The joint research and development project of GE HealthCare Kft. and the University of Debrecen will make medical diagnostics and pharmaceutical research more efficient with the help of Artificial Intelligence. Their solution, based on Natural Language Processing algorithms, supports the analysis and organisation of visual and textual findings, which can be used to target patients for inclusion in clinical research, among other things.

The development, supported by the National Research, Development and Innovation Fund, will reduce the burden on healthcare workers, improving the overall efficiency of clinical care and research.

GE HealthCare, a company that has been present in Hungary for 25 years, is an innovator of medical devices, digital developments and molecular diagnostic solutions. Its activities include the application of medical imaging, workflow management and AI in healthcare. For years, the University of Debrecen has been an important strategic partner of the company, which has a strong focus on the healthcare industry. In their latest joint project, they have developed a solution, under the professional leadership of the Faculty of Informatics, which allows the analysis of image and text diagnostic findings in a broader context, using the newly developed Natural Language Processing (NLP) algorithms designed for Hungarian-language texts. Natural language analysis combines best practices from linguistics and computer science to enable computers to process, analyse and interpret human language. These methods are not new, they are used by chatbots and digital assistants, but they have not yet been applied more widely to the analysis and research of medical findings.

The development is capable of analysing and jointly interpreting both textual findings and diagnostic imaging output, not only to facilitate smoother patient care, but also to improve the efficiency of specific drug discovery processes by targeting patients for clinical research. Clinical trials are used by pharmaceutical companies, medical device manufacturers and research institutes to test the efficacy and safety of various medicines, in addition to testing the latest therapeutic procedures. These increasingly complex processes require the identification and involvement of specific patient groups, which is a time-consuming process. Such a procedure can take a long time, partly because of the complexity and manual nature of sample collection. Thanks to the data mining platform developed by GE HealthCare and the University of Debrecen, the time spent on this can be significantly reduced, as the system is able to identify potential subjects for appropriate clinical trials through automated processes. This development benefits multiple stakeholders: patients gain access to the latest therapeutic procedures, clinical trial sites can acquire valuable professional experience while reducing costs, and the medical field as a whole advances by moving closer to solving unresolved health challenges.