Cardiovascular diseases such as heart attacks and strokes are responsible for the majority of deaths in the Western world. In Germany alone, nearly 350,000 people die from them each year. Pathological narrowing and aneurysms of the blood vessels are often the underlying causes. Currently, these diseases can only be diagnosed with expensive examinations performed by experienced specialists, meaning they are only carried out in cases of acute suspicion. The Department Life Science Engineering at the Technische Hochschule Mittelhessen now aims to develop a diagnostic device with the project "AngioDiagnostics – Early Detection of Risk Factors of the Cardiovascular System Using Homomorphically Encrypted Machine Learning Algorithms." This device will enable simple, cost-effective, and non-invasive screening – even at the general practitioner's office. Simultaneously, the data will be secured using an innovative encryption method. The Hessian Ministry for Digital Affairs is funding the project with €710,750 from its Distr@l program. Digital Minister Prof. Dr. Kristina Sinemus presented the funding notification to the team led by Prof. Dr. missing in original text] today. Stefan Bernhard presented the document and learned about the project.
“This project demonstrates that technological innovation can concretely save lives – without compromising data protection. This combination of medical Further and data security sets standards for the future of modern, patient-centered healthcare. At the same time, the project exemplifies the transfer of research into practice: an innovative diagnostic system that combines AI and data protection in an exemplary way. Such projects are crucial for further developing Hesse as a location for safe and responsible medical technology. The fact that this will be realized in a spin-off from the university after the project's completion further strengthens Hesse's start-up ecosystem,” said Digital Minister Sinemus.
Dangerous bulges in the vessel wall, known as aneurysms, often go undetected for a long time because they frequently cause no symptoms and are only discovered by chance. Existing diagnostic methods such as Doppler ultrasound, CT scans, or MRIs are expensive and complex, and are therefore only used when there is a specific suspicion. The team at "AngioDiagnostics" is therefore developing a diagnostic device that enables non-invasive, cost-effective, and data-secure early detection of aortic aneurysms. In the future, it will allow general practitioners to regularly screen their patients—and thus detect life-threatening vascular changes at an early stage.
The underlying technology combines insights from mathematics, computer science, machine learning, and medical signal processing. It is based on so-called photoplethysmography signals (PPG), which are evaluated using intelligent algorithms. The funded project focuses on the further development of these algorithms and their implementation in a functional medical device. Furthermore, the developed data security architecture is intended to serve as a model for other medical applications.
Pioneers of data protection-compliant AI solutions in healthcare
A key focus of the project is the security of sensitive patient data. The team is relying on homomorphic encryption, one of the most promising technologies currently available for combining data protection and data utilization. This encryption method allows calculations to be performed directly on encrypted data without requiring prior decryption. This enables the secure processing and analysis of medical data without third parties gaining access to the original data. Highly sensitive patient data, such as genome or diagnostic data, can thus be securely stored and processed in the cloud, and medical institutions can securely exchange and jointly analyze data without compromising confidentiality. Furthermore, AI algorithms can be trained and applied to encrypted data to recognize disease patterns – without anyone needing to view the underlying data. The practical application of homomorphic encryption in medical devices is still rare due to technical and regulatory hurdles. With "AngioDiagnostics," it is now being tested for the first time in a concrete medical application scenario – paving the way for data protection-compliant AI solutions in healthcare.
“Every research project begins with the desire to help people. Ruptured aneurysms can be immediately life-threatening, often resulting in lifelong disabilities. Low-threshold diagnostics can prevent a great deal of suffering – and certainly also reduce costs in the healthcare system,” said project leader Prof.
Dr. Stefan Bernhard, Professor of Medical Technology and Numerical Mathematics at the Department of Life Science Engineering at THM, explained that the transformation of a research idea into a market-ready technology depends not only on perseverance but, above all, on suitable partners and a good team. "Ultimately, the idea must also stand the test of reality," Prof. Bernhard said, looking ahead to the next steps toward a reliable and safe product for everyday use in general practitioners' offices.
background
As of mid-October 2025, 164 projects with a total funding volume of approximately €53 million have already been approved. This is supplemented by around €29 million in co-financing from the private sector and approximately €4.5 million from the ERDF 21+ program. Detailed project information is available on the LIDIA platform.