Global Medical Data Infrastructure for AI Systems with MedSyntra - Life Sciences Today Podcast Ep 52
Why It Matters
Medentra’s data‑infrastructure removes barriers to AI‑driven healthcare research, giving pharma, AI firms, and governments rapid, compliant access to high‑quality imaging data that can speed drug development and improve patient outcomes.
Key Takeaways
- •Medentra normalizes and de‑identifies imaging data for AI research.
- •Provides vendor‑neutral archiving, ensuring interoperable, AI‑ready datasets across modalities.
- •Connects hospitals to global research groups within hours, enabling rapid collaboration.
- •Serves pharma, AI firms, and governments while respecting national data sovereignty.
- •Client base growing fast, adding four to six new customers monthly.
Summary
The Life Sciences Today podcast introduced Medentra, a Tel‑Aviv‑based startup building a global infrastructure that transforms fragmented radiology and imaging data into AI‑ready assets for research and clinical use.
Medentra’s platform normalizes DICOM files, strips proprietary tags, and fully de‑identifies patient information while preserving longitudinal links via a common key. The service offers vendor‑neutral archiving and can be deployed on a client’s own infrastructure or in cloud environments (Azure, AWS, Google) that respect each country’s data‑sovereignty rules.
Co‑founders Kristoff Katilla and Avien explain that their LLM‑driven pipeline extracts precise disease descriptors from complex radiology reports, enabling rapid indexing of scans. They cite recent collaborations, such as supplying data for a tuberculosis‑screening model across India, South Africa, and Eastern Europe, and participation in Saudi Arabia’s Vision 2030 biotech accelerator.
By unlocking previously siloed imaging data, Medentra accelerates AI model training, supports pharmaceutical trial analytics, and empowers national health systems to develop population‑level insights, positioning the company as a critical enabler in the emerging med‑tech data economy.
Comments
Want to join the conversation?
Loading comments...