Using artificial intelligence to support clinical diagnostics and research

On-going development in the healthcare sector places an ever-growing demand on diagnostics. The aim of EMPAIA (Ecosystem for pathology diagnostics with AI assistance) is to enable the routine application by physicians of validated and certified AI-solutions for computer assisted diagnostics (CADx). To this end, the project is establishing an ecosystem for image-based diagnostics using pathology as a prime case study. The central instrument is a software platform on which suitable AI algorithms can be refined and validated. The platform will also offer AI services to support clinical diagnostics and research, the certification of which will be facilitated by a clear legal framework and the provision of a body of supporting developers, reference institutes and certifiers.

EMPAIA is developing a platform to remove the current barriers to the development, distribution and use of AI in pathology. The platform enables the development and global deployment of standardized, certified and explainable AI solutions. It delivers an infrastructure for data access and enforces the associated legal and financial regulations. Additionally, the platform provides an ecosystem for close cooperation with the responsible regulatory committees. Within these committees, regulations are developed and codified to the themes of standardization, economic viability, validation and certification, legal requirements and AI ethics.

Within the EMPAIA project, the DAI-Labor is responsible for the development of the data platform and is intensively involved in the dynamic matchmaking of data, AI solutions and experts. Furthermore, the theme of explainable AI (XAI) lies in the foreground of the project. The aim is to bring transparency to the AI black-box model, thereby supporting the validation process and providing a means to establish trust in AI-assisted diagnostics for both physicians and patients.​