Medical data is sensitive and requires the implementation of adequate technical and organizational measures to ensure its security.
Compliance with data protection regulations (GDPR/HIPAA) is paramount.
Machine learning and deep-learning models require medical datasets with sufficient size and accuracy. Data quality is hard to define and project dependent.
Peri-procedural data comes from multiple sources. Lack of standardization in how medical data is stored and formatted.
Annotation, curation, and validation of medical data requires a high-level of expertise.
Representativeness of the data used to train the models is critical to avoid bias.
To ensure adoption, involvement of the different stakeholders from the development phase is critical and clinical evidence is key.
All data collected is anonymized or pseudonymized in compliance with Data Protection regulations. Notably, data undergoes state-of-the-art encryption (dual process, dual keys). Our policy meets the highest standards of data protection.
Our algorithms have been trained on very large databases of intracardiac signals, carefully annotated by expert physicians.
Our database is continuously enriched with new procedural data that allows us to further improve our AI solutions.
Our solution has undergone significant clinical validation with a multicentric clinical trial in Europe (the EvAI-Fib Trial, JCE 2022).
We are currently conducting an international randomized clinical trial (the Tailored-AF trial, NCT04702451) to further validate its use for persistent AF.