Pilot Application 2: Algorithm for Identifying Treatment Options

Realization of a Prototype Workflow for Identifying Epileptic Lesions Requiring Treatment Using MRI Data

Starting Point

Neuroradiological sectional imaging (computed tomography; magnetic resonance imaging, MRI) of the brain is of central diagnostic importance in numerous neurological diseases. The outcome directly impacts therapeutic decisions and indirectly influences the course of disease. However, the range of epileptogenic lesions is not commonly known among radiologists, which can lead to missed treatment opportunities. Doctors who perform neuroradiological cross-sectional imaging (usually radiologists) are reaching the limits of their ability to do this job as their workload increases and examination procedures become more complex. Therefore, it is hoped that AI-supported evaluation methods will be able to support the detection of therapy-relevant findings in cross-sectional imaging.

The pilot scheme aims to achieve AI-supported MRI diagnoses and assess how radiologists and neurologists incorporate the findings into their care process. This will be tested in two studies (retrospective and prospective) for its practical feasibility and acceptance.

Approach

Part 1:
Process Model for Interaction at Human-AI-Human Interfaces in the Workflow (IAW)

Part 2:
Interview and dialogue guidelines for the participatory involvement of employees, works councils, staff councils, and other co-determination bodies (GA RUB/IGM)

Part 3:
Role-Based Training for AI Developers and AI Users Along the Competency Dimensions (IAW)

Scientific Publications

Gersch, M., Meske, C., Bunde, E., Aldoj, N., Wesche, J., Wilkens U. & Dewey, M. (2021): Vertrauen in KI- basierte Radiologie – Erste Erkenntnisse durch eine explorative Stakeholder-Konsultation. In: Bruhn, M. & Hartwich, K. (Hrsg.): Künstliche Intelligenz im Dienstleistungsmanagement. Forum Dienstleistungsmanagement (S. 309-335), Band 22, Springer Nature, 2021. https://doi.org/10.1007/978-3-658-34326-2_12

Thewes F., Langholf V., Meske C., Wilkens U. (Hg.) (2022): Towards a process model for cross-domain AI development – Insights from neuroradiological imaging. Unter Mitarbeit von Gesellschaft für Arbeitswissenschaft e.V. Frühjahrskongress 2022. Magdeburg: GfA-Press (Technologie und Bildung in hybriden Arbeitswelten).

Langholf, V.; Ranft, A.; Will, L.; Denz, R.; Schwarz, J.; Syoufi, M.; Rath-Manakidis, P.; Kämmerer, M.; Kremers, M.; Mosig, A.; Wilkens, U.; Wellmer, J.: Multi-stakeholder AI Ethics in Radiology: Implications for integrated technology and workplace design

Langholf, V.; Ranft, A.; Will, L.; Denz, R.; Schwarz, J.; Syoufi, M.; Rath-Manakidis, P.; Kämmerer, M.; Kremers, M.; Mosig, A.; Wilkens, U.; Wellmer, J. (2026): KI-Ethik im Multi-Stakeholder-Umfeld in der Radiologie. In: I4S 1/2026: Angewandte KI-Ethik am Arbeitsplatz. Eine gemeinsame Verantwortung – von der Radiologie und Sprachtherapie bis zur Montage . Industry 4.0 Science.