved in cardiac imaging who seek to improve workflow efficiency, reduce variability, and enhance the accuracy of CMR image analysis.
Features
- Fully automatic cardiac segmentation using state-of-the-art algorithms with high overlap with expert manual segmentation (Dice coefficient >= 0.8)
- Automatic quantification of left and right ventricle volumes, including end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), ejection fraction (EF), myocardial mass (LVM), and cardiac output (CO)
- Automatic identification of cases with normal vs. suspicious wall motion with high sensitivity and negative predictive value
- Significant reduction in cine reporting time, with an average response time of 86 seconds for segmentation compared to 20 minutes for manual segmentation
- Compatible with different MRI manufacturers, including Siemens, GE Healthcare Systems, and Philips Medical Systems
- Cloud-based service offering high availability, scalability, and reliability
- Easy communication with other devices via standard protocols
- Seamless integration with PACS or VNA viewers, eliminating the need for dedicated viewer software